INTERNATIONAL SOCIETY FOR SOILMECHANICS

AND GEOTECHNICAL ENGINEERING TC 40.

FORENSIC GEOTECHNICAL ENGINEERING OCTOBER 2009

INTERNATIONAL SOCIETY FOR SOIL MECHANICS AND

GEOTECHNICAL ENGINEERING

FORENSIC GEOTECHNICAL ENGINEERING

EDITED BY

Dr.V.V.S.Rao

CHAIR 2005-2009 TC40: FORENSIC GEOTECHNICAL ENGINEERING

OCTOBER 2009

FORENSIC GEOTECHNICAL ENGINEERING

CONTENTS

PREFACE

1. INTRODUCTION TO FORENSIC GEOTECHNICAL ENGINEERING V.V.S.RAO

2. WHAT IS FAILURE PETER DAY

3. CHARACTERIZATION OF DISTRESS ANIRUDHAN,I.V.

4. COLLECTION OF DATA PETER DAY

5. COMPILATION OF DATA V.V.S.RAO

6. LABORATORY TESTS ROBINSON,R.G

7. BACK ANALYSIS SIVAKUMAR BABU

8. BACK ANALYSIS OF SLOPE FAILURES POPESCU,M.E.et.al

9. BACK ANALYSIS IN FGE RICHARD N. HWANG

10 GEOTECHNICAL FAILURE:THE CAUSE JAN HELLINGS

11. INSTRUMENTATION AND MONITORING IWASAKI,Y.

12. LEGAL PROCESS AND JURISPRUDENCE SAX SAKSENA

13. POTENTIAL ROLE OF RELIABILITY PHOON,K.K,et.al.

CASE HISTORIES:

1. TURNING HINDSIGHT INTO FORESIGHT MIKE MARLEY

2. FAILURE IN CONSTRUCTION OF AN UNDERGROUND STATION RICHARD HWANG

3. FORENSIC ENGINEERING FOR UNDERGROUND CONSTRUCTION BROWN,E.T.

4. IMPORTANCE OF UNDERSTANDING LANDFORMS THOMPSON,R.P.

5. MONITORING IN FORENSIC GEOTECHNICAL ENGINEERING IWASAKI,Y.

6. EXPERTS’ DILEMMA DIRK LUGAR

7 R.E.WALL DISTRESS SIVAKUMAR BABU

8. DISTRESS TO REINFORCED EARTH WALL EMBANKMENT SITHARAM,T.G, etal

9. ANALYSIS OF SOIL NAILED STRUCTURES RAJAGOPAL,K.,etal.

10. FAILURE ANALYSIS OF DAMAGED RE WALL MURTHY,B.R.S.

11. APPROACH EMBANKMENT ON SOFT GROUND MADHAV,M.R.,et al.

12 COLLAPSE OF SOILNAILED WALL SITARAM,T.G.

13. THE CASE OF TILTED CHIMNEY RAO,V.V.S.

14. INFLUENCE OF VIBRATIONS SANTOSH RAO, N.

ANNEXURE.

PREFACE

Forensic Geotechnical Engineering (FGE) deals with the procedure to be followed while analyzing

a distress / failure in a structure which is attributed to geotechnical origin, not only from technical,

but also from legal and contractual viewpoints. In cases of remedied installations the analysis and

evaluation of adopted remedial measures may be subjected to legal scrutiny with regard to their

effectiveness and economy. Geotechnical based distress in structures due to natural hazards

including seismic damages also come under this purview. The commonly adopted standard

procedures of testing, analysis, design, and construction may not be adequate for forensic analysis.

The test parameters and design assumptions must simulate the actual conditions encountered at

site. Thus, in the forensic investigations, every micro aspect of the design, construction and

maintenance actions are studied in detail to analyze what, when, how, and why something went

wrong and more importantly, who is responsible for it. This procedure not only assists in

litigations, but also helps in improving the standards of geotechnical aspects of a project.

In order to develop this subject, Prof. Pedro Seco e.Pinto, President, ISSMGE established during

February 2006 a Technical Committee (TC40) on Forensic Geotechnical Engineering. The

composition of the committee is given in the annexure. The committee proposed to prepare a book

on FGE as a guide to geotechnical professionals.

The present book highlights the aspects of a forensic investigation illustrated by a number of case

studies. This is a first attempt and more detailed one has to follow.

While compiling the chapters, the papers presented in the workshop on FGE conducted in

Bangalore on 12th Sept.2009 have been included. This w/s was organized by the Indian Institute of

Science in collaboration with TC 40 and Karnataka Chapter of IGS.

Chapter 1.

INTRODUCTION TO FORENSIC GEOTECHNICAL ENGINEERING

Dr.V.V.S.Rao Nagadi Consultants Pvt. Ltd., New Delhi, India,

dr_rao@nagadi.co.in INTRODUCTION Forensic analysis in geotechnical engineering involves scientific and legalistic investigations and deductions to detect the causes as well as the process of distress in a structure, which are attributed to geotechnical origin. Cases of remedied installations where the analysis and evaluation of adopted remedial measures with regard to their effectiveness and economy may be subjected to judicial scrutiny also fall under this purview. The normally adopted standard procedures of testing, analysis, design and construction are not adequate for the forensic analysis in majority of cases. The test parameters and design assumptions will have to be representative of the actual conditions encountered at site. The forensic geotechnical engineer (who is different than the expert witness) should be able to justify the selection of these parameters in a court of law. Hence he has to be not only thorough in his field of specialization, but should also be familiar with legal procedures. This paper presents principles of planning and executing a forensic investigation SCOPE While investigating any distress, the engineer should meticulously follow a well planned programme. The scope of work would broadly be under the following heads:

a. Compulsory tasks i. survey and documentation of the distress ii. scrutiny of all design documents including design criteria chosen iii. review results of original geotechnical investigations, their analysis, and selection of

design parameters. iv. study the field reports of construction v. interview persons involved in planning, design, construction and

performance monitoring, etc. b. Optional tasks:

i. perform additional investigations ii. develop and conduct special tests iii. non-destructive testing of structural element

c. Analyze all data and evaluate i. the distress history ii. causes of distress iii. identify the shortcomings in the original investigation and analysis

d. Report i. authority and scope ii. history iii. summary of original documents iv. data collected v. interviews vi. meteorological information vii. earthquake viii. investigations performed , their methodology and their results ix. analysis x. conclusions.

TYPES OF DISTRESS The visual distress which commonly occurs in normal structures and the most probable causes are listed in the following table

Structure Visual distress Causes Buildings and bridges Cracks Structural Geotechnical: settlements heaving uplift vibrations Tilts Differential settlements Earthquakes Collapse Excessive loading Soil erosion Repetitive loading, fatigue Retaining structures Lateral movement Inadequate base resistance Tilting Differential settlement between toe and heel Excessive surcharge on the backfill Excessive water pressure due to poor drainage Slopes Excessive settlements Improper compaction Slope failure Settlement of virgin strata Longitudinal cracks Erosion due to water, rainfall Inadequate drainage

Apart from the above causes, most important causes would be inadequate and/or inappropriate soil investigation, selection of design parameters and use of inappropriate theories DIAGNOSTIC TESTS. After identifying the cause of distress, the following questions arise :

a. Has the distress fully occurred and if not, how much more can be expected? Quantify. b. What were the precise causes for distress? c. Whether the soil underwent same stress-deformation history as was anticipated? If not, what was

the actual history at site? d. The effect and efficacy of remedial measures on the soil+structure behaviour.

To answer these questions, detailed tests have to be conducted both in the field and in the Laboratory. The choice of tests will normally be from among the following tests depending upon the problem. A. Field tests

i. Borehole investigations including SPTs to a depth deeper than the influence zone ii. Cone penetration tests iii. Load tests iv. Special tests like, pressuremeter tests, vane shear tests, seismic or dynamic tests

B. Laboratory tests i. Triaxial shear tests; to simulate the actual field conditions, these tests should be done on stress

increment basis on partially saturated sample. The effect of fluctuation in degree of saturation on the deformation behaviour of the soil should also be investigated. In case of clays, the field stress history is also to be considered

ii. Repeated cyclic shear tests, in cases like water towers, bridges,etc.

iii. Large deformation tests, to assess the residual strength and magnitude of final deformations in cases of slopes,etc..

iv. Compaction tests with different compaction energies v. Permeability tests

In case of residual and wind deposited soils, the effect of change of soil structure due to loads, both static and moving ones In all cases it is advisable to conduct regular borehole investigations and evaluate the sub-soil profile. ANALYSIS After collecting all data detailed analysis can be done to evaluate the design parameters. Use of empirical relationships should be avoided, unless their validity in the particular site is established. The analysis should be based on: - limit conditions - partial factors of safety - equilibrium state vis-à-vis flow state - liquefaction potential - critical void ratio in compacted fills With these design parameters the load – deformation history of the soil+structure combine can be reconstructed. This process will lead to identification of the causes of distress. A suitable and economically viable remedy can then follow. LEGAL ISSUES In the entire process of investigation, the forensic engineer should be careful to ensure that all the experimental and analytical procedures as well as the selected parameters for tests and analysis fully conform to the field conditions. The report should be comprehensive and intelligible to a legal person also. It is advisable to avoid “hi-tech” terminology and strong verbs like-should, must, etc. As far as possible, it is better to avoid too many details in the main text. At the same time, the report should have sufficient details for the client to give a comprehensive brief to the executing agency. One should realize that the association of the engineer with the project is based on the principle of-“contract of skill”. Hence the consultant should ensure competent and reliable advice to the client. It is also imperative for the consultant to explicitly detail the risks that might be involved or expected in using the conclusions and recommendations. The consultant should be aware of the importance and implications of the following guidelines (ref: Guidelines for the provision of Geotechnical Information in Construction Contracts- The Institution of Engineers, Australia, 1987.): Facts: These have to be true and should not be erroneous.

- exploration locations - samples and cores available for inspection - lithological descriptions of soils and rocks - measured water tables - test results

Interpretations: the skill of the engineer is judged here

- borelogs - inferred stratigraphy between boreholes - properties of various layers for use in the analysis

- seismic interpretations yielding velocity and layer depths Opinion: may or may not be disputable

- assumptions - judgement based on facts and interpretations

Negligence: obviously, very serious - performance of investigations

- description and analysis of information - communication - accuracy, in general

Overall, it is emphasized that application of standard of skill and care is expected from a professional, irrespective of quantum of remuneration. However, the liability of the consultant is also limited as the owner pays for the “skill and service” and not for “insurance”. REFERENCES Leonards, Gerald A., “Investigation of failures”, Journal of Geotechnical Engineering Division, ASCE, GT 2,

Feb. 1982. Task Committee on Guidelines for Failure Investigation, “Guidelines for Failure Investigation”, ASCE,1989. Green, D.C., “Principles for providing Geotechnical Data in Construction Contracts”, Conference on Dams, Queenstown, Tasmania, 1988, (also in Ancold Bulletin No. 81.) Robert W Day, “Forensic Geotechnical and Foundation Engineering” Mc Graw Hill, 1998.

Chapter 2

WHAT IS FAILURE

Peter Day Jones & Wagener Consulting Engineers,

Livonia, South Africa, 2128 day@jaws.co.za

1. BACKGROUND

Forensic geotechnical engineering involves the application of scientific methods and engineering principles in the investigation of failures of geotechnical origin, not only from a technical view point but also with the possibility of legal proceedings in mind(Rao,2005)

Forensic investigations differ from conventional geotechnical investigations in that they are retrospective. They seek to explain what has happened rather than to predict future performance. A further distinguishing factor is that, following a failure, there is an urgency to clean up the site and rebuild or repair the works. This limits the time available for investigation and makes it essential that all relevant data is recorded before the evidence is removed. There is also the added difficult that the ground conditions may have been altered by the failure and testing of the ground in areas that have not failed is not always representative.

2. WHAT CONSTITUTES FAILURE ? 2.1 General

When a structure collapses, there can be little doubt that a failure has occurred. However, legal disputes often arise in cases whee the distress is considerable more subtle. Leonards(1982), defines failure as an “unacceptable difference between expected and observed performance”1. This performance may either involve the stability of the structure, its appearance or its ability to fulfil its intended function in either the short or long term. The distress of geotechnical work can be classified in a number of ways ranging from the type of structure involved, the nature of the distress, the consequences of failure and many others. However, from a legal perspective, it is preferable to have a classification that corresponds with the performance requirements laid down by international standards or codes of practice. Classification of distress in this manner will assist in determining the acceptability or otherwise of the observed performance of the structure.

Many modern codes of practice are based on limit states design principles. These codes clearly define the standards of performance required for various design situations or limit states. Eurocode En 19902 defines the two main limit states as the ultimate and serviceability limit states.

1 Leonards,G.A. (1982). “Investigation of Failures.” Journal of the Geotechnical

Engineering Division, ASCE, 108 (GT2):187-246

In broad terms , the ultimate limit state deals with the stability of the structure(or of its

component parts) whereas the serviceability limit states must be satisfied. This firmly establishes the principle that unsatisfactorily performance of a structure in terms of serviceability is equally as much a failure as its collapse.

The criteria applied in adjudicating compliance with these limit states differs as described below.

2.2 Instability

According to EN 1990, the ultimate limit state concerns the safety of people and/or the safety of the structure. Ultimate limit state failures are generally easy to recognise as they involve visible collapse or instability of a part or the whole of the structure3.

For geotechnical engineering, EN1997-1 (Geotechnical Design) 4 requires consideration of the following ultimate limit states:

• loss of equilibrium of the structure or the ground, considered as a rigid

body, in which the strengths of structural material and the ground are insignificant in providing refinance (EQU);

• internal failure or excessive deformation of the structure or structural

elements. Including e.g. footing, piles or basement walls, in which the strength of structural materials is significant in providing resistance (STR);

• failure or excessive deformation of the ground, in which the strength of

soil or rock is significant in providing resistance (GEO)

• loss of equilibrium of the structure or the ground due to uplift by water pressure (buoyancy) or other vertical actions (UPL)

• hydraulic heave , internal erosion and piping in the ground caused by

hydraulic gradients (HYD)

Although the design requirement to be satisfied in the case of each particular ultimate limit state differs, the basic requirement is that the design action effect (typically the effect of loads or deformations ) should not exceed the design resistance (typically its strength).

In the case of abnormal events (such as accidental impact, fire, explosions, earthquakes, and the consequence of human error), there are additional requirements for structural integrity and robustness in terms of which the damage to the structure should not be disproportionate to the original cause. The structure should be capable of withstanding local damage this causing widespread collapse.

2.3 Serviceability

EN 1990 describes serviceability limit states as those that concern the functioning of the structure under normal use, the comfort of people and the appearance of the works.

Verification of the serviceability limits states requires consideration of : • deformations that affect the appearance the works, the comfort of the users and the

functioning of the structure. • Vibrations that cause discomfort to people limit the functional effectiveness of the

structure, • damage that is likely to adversely affect the appearance, durability or functioning of

the structure.

In the case of the serviceability limit state, the criteria to be applied is that the design action effect must be less that the limiting design value.

It is often considerably more difficult to adjudicate whether a serviceability limit state has been transgressed than is the case with the ultimate limit state as the assessment of the serviceability limit state requires the establishment of the performance criteria (or limiting design values). Although no such criteria are given in the Eurocodes., they may be specified in the National Annexes or in other national standards. An example of this is the categorisation of damage to single storey masonry strucutres5 or the performance requirements and categorisation of damage given in the Australian code of practice for foundations for residential buildings6.

3. CAUSES OF FAILURE

During a forensic investigation, it is important not to anticipate the outcome of the investigation by pre-judging the cause of distress or failure. However, the investigator should be aware of the most common causes of geotechnical failures to ensure that the data collected is sufficiently comprehensive to enable the problem to be analysed from all angles.

A study of several published case histories suggests that the most common causes failure are as given in Table 1. It is often found that failures involve two or more of the elements given in Table 1 acting in combination. In geotechnical failures, water or water pressure is one of the most frequent contributing factors.

Table 1 : Most common causes of geotechnical failures

Cause Description

Inadequate geotechnical investigation

Budgetary or programme restraints can result in insufficient investigation being undertaken to adequately model of the conditions on the site. Alternatively, even the most comprehensive investigations may fail to reveal critical conditions that affect the geotechnial behaviour of profile.

Incorrect parameters This can occur for many reasons, including : - poor sampling and testing procedures - selection of inappropriate parameters for particular desig situation (e.g. mean values, lower characteristic values or upper characteristic value) - underestimation of variability of soil properties

Inappropriate analysis model

Failure to recognise the critical failure mechanism , e.g. drained v. undrained failure of slopes or foundations, internal stability v. external stability of reenforced fills.

Underestimation of actions Either the magnitude, distribution or combination of actions (forces or displacements) incorrectly assessed, particular load case or combination not considered, use of structure changed over lifetime.

Unexpected groundwater regimes or changes in moisture content

Changes in ground water levels can increase the loading on the structure and decease the shearing resistance of the soil. Seepage forces can also have an adverse effect on stability. Changes in the moisture content of partially saturated soils can cause softening, heave or collapse settlements.

Sub-standard workmanship or materials

Required construction procedures (including sequence and timing) not followed, specification requirements not met, inappropriate construction techniques employed, material properties not in accordance with design assumptions.

Abnormal events not catered for in design

Extreme meteorological events (including temperature, precipitation or wind), accidental impact, errors in construction or use of structure.

3. CLASSIFICATION OF DISTRESS 4.1 General

Classification of distress should not be confused with the identification of its likely causes. It is simply a method of classifying the damage to the works prior to any

investigation into its origin. It is simply a case of what has happened to the structure and not why it has happened.

When viewed in this manner, distress can be classified in terms of two criteria, namely the manifestation of the distress and its severity.

4.2 Distress of a structure or geotechnical works can manifest itself in two ways, namely

loss of stability and loss of serviceability. These are may be broadly linked to the ultimate and serviceability limit state requirements of most codes of practice.

Ultimate limit states conditions include the manifestation of distress given in Table 2. Manifestation Geotechnical examples

Failure of excessive movement of the ground

Bearing capacity failure, slope instability

Failure or excessive movement of the structure or structural component

Failure of a retaining wall in bending or shear, structural failure o fa pile shaft

Loss of equilibrium of the structure or ground considered as a rigid body

Overturning of a rigid structures or bodily translation of a reinforced soil mass

Loss of equilibrium of a structure due to uplift by water pressure

Floatation of a tanked basement structure or buried tank

Hydraulic heave, internal or piping in the ground due to hydraulic gradients

Base instability or upward movement inside supported excavations or caissons, headward erosion of dam embankments

Serviceability limit state conditions include the manifestations of distress given in Table 3. Table 3 : Manifestations of distress - serviceability limit state

Manifestation Geotechnical examples

Deformations that affect the appearance of the works or their functionality

Total and differential settlement of foundations causing cracking, jamming of doors etc.

Vibrations of transient movements that cause discomfort to people or the function of structure

Dynamic response of machine foundations or supporting structures

Damage that affects the appearance, durability or functioning of the structure

Water ingress into structures, excessive seepage from dams, cracks in piles or retaining structures that permit corrosion of reinforcement, etc.

4.3 Severity of Distress The severity of distress may be classified according to the degree to which the functionality and stability of the structure is impaired and the ease with which it can be repaired as indicated in Table 4. Table 4 : Severity of Distress

Severity Degree of Impairment Remedial work required

Negligible Some visual impairment but structure/works fully functional

Cosmetic only

Slight Structure/works still serviceable but users inconvenienced. Appearance affected.

Minor non-structural repairs and redecoration

Moderate Structures still deemed safe but use of structures/works restricted. Damage clearly visible.

Structural repairs required

Severe Structures still standing but no longer serviceable. Access restricted

Significant structural repairs of partial reconstruction required.

Very severe Structure in state of partial or complete collapse

Demolish and rebuild

References:

1 Leonards,G.A. (1982). “Investigation of Failures.” Journal of the Geotechnical Engineering Division, ASCE, 108 (GT2):187-246

2 EN 1990:2002. Eurocode - Basis of structural design, European Standard. European Committee for Standardisation, Brussels. 3 In this context, the term structure is taken in its broadest sense and includes

constructed works such as slopes and embankments, or the natural ground in the case of landslides.

4 EN 1997 -1:2004. Eurocode 7 : Geotechnical Design - Part 1 : General Rules, European Standard. European Committee for Standardisation, Brussels.

5 SAICE ? Joint Structural Division (1995) Code of Practice for Foundations and Superstructures for Single storey Residential Buildings of Masonry Construction. 1st Edition.

6 AS 2870-1996: Residential Slabs and Footings Construction. Standards Australia,

1

Chapter 3 CHARACTERIZATION OF DISTRESS AND SOME CASES

Anirudhan I.V. Geotechnical Solutions

19 Usha Street, Dr. Seethapathy Nagar, Velachery, Chennai – 600 042 anirudhen@eth.net

ABSTRACT

Types of distress found in structures due to different geotechnical reasons are briefly outlined in this paper. The patterns of distress generally found in near failure cases or at a stage where a remediation is required are typical and these patterns can be used to arrive at the possible remedial measures. Legal disputes arising out of the distress cannot be simply resolved by these patterns and in such cases, the patterns are used for formulate further investigation of the problem and then fixing the responsibility. This step of the forensic geotechnical engineering is not dealt with in this paper. Three cases of distress identification during construction are discussed in the second part of the paper. Deficiencies in the design, construction methods and the deficiencies in the geotechnical investigation are the causes for these distresses. The attempt to identify the causes for distress and then to recommend remedial measures in these cases may not be very scientific and legally acceptable as the problems were not looked into from a legal point of view.

1 INTRODUCTION

Any geotechnically induced distress in a structure is a result of deformation or displacement of the soil over which the structure is supported. The relative stiffness of the structure vis-à-vis the soil supporting it plays a major role in most of the distress. Unfortunately the relative stiffness itself is constantly changing due to several factors, which includes the influence of the structure itself. Deformations and displacements due to swelling and shrinkage point to a continuing, but erratic change in the relative stiffness in the soil by the influence of environmental factors. Changes in the otherwise harmonious relative stiffness because of the changes in in-situ stress conditions due to excavations, dewatering, blasting, etc., also cause distress. Earthquake induced distresses are the results of such changed stress environments, but with a distinct feature of stress reversals at very rapid successions.

Most of the structures have complex dimensions and properties that cannot get along with more complex behaviour of soil due to the changes imposed by the structure. The limits put forth on such several changes within the soil as well as in the structure to some extend prevent the distress. Several studies on these limits revealed serious limitations on generalising them and only resulted in more and more limiting parameters (Boone, S.J., 2001).

While the analysis of distress is complex and requires in depth case to case study, patterns of distress can point to an initial assessment of the possible cause that initiated the distress. Characterising the distress with respect to such different causes may be the first step towards such detailed analysis.

2

In forensic geotechnical engineering, emphasis is given on the study of causes and remedial measures after a structure is undergone undue distress and deformation significantly reducing the utility of the structure. Often, the symptoms and initial stages of distress are ignored and patched up leading to a major distress making it more complex for further remediation. Most of the classical examples of back analysis cited by Madhav (2003) are well planned failures helping a step to step back analysis leading to constructive conclusions. The slope failures, testing of foundations to large deformations (footing load tests, pile load tests, retaining wall load tests, etc.), piping failures, etc. have less complex modes or the modes can be simplified significantly.

The symptoms that caution the designer can be identified during several stages of construction itself helping the designer to take corrective steps to prevent a major distress after the completion of the structure. The defect can be in the basic design, in the investigation of sub-soil conditions that can be revealed during a deep excavation, and also in the construction of foundation including the defect in the workmanship and wrong methods. The back analysis in such case is not very complex in terms of geotechnical engineering. Such cases are complex mostly because of concealment of various factors relating the construction procedures and the symptom of distress.

This paper attempts to briefly put the reasons for distress and then profiles possible distresses in various geotechnical problems cases. Since the distress is often exhibited in the form of cracks, some of the common features of these crack patterns in relation with the reason for such distress are identified and explained. Starting from lightly loaded conventional buildings and multi-storeyed buildings built on conventional shallow foundations and special foundation, distress patterns for structures like retaining walls, deep excavations, bridge supports, high embankments, roads, etc. are outlined. Earthquake induced distresses are not reviewed here.

The second part of the paper discusses three cases where the major distress is prevented identifying (i) a wrong construction procedure. (ii) a design defect, and (iii) a wrong assessment of the soil condition.

2 CAUSES OF DISTRESS

It is well understood that a construction, which does not undergo displacement, settlement, or deformation is impossible. The practice is to adopt a design that limits these to allowable levels.

‘Limits’ to foundation movements / rotations / displacements, etc. were developed from various studies undertaken all over the world. These limits, as on today, have wide ranges and the debate is still on about the most significant parameter that would explain the cause of distress in structures. Skempton and MacDonald (Burland 1977) relate distress to angular distortion β. This implies that the damage results from shear distortion within the building, which is not necessarily the case. These studies separate load bearing brick walls from framed structures. Some of the studies introduced deflection ratio Δ/l into the realm of criteria for damage (Burland 1997). Burland in his earlier studies (Burland and Wroth, 1974) related the building damage to limiting tensile strain εlim along with L/H and E/G, the ratio between Young’s modulus and shear modulus, as the deciding parameters. The typical damage criteria with respect to individual crack width have been in use for a long (Burland & Wroth, 1974). Burland had however cautioned about the use of crack width alone as the criteria of damage. Studies by Boscardin and Cording (1989) later introduced limiting lateral strain εh also into the damage criteria.

3

Boone (2001) emphasised that simplified procedures using angular distortion, deflection ratio, lateral strain and limiting tensile strain would not be sufficient in the damage related studies of more complex structures we have today. Figures 1 & 2 reproduced from his case studies, will suggest that all these formulations do not really predict damage by distress and in turn it suggests that there is a risk of damage even after painstakingly taking care of these possible hazards.

‘Strain superposition method’ suggested by Boone (2001) assumes influence of the building stiffness on the final ground profile rather than accepting the general assumption of structure deforming to match the ground movement. The later may be conveniently accepted, though bit conservative, in the case of one to four storied structures usually found in urban areas. He analysed many cases and found that the said procedure agrees well in the case of these relatively simple structures. However exact ground movement cannot be predicted in a simple way.

3 DISTRESS PATTERN AND MODE OF DEFORMATION

Distress in a structure is often related to the foundation displacements due to various reasons. A distress is ‘sighted’ only when it exceeds a certain limit that can cause ‘cracks’ in the elements of the structure. It is understood that all the structures undergo deformations for various reasons and the ‘undue’ deformation not confirming to safety standards are usually referred as ‘distress.’ Most of the studies discussed above consider three modes of deformation, bending, shear and extension, for describing the crack pattern. These modes are illustrated in Figure 3, (Boone, 2001). The crack patters are typical. However, what one sees in a structure with undue distress are many cracks without a definite pattern that would fit into any of these three. Thus what one sees in a distressed structure are manifestations of two or more modes of deformations. Just observing the cracks and even measuring different

Horizontal strain εX103

Δ/l (

%)

Figure 2: Horizontal strain and deflection ratio and damage category and reported damages

(Burland 1997)

Angular distortion, βX103

Hor

izon

tal s

train

εX

103

Figure 1: Angular distortion, horizontal strain and damage category and reported damages

(Boscardin and Cording, 1989)

Figure 3: Separation of deformation modes (Boone, 2001)

4

parameters like angular strain, deflection ratio, lateral strain, etc. will not explain the root cause for the distress. Behaviour of some uniformly loaded structures such as storage tanks can be explained by standard deformation theories (Marr, Ramos and Lambe, 1982).

4 DISTRESS IN RELATION WITH MODE OF ‘FAILURE’ AND CAUSE FOR DEFORMATION

4.1 Bearing Capacity Failure:

Bearing capacity failure often occurs under undrained condition especially in the case of clayey soil at foundation. Punching through very loose fine sand is also possible when the foundation is over loaded almost instantaneously. The displacement resulting from a bearing capacity failure is often very excessive and the resulting distress is failure in most cases. There is an element of rotation displacement in such failures. Slope failure is one example. Bearing capacity failure will cause excessive differential movements that can cause shear. The result is vertical cracks or rotation failure. Simultaneous and uniform bearing capacity failure of all foundations of a structure is just not possible.

4.2 Compression Settlements - During Construction:

Loose sand deposits as founding soil can cause large settlements during construction. Large differences in the loading of different components of a structure can lead to undue differential settlements that are sudden and prohibitive. The resulting distress is also like bearing capacity failure because of large shear in the structure. Results almost vertical shear cracks and failure.

4.3 Consolidation Settlements – After Construction:

On the other hand, long term consolidation settlement in excess of permissible limits can cause differential settlements, during which the structure is allowed to adjust itself depending up on its capability to do so. Large storage tanks allowing to settle gradually under consolidation settlements, but avoiding a bearing capacity failure by not loading suddenly, are common. Large consolidation settlements can cause differential settlements in a structure with non-uniform loading and with the subsoil having varying thickness of the consolidating layer. The result is often diagonal cracks. Non-uniform foundation sizes, non-uniform column loading and relative positions of the foundations are the cause for large differential settlements. Prolonged settlements under secondary consolidation of grounds improved by pre-loading causing distress to structures with non-uniform loading are also reported.

4.4 Delayed Compression – Collapsible soils:

Ramanathan Ayyar & Jaya (2003) describe delayed failure in laterite as one of the reasons for mass collapse of wells in Kerala during 2001. Large cavities in lithomarge clay below laterite crust due to continuous water seepage through this soft soil is common. Collapse of lithomarge and the top crust is possible due to sudden rise in ground water level. Figure 5 shows such cavity of roughly 10.0m x 12m at a depth of about 2.50m below the surface found in construction site in North Kerala. The soil around the cavity was very soft and wet, whereas the crust, the soil above the cavity was very hard laterite that required mechanical excavation. The reason for such large cavity was then deduced as the loss of lithomarge (fine soil mss) through seepage into large excavations for

5

quarrying the laterites stones very close to the site. As seen form the picture it is very difficult to assess such situation in a large site where the exploratory boreholes are made at a normal close spacing 25m to 30m and the presence of such cavity just below a heavily loaded column will be a disaster.

Similar collapsible formations are reported in many parts of western Tamilnadu also. Some sand deposits in Rajastan are also reported to be collapsible. Distress in these cases may be leading to failure.

4.5 Squeezing of Soft Soil – Tooth Paste Phenomenon:

This ground movement is different from settlement or punching. Lateral displacement of very soft soil sandwiched between relatively strong soil / structure above and below is possible when the incremental stress in this soft soil is more. The displacement in this case also is often ‘sudden’ and the result is lateral displacement of the foundation apart from vertical movements. Tilting can be an associated feature.

Poulos (2003) describes a case of lateral flow of very soft soil due to adjacent excavation that caused undue lateral forces resulting failure in the deep foundation supporting a huge structure. Though this case is not strictly a case of squeezing of soft soil, sudden relief of confining pressure giving rise to increased K could be the reason for this flow.

4.6 Rotation and Sliding:

Rotation is related to bearing capacity failure and slope stability. Sliding of retaining structures can be independent of rotation, while sliding in a slope is also a rotation with very large curvature even though some models assumed a wedge failure. Wedge failure shall be termed as sliding. The distress noticed under these cases is in the form of relative ground movement, failure of retaining structures and embankments. In most of the cases, the distress is sudden. Many rotation failures of abutment piers manifested by bearing capacity failure because of overweighing back fill are reported.

4.7 Hydraulic Fractures such as Piping:

Continuous removal of fines from soil due to seepage or piping causes deformation in earthen structures such as reservoir bunds, earth dams, etc. The distress is noticed in the form of unequal settlements and local slope failures. These indications are warning of a major failure such as slope failure and sliding. A case of failure of a wide RCC channel due to piping is presented in Figure 5. Figure 5: Failure of a canal (and adjacent large sump)

because of erosion of soil initiated by piping due to seepage of water stagnated on the side of the channel, Ludhiyana

Figure 4: large natural cavities found in laterite formations, Pattuvam, Kannur, Kerala

6

4.8 Heaving:

Heaving is associated with deep excavations in soft clay or very stiff soil or excessive filling over a desiccated crust of firm soil of small thickness over a thick layer of soft clay. Heaving of excavation bottoms in the case of soft clay formations is common because of reduced effective stresses. Bulging of excavation sides is also common in the case of very soft deposits. Lateral earth pressure exceeding the effective vertical stress is common in such cases.

Heaving of ground due to lateral and upward movement of very soft deposits below relatively thin desiccated crust or better soil occurs when heavy area loads are placed. Heave occurs away from the loaded area and often develops gradually. Indications are displacements and resulting distress in small structures in the heaved area.

Deep excavations into very hard over-consolidated clay formations or sedimentary rock formations like shale and mudstone can result large up-heaving of excavation bottom. This is apart from possible swelling under saturation if the soil in-situ is partially saturated. Shrinkage cracks developed within the natural soil immediately below the excavation bottom and sides because of increased σ3 and subsequent saturation can cause considerable reduction in bearing capacity and increase in compressibility. Small footings placed over large common excavation area can be affected by this softening.

Even very shallow excavations in expansive soils results significant heaving of the excavation bottom by the migration of moisture under changed stress conditions. The heaving is aggravated by the presence of ground water within the capillary range. Lowering of ground water table alone is not reducing the evil of heaving in such cases as the process of heaving is not merely related to the presence of ground water table, but significantly related to moisture locked in the soil itself. Examples of disintegration of highly weathered shale upon shallow excavations are plenty (Figure 6).

4.9 Subsidence:

General ground subsidence can happen due to ground water lowering in relatively loose and soft deposits. These movements are typically gradual and slow giving enough warning to possible damage to structures in the affected area. The net result is however a more dense sub-soil, and this does not help in avoiding distress in the existing structures because of possible differential movements Mair (2003).

Subsidence due to ground movements during tunnelling and mining are often sudden because of the involvement of pore pressure. Such ground movements usually result in loosened soil. Similarly the ground movements due to reduction in density due to particle loss along seepage during heavy ground water lowering by dewatering are also sudden. Caving in of several wells in Kerala during 2001 after a relatively intense rains is an example (Ramantha Ayyar & Jaya, 2003).

Figure 6: Disintegration of weathered shale upon one cycle of drying (after excavation) and wetting (during construction)

Oragadam, Chennai

7

4.10 Swelling and Shrinkage:

Foundation displacements and slope failures are common where swelling soil is encountered. Moderately swelling soil experiences large volume reductions by loss of moisture. Displacements due to swelling and shrinkage will be highly varying between different parts of a building because of varying moisture build up or loss below different foundations. Usually footings in the corners are affected more because of relatively small loading and more exposure to environmental changes. Moisture variations within the building area in normal conditions will be small leading to small movements for foundations in the interior developing large differential movements between exterior and interior foundations.

The differential heaving associated with the swelling of soil has resulted distress in many structures Ramaswamy and Narasimhan (1978) had investigated a single storey building at Anna Nagar which has undergone severe cracking. The differential heave pattern of the building is shown in Figure 7 as indicated by the plinth line at the time of inspection.

Progression of moisture towards interior of buildings where swelling soil is used in backfilling the excavations and plinth can cause swelling of soil in the interior after a prolonged period. Interestingly the cracks usually associated with swelling of foundation soil is vertical and horizontal. Horizontal cracks between roof slab and load bearing wall on one side results when the cross wall foundation is subjected to swelling displacement. Heaving of plinth beams supported on swelling soil can cause hogging in the walls allowing to develop tensile cracks at the top portion of the walls. Outward movement of walls are also noticed due to rotation of plinth beams along horizontal axis.

Displacement due to shrinkage is also gradual similar to that due to consolidation settlement, but can be very erratic. Alternate shrinkage and swelling displacements cause distress in the structure over a period of time and this later manifest into a major crack or deformation. By the time this severe crack is noticed in the structure, many minor unnoticed cracks due to stress reversals could have been inflicted weak zones in the structure. Shrinkage of topsoil under drought or ground water depletion causes sinking of floors and development of additional stresses in plinth beams designed as ground beams. Foundations initially resting over medium stiff clays of moderate shear strength can displace due to large volume reduction under loss of moisture. Cases of failures are reported due to shrinkage of foundation soil under boiler foundations under severe moisture loss due to induced heat. These are occurring after a period and can be differentiated from other types of distress.

Figure 7: Differential Heave Pattern of The Building at Anna Nagar, Chennai (Ramaswamy and Narasimhan, 1978)

8

5 DISTRESS PATTERNS

5.1 Lightly Loaded Buildings on Shallow Foundation

5.1.1 Due to Differential Loading:

The most common distress found in lightly loaded buildings on shallow foundation is vertical and diagonal cracks below the window sills. This, usually very prominent and eye catching distress, is not strictly due to any geotechnical issue. The imbalance in the load distribution towards foundation due to large openings provided for the windows and the resulting tensile stresses in the generally brittle wall is the reason for such distress. This distress happens within few months from the construction (Figure 8).

5.1.2 Due to Differential Settlements:

If the diagonal cracks are towards the corner or to the edges of the walls, such distress could be associated with differential settlement within or between the foundations. Usually such distress appears in one or two years after the construction and involves the settlement of underlying clayey stratum. The floors are generally intact. The floor portion close to the walls may try to move along with the foundation causing distress little away from the intersection of the floor with the wall. This happens when the settlement is relatively large.

More settlement for heavier interior column footings is a clear indication of consolidation of underlying clay stratum. The apron provided around the outer walls is usually detached and cracked when the foundation settlement is excessive. In the case of severe settlements, several horizontal cracks associated with the main diagonal cracks will be sighted (Figure 9).

5.1.3 Due to Shrinkage in Soil:

Differential settlements occurring after a long period without any change in the loading of the structure shall be attributed to volume changes in founding soil due to desiccation or shrinkage. The foundations below the exterior walls and columns are more affected because of more exposure to environmental changes.

SHRINKING SOIL

SHRINKING SOIL

Figure 10: Distress in walls due to differential settlements (shrinkage in founding soil)

Figure 11: Distress due to differential settlements (shrinkage in founding soil)

Figure 8: Distress in walls due to differential loading

Figure 9: Distress in walls due to differential settlements (consolidation)

9

Shrinkage of founding soil aggravated by the water consuming trees and bushes rather than physical interference by the tree roots is also common. There is no definite pattern for such distress, but can be identified from the period at which it occurs. The floors are affected badly in such cases mainly because of the detachment from supporting soil that is dried up. The plinth protection around the outer walls shows severe distress by way of detaching from the walls and settles more than the walls (Figures 10 and 11).

5.1.4 Due to Swelling of Founding Soil:

Distress due to swelling / heaving of founding soil is more haphazard in comparison to the distress resulting from increased differential settlements. Generally wide cracks develop between the interface of the columns and walls or between the roof slab and the interior walls. Shallow foundations supporting the outer columns and walls are generally affected and the problem is aggravated by the fact that these columns and walls are with less loads in comparison with the interior ones. Hogging of the foundation portion below large window openings will result in the case of strip footing placed over expansive soil. This distress is similar to the one attributed to imbalance loading over the wall below the window sill. However, one of the major differences between these two distresses is sort of a cyclic nature of the distress due to swelling. Another difference is the appearance of distress above the window opening or at the interface of the lintel beam and the wall (Figure 12).

Lateral separation between window / door frames from the walls is a clear indication of differential heaving in the continuous footing (Figure 13).

Complete shear of the walls displacing it from its alignment is an indication of swelling of founding soil. Severe uplift pressure caused by the heaving soil can cause shear of infill walls. The distress is spread to upper floors when the structure is with load bearing walls and more distress will be noticed near the window openings. Detachment of interior cross walls from the outer ones is also very common in such case. Such distresses make the load bearing walls considerably weak and the in-fill walls loosing their confinement.

5.2 Lightly Loaded Buildings Supported on Short Piles / Under-reamed Piles

The major cause for distress in such structures is improper selection of the founding level for the supporting system. Even though the active zone with respect to severe moisture changes can easily be identified from local data, severe draughts and measures like rain water harvesting, etc. can trigger large changes in the active zone. Many case of shearing of wall and column interfaces are reported due to uplift of the supporting system because of changed environments. An extended draught period will result loss of moisture below the active zone causing tremendous swell pressure on the supporting system during next wet season. The floors and plinth walls properly treated for swelling pressures from the soil from shallow depths may remain intact in such cases triggering

BRICK WORK

BOWING WALL

DETACHING THE DOOR FRAME

Figure 13: Distress due to differential displacements by swelling of founding soil

SWELLING SOIL

Figure 12: Distress due to differential displacements by swelling of founding soil

10

distress in the interfaces. Horizontal cracks in upper portions of the infill walls, separation of infill walls from the roof, etc. are other distresses noted in such cases.

5.3 Structures Supported on Pile Foundation

Distress due to settlement of pile foundations is not generally expected unless very large pile groups are involved the supporting system. Large differential settlements between medium and large pile groups are reported because of underlying weaker layers. Excessive and continued settlements are also reported because of inadequate bearing stratum. Severe distress in the form of shear cracks on the beams and the columns supported on piles is reported where one structure on pile foundation abutting another structure on spread foundation is constructed. Consolidation of upper soft clay layers under the weight of the structure with spread footing transfers considerable amount of load through negative drag on the piles of the adjoining structure.

Failure of end bearing pile foundations under a load much less than the design load was reported from a site where the filling required for reclamation was the order of four metres. The piles failed in shear just under the tremendous negative drag caused by the consolidation of roughly 6.0m soft clay under the gravely soil fill of 4.0m. The failure occurred just in three to four months

5.4 Bridge Supports

The most common distress found in bridge structures is the level difference between approach length and the deck portion. This is definitely due to settlement of the approach portion. Generally there is a thick embankment or fill below the approaches and a well laid fill or embankment is not expected to undergo compression. Consolidation settlement of the founding soil is the major cause. The level difference or opening up of the deck joints over a pier suggests differential settlement between the piers or tilting of piers. Joints opening up between the deck slabs can also indicate uplift of piers founded on very hard clayey soil or shale formation when the river swell after a long draught period. Lifting of the first deck slab supported on the abutment pier is an indication of the rotation of the abutment pier under the embankment load.

6 CASE STUDIES OF DISTRESS DURING CONSTRUCTION

A Geotechnical Engineer will

1. ‘Foresee’ several scenarios 2. ‘Assume’ most reliable design parameters including the loading (with the help of an

investigation) 3. ‘Anticipate’ deviations from the assumptions 4. ‘Analyse’ such possible deviations 5. ‘Set the limits’ 6. ‘Estimate’ the stresses and strains for all the possible scenarios, and 7. ‘Reworks’ if the stresses and strains exceed the limit

before finalising a geotechnical design. However, often the inadequate data on the structure itself, continuous changes in the overall planning of the structure by the architects and owners, etc. compel the geotechnical engineer to finalise a design well before the finalisation of the final project the owner decides to execute. Very rarely the geotechnical design is re-looked or re-analysed to fulfil all

11

the requirements of the finalised project. At the same time there are instances where the geotechnical engineer failed to foresee the problems of a particular construction procedure that lead to unsatisfactory performance of the foundation. There are several cases in which the geotechnical investigation was inadequate of inaccurate that led to failures of near failures. The three cases described below try to illustrate some of these lapses.

6.1 The Case of ‘Lifted’ Piles

6.1.1 Problem

In 1994, in Chennai, about 1100 driven cast-in-situ piles were constructed to support a multistoried building with two basements. The construction of piles was done from (-) 4.0m level, whereas the basements were to be as deep as (-)9.5m and (-)11.5m. 600mm diameter and 550mm diameter piles extending to (-)19.0m were driven. Since the piles were to be cut-off below the basement levels, the concreting depth was limited to (-)8.0m, while the remaining length up to the working level was filled with sand.

When the excavation started after the construction of piles, the concrete top levels were found to be at much higher level than planned. The owner’s engineer thought that the piles were somehow lifted and there could be gap between the pile tip and the soil. Some piles were then stripped up to the cut-off levels and found reduction in pile diameter over a large length (Figure 14) and also discrepancies in the levels of reinforcement. The reinforcement top levels were however lower than the expected ones in most of the cases of discrepancy, while the cases of ‘cage lifting’ were also significant (Anirudhan, 1997).

6.1.2 The Distress

The major distress apprehended here was the loss of contact between pile tip and the supporting soil. The discrepancy in reinforcement top level brought the fear of cavities in the pile concrete at deeper levels.

6.1.3 Perusal of Construction Method

The foundation consultants had cautioned about ‘necking’ (reduction in pile diameter) in pile concrete over a portion between (-)6.5m and (-)10.0m where soft clay existed. Ground water table was at (-)4.50m during construction and the excess pore pressure developed in this soft clay could exert tremendous pressure on the green concrete forcing it to squeeze. Ideal option for preventing this was to provide concrete up to the working level irrespective of deep cut-off levels.

However, the concrete top level during construction was decided as roughly 2.0m above the cut-off level and filling of sand in the rest of the length was adopted to save cost. This remedial measure proved ineffective because of arching of sand within the casing pipe completely relieving the load over the green concrete that was supposed to have been provided with an overburden.

Similarly the excess hydrostatic pressure developed in the very dense clayey sand (residual type) at the pile tip during driving could be very significant, while the hydrostatic pressure inside the casing is zero. The minimum height of concrete column within the casing pipe was hence suggested

Figure 14: Necking in driven cast-in-situ piles, Chennai

12

as 11m before allowing even a small lift of the casing pipe so that the weight of green concrete is adequate to counter the hydrostatic pressure developed at the pile tip.

6.1.4 Actual types of distress

Now the types of distress in question were modified to the following

Reduction in pile diameter due to necking Possible reduction in pile diameter below the soft clay levels also Deterioration of concrete at pile tip due to excess pore pressure

Vertical displacement of reinforcement cage

6.1.5 Issues to be Investigated

1. Whether the necking was limited to the soft clay between (-)6.50m and (-)10.0m alone?. Even though there was no weak layers below (-)10.0m, the soil layers up to (-)16.0m were not very dense or stiff.

2. Whether the pile really lifted from the base? 3. Weather the reinforcement was intact in its position or was there a lift of the

reinforcement on account of the necking? 4. Whether the piles were deficient in their structural capacity due to reduction in cross

section? 5. Whether the piles had deficiency in end bearing resistance due to bad concrete at pile tip?

6.1.6 Investigation of Distress

A systematic investigation of these distresses was executed in the site.

The extend of necking in pile shaft: This task was achieved by systematic volume measurements with the aid of reasonably thorough pile driving records and, concreting & reinforcement details available for each pile executed. The top level of pile concrete was measured using precision level. Diameters at every 300 to 500mm length interval were then measured to a level at which the pile diameter was equal to the design diameter after the occurrence of first necking. The total volume of the concrete within the necking portion was found to be equivalent to the theoretical volume of the pile up to the theoretical concrete top level. Such observations were made for almost all the piles suggesting that there was no reduction in pile diameter (necking) below the first level of necking found at about (-)10.0m.

Lifting of pile from the base: The above measurements also expelled the fear of lifting of pile from the base

Reinforcement displacement: The fear was that when the necking in concrete took place resulting an upward movement of the green concrete, this concrete might have also lifted the reinforcement cage. The length of reinforcement cage above the level up to which the necking took place is only one sixth of the total reinforcement length. The resistance offered between the remaining length of the reinforcement and the concrete around this length is much more than the lifting force that can be offered by the concrete moved upward while necking took place.

Physical verification of the top level of the reinforcements for each pile were made. Interestingly the observations revealed that the top levels of the reinforcements for almost all the

13

piles in a group remain same. There were groups of 6 piles to 18 piles. Very rarely there was relative difference within a group. There was a drawback of not recording the actual ground levels at each pile location that complicated this review process. An average ground level of (-)4.50m was mentioned in the records of all the piles which could not have been the case in such a large piling area. The length required for reinforcement cage in each pile was estimated based on this average ground level. This caused small differences between the levels of reinforcement in piles of different groups.

About 1% piles found suffering from relative displacement of reinforcement by 50mm to 350mm, in which most of the cases were less than the lengths as per record. A slip at lap joints when the laps were towards the bottom of the cage was suspected to be the reason for this phenomenon and this was verified from some records. Only 5 piles recorded lifting of reinforcement. One pile recorded a cage lift of 560mm.

Compression capacity of piles with respect to bearing: Direct measurement of static vertical compression capacity by means of maintained load tests was resorted to. Eleven piles out of about 1000 piles were subjected to load tests apart from two initial load tests carried out from a higher level. Eight piles recorded settlements within the permissible limits and two piles recorded slightly excessive settlements. One pile failed at very small load and this pile had the history of cage lifting by 560mm. However, the range of settlements recorded in some of the load tests were towards the maximum limits and there were deficiencies when compared with the results from initial load tests. The piles were driven to a very hard set of 3 to 5mm for ten blows of 4 tonnes hammer falling from 1.20m height at the time of construction and such large settlements were not expected under static load.

The effect of excavation for basement: The relief of roughly 6.0m thick soil for the construction of basement will cause a ‘heave’ in the soil below the excavation level since these layers are relatively stiff and dense. This heave might cause lifting of piles through the friction between the heaving soil and the pile. This might also cause a relative movement between the pile tip (flat shoe in this case) and the soil at this level. However, such relative displacements shall be fairly uniform and the load test results should also have resulted a uniform pattern. Large differences in the behavior of piles pointed to different reason/s.

The excess pore pressure at the time of pile driving: The completely weathered rock present at the pile bearing level has reasonable amount of clay suggesting low permeability. Large amount of excess pore pressure could develop during such hard driving of pile. Anticipating that the entire excess pressure could not have dissipated before completing the concreting and lifting of the casing pipe, the piling instructions included a direction not to lift the casing pipe before placing minimum length of 11m concrete column within the casing. This concrete column provided a pressure of 22 t/m2 against excess pore pressure of about 8 t/m2 plus the static water pressure of 14.5m at the pile tip. It was reported that some of the piling rig did not have the capacity to lift the casing pipe carrying a large column of concrete. There was also significant resistance from the soil around the casing. Therefore, the piling agency resorted to an initial lifting of 500mm after filling only about three metre concrete in the casing pipe. The concrete exposed to the existing excess pore pressure should have suffered loss of cement particles resulting a weak concrete at the pile tip. Then the real worry was to identify such piles since the piling records did not mention the initial lifting of casing pipes.

14

Investigation by pile integrity test, PDA: A large number of piles were subjected to low energy PDA to confirm the pile diameter, length of pile and the quality of concrete. These factors were determined with some accuracy, but the results were not conclusive on two aspects. One is the quality of concrete at pile tip and the other was inconsistencies in shear wave velocities. In almost all the cases intact pile tips were recorded, which was proved wrong by other means later.

Investigation by high energy hammer tests: It was decided to conduct some high energy tests by ‘driving’ the piles using a 4 tonne hammer. A suitable frame with leader was devised and initially around 100 piles were driven. While about 70% piles did not penetrate more than 4mm under three or four hammer blows, 30% piles penetrated more than 10mm in one or two blows. This was attributed to poor tip conditions caused by premature lifting of the casing pipe since the piles were initially driven to very hard set. Most of these piles, however, offered resistance after this ‘initial’ displacement. These 30% piles would have settled more under the working load while the pile are expected to settle only 6 to 8mm under the working load.

These observations were not in comparison with the PDA test results, particularly in the case of quality of pile tip. Therefore, PDA tests were discontinued. However, driving the piles using high energy drop hammer was resorted for identifying the piles with bad tip conditions and also for driving it further to achieve adequate seating of piles with relatively large displacement during initial blows. Except for 5% piles, the piles could be driven to a hard set with maximum movement of 10 to 12mm.

Piles with more set: The 5% piles with more set had to be underrated from its safe capacity and some piles were to be rejected altogether because of continued displacement. About 1% of the total piles were rejected and new piles were driven to compensate the deficiency.

6.1.7 Conclusions

Concreting the piles up to about 3.0m above the cut-off level and filling the remaining length up to working level using sand did not work against necking of concrete within soft clay potion. This was identified to be attributed to arching of sand within the casing pipe completely relieving the intended load over the green concrete. However, in this particular case, the cut-off level was very close or below the portion affected by necking. Foundation consultants had to learn this lesson.

The piling agency failed to comply with the requirement of minimum length of concrete column within the casing for countering the excess pore pressure developed during driving. This resulted bad concrete at the tip of a significant number of piles

The piling agency also failed to take ground level measurements at each pile location and instead resorted to approximate measurements leading to misjudgment of reinforcement lengths. This resulted difficulty in identifying piles with real problem of reinforcement cage lifting.

Placing un-welded laps on main bars towards the tip of the reinforcement cage had the risk of slipping at the lap by self weight of remaining cage above the lap.

These lapses resulted huge investment on rectification measures like re-driving of all the piles and, underrating of few piles and rejecting some piles apart from heavy loss of time.

6.2 The Case of Settling Bungalows

6.2.1 The Problem

15

A vast residential project site was developed from a lowland by filling good quality murrum (residual gravely soil) and then by preloading, supplemented with band drains for increased rate of consolidation settlement. The effect of preloading was confirmed by time-settlement measurements using plate settlement markers. The observed settlements were 780mm to 850mm, close to 90% of the total estimated settlement of 900 to 950mm. The time settlement curves also suggested 90 % consolidation under the prescribed pre-loading period of 60 days. Construction of two bungalows was taken up immediately and a systematic measurement of settlement under every stage of construction was carried out.

Two bungalows were then constructed and a detailed monitoring of the settlement was made. One bungalow settled by about 145mm and the other one settled by 80mm, whereas the expected settlement of the structure after ground improvement was less than 40mm. Fortunately, the settlements were uniform because continuous RCC strip footing foundation was provided for these buildings, whereas the initial design was based on RR masonry strip footing. Even though the structure did not experience any distress because of uniform settlement, the excessive settlement far beyond the expected settlement, was considered as a ‘failure’ in the design of ground improvement programme.

6.2.2 Distress and Owner’s Worry

Even though the settlements were uniform, there were apprehensions about settlements larger than acceptable limits. The differential settlements at every stage were negligible because of fairly uniform construction schedule adopted for these two buildings. The foundation consultant and the ground improvement agency were called upon to explain the large settlements irrespective of the commitment that the settlement of finished structure not to exceed 40mm.

6.2.3 The Soil Profile

The detailed investigation prior to the foundation design revealed presence of 1.20m thick residual soil fill followed by soft to very soft marine clay of 4.50m to 7.00m thick below which relatively stiff residual clay existed. Weathered rock stratum followed. Compressibility of plastic clay was high and long term settlement due to consolidation of these layers under the weight of existing fill and another 0.80m fill proposed was estimated.

Shallow foundations resting in the fill and deep foundation resting in weathered rock suffered doubts of long term performance. The distress in infrastructure like roads, sewage lines, water pipe lines, etc. were anticipated and ground improvement using pre-loading was considered more appropriate.

6.2.4 Original Design

The amount of preload was decided based on a pre-determined finished ground level of 0.80m above the existing ground level. There already existed an original fill of 1.20m thick that has not undergone adequate compaction. A fill of about 3.80m above the existing level imposing a load of about 7.0 t/m2 was the recommended pre-load as the average load intensity from the proposed construction including the weight of the 0.80m fill required to reach the finished ground level was close to 6.2 t/m2 (roughly 90% of the pre-load). Vertical band rains at an interval of 0.90m were introduced for the depth of soft clay for accelerating the consolidation process. It was expected that about 2.20m thick preload fill (equal to a load of 4.2 t/mm2) could be shifted to other locations

16

leaving the FGL as required. Figure 15 illustrates the preload fill and the expected settlement under the preload. Figure 16 illustrates the design principle of pre-load for ground improvement.

6.2.5 Changes in the Design

The owner and architect meanwhile decided to raise the ground level by another 0.80m for better appearance and also to avoid any possibility of flooding of the premises. But this was informed to the consultant and the ground improvement agency only after the removal of the pre-load and the strat of foundation excavation.. The consultant advised further pre-loading to compensate the deficiency arose because of revision in the finished ground level. However, the owner decided to compact the soil at founding level and carried out plate load test on the compacted soil. Satisfied with the small settlements measured in the plate load test, the owner proceeded without further pre-loading.

The plate load test using a 450mm x 450mm plate naturally showed less settlement since the test was done on the compacted fill. Based on the results the construction proceeded, but after agreeing to provide continuous strip raft as foundation (Figure 17).

LOAD FROM PRE-LOAD + FILL UP TO FINISHED GL

NET SETTLEMENT

TOTAL SETTLEMENT

PROCESS OF CONSOLIDATION

(EXPELLING WATER FROM VOIDS)

PRE-CONSOLIDATION-(LOADING HISTORY)

PRESENT LOADING

REMOVAL OF PRE-LOAD

EQUAL TO LOAD FROM STRUCTURE

SETT

LEM

ENT

PLAS

TIC

YIE

LD

Figure 16: Design principle of the preload

Original ground level

Area for building

Other areas

Finished ground level

Preload for removal

Soft clay compressed by 800 to 1000mm under pre-load

Pre-consolidation improved by 2.5 t/m2

Soft clay compressed by 400 to 500mm under pre-load

Figure 15: Pre-load fill and the settlement

Figure 17: RCC strip raft in the form of inverted ‘T’ adopted for the pilot bungalows

17

The construction of both the buildings was fairly uniform and took about 6 months. Regular measurements of settlement at different corners of the buildings were made. The band drains functioned very effectively during the construction and the settlements were very rapid. The settlements measured after completing the construction were 145mm and 80mm for the two buildings. The Figure 18 illustrates the settlement of one villa. Settlement stabilised within fifteen days from the completion of the construction. Uniform construction schedule helped in resulting a uniform settlement without any significant differential settlement.

6.2.7 Analysis

Very careful study of the load settlement curves will show that the rate of settlement significantly increased after 20 to 30mm settlement suggesting the major settlement is resulting form virgin compression and not from re-compression expected from pre-loaded soil. This means that there is a deficiency in the pre-load intensity. In this case, the pre-load fill that could be removed after the preloading period was limited to 1.40m because of the upward revision of finished ground level. Thus the effective preloading suffered a deficiency of about 1.5 t/m2 equal to 0.80m thick fill (Figure19) that caused this additional virgin compression.

However, it is very clear from the settlement observation in the front and rear of the building, the settlements were very uniform. The main reason for such uniform settlement was very uniform vertical progress of the construction possible for a load bearing wall construction. The initial design was to have independent footings with RCC frame and infill walls that would have resulted very non-uniform construction pattern. Having the vertical drains in position, the settlements could have been very fast (also as seen from the time settlement curve in Figure 18) resulting large differential settlements. Another reason for such uniform settlement is that the inverted ‘T’ RCC strip footing with much higher stiffness contributed to a more uniform load distribution.

A detailed settlement analysis revealed that the probable settlement due to this extra loading is 80 to 105mm. More settlement for one villa could have been the

EXTRA LOAD FROM THE REVISED FGL

SETTLEMENT EXPECTED

ACTUAL SETTLEMENT

SETTLEMENT VILLA # 42

Roof tiling completed

First floor roof completed

0

20

40

60

80

100

120

140

160

180

200

0 20 40 60 80 100 120 140 160 180 200

NO OF DAYS

CU

M S

ETTL

EMEN

T M

MREARFRONT

Figure 18: Settlement observation of Villa # 42

18

result of compression of initial portion of the preload fill that remained below the founding level.

6.2.8 Further Constructions

However, such large settlements could not be allowed because of the fact that the RCC strip raft foundation was relatively expensive and a uniform construction pattern could not be enforced in the case of RCC frame construction.

Even though the owner decided not to listen to the arguments, adopted required additional pre-loading for the remaining sections of the project. The settlements observed during preloading increased to 900 to 1000mm and the completed bungalows recorded settlements less than 30mm.

6.2.9 Conclusions

Ground improvement programmes like pre-loading with settlement accelerators require good planning with respect to site development. All the possible changes in the development programme shall be thoroughly investigated and the improvement shall take care of the worst case. Co-ordination between the foundation consultant, the architect, the structural engineer and the owner is a must in such programmes.

A thorough and systematic monitoring of improvement in terms of settlements versus the load and time is needed.

Adoption of expensive RCC strip raft foundation for the area with under-improvement helped in preventing large distress in the structure. However, adoption of similar measures for all the villas could have affected the economy of the project.

6.3 The Case of Rotating Abutment Pier

6.3.1 Problem

Two abutment piers and three intermediate piers for a rail over bridge was constructed on shallow footings placed at about 2.60m below the natural ground level. The construction was planned without hindering the rail traffic and hence launching of main girders through rails supported on approach embankments was in the offing. The general layout of the piers and the embankment is shown in Figure 20.

Embankment fill

Embankment fill

Piers

Failed pier

Figure 20 Layout of the abutment piers and intermediate piers

19

The embankment construction progressed from both the sides and when the embankment near one of the abutment piers reached almost maximum height, the embankment fill failed. The top of the abutment pier moved towards the embankment fill while the embankment cross section close to the pier slipped by more than 1.50m (Figure 21). The embankment fill did not reach the abutment rear face when failure occurred.

6.3.2 The Distress

It was a clear case of base failure of almost vertical embankment section facing the abutment pier, without a support by the pier. A wedge shaped gap was provided between the abutment pier and the embankment for the provision of conventional filter layer. Rotation of the abutment pier about its base and towards the embankment clearly suggested the involvement of soil beneath the pier foundation in the embankment base failure. The major distress was failure of the abutment foundation because of its rotation.

6.3.3 The Investigation

A possible slip circle was constructed based on the shape and position of the slip line. The slip circle passed very close to the abutment pier foundation and the diameter of the slip circle was roughly estimated as 36 metres. The design of pier foundation was done on the basis of net safe bearing capacity equal to 18 t/m2 arrived at on the assumption that stiff sandy clay with good shear strength present 2.60m below the ground level is continuing towards depth.

Further investigation through four exploratory boreholes and five dynamic cone penetration tests revealed presence of soft clay and soft sandy clay with undrained shear strength in the range of 0.20kg/cm2 to 0.28 kg/cm2 between 3.0m and 8.0m below the ground level. The thickness of relatively good bearing stratum between the founding level and the weak layer was less than 0.50m. The soil profile and the failure imagined are illustrated in Figure 20. The detailed geotechnical investigation data from one borehole is presented in Figure 21.

15 30 45 60 N**

5.0

6.0

1.0

2.0

Soil

Prof

ile

3.0

4.0

Field Description

Test

Dep

thm

SPT / VST

Blow Counts

Dep

th b

elow

G

L (m

etre

s)

1

1.50

0.75 2 1 0 0

0

3.00 0

2.25

0

4.50 0

3.75 Su=0.24kg/cm2

Sunk @63.5kg

0

6.25 0

5.50

Su=0.38kg/cm2

Sunk @63.5kg

Su=0.44kg/cm2

Sunk @63.5kg

Su=0.31kg/cm2

0.5

1.5

5.1

6.2

Greyish dry sandy clay

Brownish grey soft clay with sand pockets

Light grey soft clay with fine sand

Greyish soft sandy clay/clayey sand

Dark grey soft clay

Fill sandy clay, cu = 0.35 kg/cm2 Fill clayey sand, φ = 28º Sandy clay, cu = 0.70 kg/cm2 Sandy clay, cu = 0.30kg/cm2

sandy clay, cu = 0.28 kg/cm2

Soft clay, cu = 0.21 kg/cm2

Sandy clay, cu = 0.65 kg/cm2

Figure 21 Slip of the embankment

20

6.3.4 The Analysis

Standard slip circle analysis resulted a minimum factor of safety of 1.29 based on the soil data revealed from the investigation. Relatively high shear strength of 0.70kg/cm2 for the sandy clay available up to 3.0m from original ground level is providing considerable resistance to failure. This called for a re-look into the back analysis. It was then observed that the possible slip circle passed through the interface between the backfill of pier foundation excavation and the original stiff sandy clay. The large excavation made for the construction of the abutment pier left a weak zone having very poor backfill in place of the stiff clay. A slip circle analysis was then carried out assuming only about one sixth of the shear strength of the stiff sandy clay along the foundation excavation line, and a factor of safety 0.83 was obtained.

The soil layers below the abutment foundation have very high compressibility and any load over the abutment pier beyond the self weight of the pier should have resulted excessive settlement and a bearing capacity failure.

Figure 22 Soil profile 15m away form the failed pier foundation

21

6.3.5 Conclusions

The foundation provided for the abutment pier was inadequate in view of soft clay layers immediately below the founding level. Very large load from 8.0m high embankment triggered the failure. Major load over the abutment pier from the main girders, road formation and the road traffic was due. Even if the failure of abutment was not happened during the embankment construction, a major failure when the abutment was loaded fully was sure to happen.

Changes in the shear strength parameters because of possible construction activities like foundation excavation and backfilling of such excavations are very relevant in similar cases.

This case illustrates clear failure from the part of the designer and the owner who went ahead with the design and construction without ascertaining the soil conditions below the embankment and the piers.

REFERENCES

Allen Marr, W. Ramos J.A, Lambe T.W. (1982), ‘Criteria for Settlement of Tanks’, Journal of the Geotechnical Engineering Division, Proc. ASCE, Vol 108, No: GT8, Aug 1982, 1017-1039

Anirudhan I.V. (1997), ‘Driven cast-in-situ piles – Execution and performance’, Proc. Indian Geotechnical Conference, IGC 1997, Vadodara, pp 233-236

Anirudhan I.V. (2005), ‘Types of distress in Geotechnical Structures’, Proc. Indian Geotechnical Conference IGC 2005, Ahmadabad, pp 165-168

Boone S.J. (2001), ‘Assessing Construction and Settlement-induced Building Damage: A Return to Fundamental Principles’, Proceedings Underground Constructions, Institution of Mining and Metallurgy, London, 559-570

Boscardin M.D. & Cording E.J. (1989), ‘Building Response to Excavation Induced Settlement’, Journal of Geotechnical Engineering, ASCE, 115(1), pp1-21

Burland J.B. (1997), Assessment of Risk of Damage to Buildings due to Tunnelling and Excavation’, Earthquake Geotechnical Engineering, Ishihara (ed), Balkema, Rotterdam, 1189-1201

Burland J.B. Wroth C.P. (1974), ‘Settlement of Buildings and Associated Damages’, State of Art Report, Proc. Conference of Settlement of Structures, Cambridge pp 611-654

Burland J.B., Broms B.B, DE Mello V.F.B. (1977), ‘Behaviour of Foundations and Structures’, State of Art Report, Proc. 9th ICSMFE, Tokyo, Vol 2 495-546

Madhav M.R.(2003), ‘Modelling Methods in Geotechnical Forensic engineering’, Proc. of a Workshop by Committee on Professional Practice of Indian Geotechnical Society, Chennai, Feb28 –March 1, 2003, pp 75-81

Mair R.J. (2001), ‘Research on Tunnelling Induced Ground Movements and Their Effects on Buildings – Lessons from the Jubilee Line Extension’, Proc of the Intnl Conference on Response of Buildings to Excavation Induced Ground Movements, Imperial College, London, UK, July 17-18, pp 3-26

Poulos H.G. (2003), ‘Á Framework for Forensic Foundation Engineering’, Proc. of Workshop on Forensic Geotechnical Engineering, Committee on Professional Practice of IGS, Feb 28-Mar 1, 2003, pp 7-13

22

Radhakrishnan R and Anirudhan I.V. (2003), ‘Ground improvement with pre-fabricated V drains and pre-load – A case study, Proc. Symposium on Advances in Geotechnical Engineering, SAGE 2003, IIT Kanpur, pp 426-430

Ramanatha Ayyar T.S. & Jaya V (2003), ‘Geotechnical Aspects of Mass Collapse of Shallow Wells in Kerala During 2001’, Proc. of workshop on Forensic Geotechnical Engineering, Committee on Professional Practice of IGS, Feb 28-Mar 1, 2003, pp 33-36

Ramaswamy, S.V. and Narasimhan, S.L. (1978), “Behaviour of Buildings on Expansive Soils – Some Case Histories”, Jnl. Institution of Engineers (India), Vol. 58, Pt. CI 4, pp 141 – 46.

Chapter 4

FORENSIC GEOTECHNICAL ENGINEERING INVESTIGATIONS:

DATA COLLECTION

Peter Day Jones & Wagener Consulting Engineers, Rivonia, South Africa, 2128

day@jaws.co.za ABSTRACT: This paper describes the initial stage of a forensic geotechnical investigation, namely the gathering of data on the site, the works and the failure for further detailed analysis. It has been written as part of the effort by the ISSMGE Technical Committee TC40 to prepare a handbook on forensic geotechnical engineering. The paper provides guidance to the investigator on the objectives of the investigation, the nature of the data required, sources of information to be considered and the recording and storage of data.

1. BACKGROUND

Forensic geotechnical engineering deals with the investigation of failures of geotechnical origin, not only from a technical viewpoint but also with the possibility of legal proceedings in mind (Rao, 2005). Forensic investigations differ from conventional geotechnical investigations in that they are retrospective. They seek to explain what has happened rather than to predict future performance. A further distinguishing factor is that, following a failure, there is an urgency to clean up the site and rebuild or repair the works. This limits the time available for investigation and makes it essential that all relevant data is recorded before the evidence is removed. The ideal outcome of the data collection stage of a forensic investigation would be to have a body of information that is (a) as complete as reasonable possible, (b) accepted by all parties as an accurate record of the facts and events and (c) is stored in an accessible and readily understood way. This paper describes procedures aimed at achieving this outcome.

2. SCOPE OF INVESTIGATION Although some geotechnical failures such as landslides occur in the absence of any human intervention, most geotechnical failures involve both the ground (soil, rock and groundwater) and the works (some man made structure or intervention). The works may be a structure that imposes loads on the ground, the alteration of the surface geometry (cuts or fills), alteration of drainage patterns or the creation of underground openings. Thus, the forensic geotechnical investigation must include a study of the event or failure which gave rise to the investigation, the site on which the failure occurred and the nature of the works. Only after these three aspects have been investigated and recorded can post-failure diagnostic testing and back-analyses commence.

2.1 The Failure In any forensic investigation, it is essential that the circumstances and events surrounding the failure are investigated and recorded as soon as possible, before any evidence is removed. Details of this stage of the investigation will vary from case to case. Nevertheless, there are common aspects that apply to all failure investigations. These include the conditions that prevailed immediately prior to the failure, the sequence of events, and the condition of the works and surrounding areas following the failure. 2.1.1 Circumstances prior to failure Before one can establish the cause of the failure, it is essential to investigate and record the condition of the works immediately prior to the failure. Typical factors to be recorded include:

the stage of completion of the works at the time (see sub-section 2.3.3 below), the occurrence of accidental actions (impact, explosion, earthquake, flooding or water leakage,

etc) or abnormal loading, abnormal meteorological conditions (wind, snow, rainfall, temperature, etc), the results of any monitoring (pore pressures, deformations, settlements, anchor loads, etc),

and any early warnings of incipient failure (cracking of ground or structure, falls of ground,

changes in anchor loads, etc). 2.1.2 Sequence of events Obtaining an accurate record of the sequence of events that lead to the failure will assist greatly in determining the failure mechanism and, in many cases identifying the trigger. The information to be recorded will vary from site to site. However, an attempt should be made to obtain as much information as possible from the time when the first signs of distress were noted. An attempt should be made to draw up a time line from which the sequence of events and the speed of progression can be ascertained. 2.1.3 Resulting distress The two sub-sections above deal with the “before” and “during” situations. This sub-section deals with the “after” or post-failure state of the works and surrounding areas. An accurate description of the distress caused by the failure may be pivotal in determining the value of any claim for compensation or damages which may follow. Unlike the “before” and “during” situations which must be investigated by reference to historical records, the “after” situation can be directly observed, photographed and recorded. Note that the recording of the condition of the works after the occurrence as described in this paper is distinct from the detailed diagnostic tests that may be required to provide parameters for back-analysis of the failure. Typical information to be recorded may include:

the extent and severity of distress, the magnitude of deformations and trajectory of movement, any signs of where rupture may have occurred (slickensided shear zones, yielded construction elements, etc),

indications of abnormal surface or subsurface water conditions (seepage, high water marks, etc), condition of any exposed rupture surfaces whether on the structure or in the ground, any deviations from expected ground conditions (paleo channels, intrusives, faults,

adverse jointing, seepage, etc), and damage or changes in surrounding areas (physical damage, settlement, lateral movement,

drop in groundwater levels, rupture of services, etc). 2.1.4 Sources of Information When determining condition of the works prior to the failure, reference should be made to any available reports (geotechnical, structural, etc), to construction records, progress payment certificates, photographs, as-built drawings and any information that can be provided by site staff.Depending on the circumstances, it may be necessary to obtain additional information from outside the site such as water meter readings, flood levels, metrological data, etc. Recent aerial photos or satellite images may also be of assistance. Determining the sequence of events normally relies on eye-witness accounts and photographs taken immediately prior to, during and immediately after the event. Care should be taken when interpreting eye-witness accounts as vested interests may be involved. For this reason, it is preferable to obtain information from as many site personnel as possible and preferably not only from personnel employed by either the engineer, the owner or the contractor. Any discrepancies in the information provided by witnesses should be revisited and clarified at the time. Written records should be kept of eye-witness interviews and these should preferably be signed by both the investigator and the witness. Where appropriate, verbatim transcripts or recordings of eye-witness evidence should be kept. Information of the condition of the works after the occurrence is generally obtained by direct observation. This will typically include photography, sketches, written descriptions, post-failure survey drawings, etc. It is preferable that the post-failure information is jointly recorded by the parties involved (owner, engineer, contractor, insurer etc) to minimise disputes at a later stage. 2.1.5 Potential failure mechanisms The scientific method relies on a process of postulation and verification. Many scientific endeavours have floundered as a result of failing to consider alternative postulates and gathering only information which supports a particular point of view. In any scientific investigation, it is as important to record both supporting data and data which is inconsistent with various postulates. The purpose of this first phase of the investigation is to gather data. Back analyses and the identification of the most likely failure mechanism(s) will follow at a later stage. In fact, forming a fixed opinion on the cause or mechanism of failure may result in certain essential information relating to possible alternative causes being overlooked. Nevertheless, it is recommended that investigators should seek to identify all the potential triggers, sequences of events and failure mechanisms. Simple logic is likely to eliminate many of these from the start. An attempt should then be made to obtain the data which will enable the likelihood of any plausible failure mechanism to be assessed at a later stage.

2.2 The Site In order to carry out a competent analysis into the causes of the failure, information is required on the site on which the works were undertaken. Much of this information should already be available in the form of existing reports and other documents. 2.2.2 Essential Data The essential data required will vary from site to site. In most instances, it will include:

the location and extent of the site, surface topography and alterations thereto,

surrounding services and development site description including vegetation, drainage, climate, previous land use, existing

development, etc, geological setting, site geology, regional geological structures (faults, folding, etc) and

seismicity, site stratigraphy including the identification of typical soil profiles for various areas of the site, detailed soil and rock profiles from boreholes, pits or other exposures at particular locations on the

site, information on groundwater including depth of water table(s), seasonal fluctuations, gradients,

flow characteristics, etc, and results of field and laboratory tests.

2.2.2 Sources of information The main source of geotechnical information on the site should be the geotechnical and geological reports prepared for the project. The absence of such reports may, in itself, be a contributor to the failure. The next most likely source of geotechnical information on the site is the construction records. Much valuable information may be gleaned from site instructions, recorded founding depths, tunnel face maps and other similar information.

On some projects, borehole core or samples from the original investigation may still be available. This creates the possibility of re-inspecting the core or re-testing of samples. 2.3 The Works In considering the works, there are three main aspects which require consideration. These are (a) the works as designed, (b) the works as constructed and (c) if appropriate, the state of completion of the works at the time of failure. 2.3.1 Works as Designed Information on the works as designed is generally contained in the construction drawings and project specifications. These documents specify the work to be carried out by the contractor and provide a basis from which any deviations may be assessed. Of equal importance are the calculations on which the design of the works was based. A review of the design calculations will enable the investigator to ascertain whether the conditions on site are consistent with the those assumed by the designer during the design process. 2.3.2 Works as Constructed The works as constructed may differ from the works as described on the construction drawings and project specifications for a number of reasons such as on-site design modifications, substandard materials or workmanship, geometric deviations (both construction tolerance and setting out errors) and concessions granted to remedy non-compliance. Identification of differences between the “as designed” and “as constructed” works will generally require physical measurement, survey, inspection and testing on site. Supplementary information can be obtained from site documents such as non-conformance reports, design modification reports, site instructions, minutes of site meetings, etc. 2.3.3 State of completion

A significant proportion of all geotechnical failures occur during construction. Where this is the case, it is essential to determine:

the state of completion of the works at the time, applied loading (both self weight and imposed loads),

changes in loading or construction activities

immediately prior to failure and strength of materials that show time dependency at the time of failure (eg clays

undergoing consolidation, freshly cast concrete or grout, etc). 3 RECORDING OF DATA

3.1 Attention to Detail The forensic investigator should strive to ensure that all information obtained during the course of the investigation is complete in every respect. Not only does this require that the scope of the investigation is adequate, that all plausible failure mechanisms are investigated and that all relevant documentation is considered, it also requires meticulous attention to detail. In this regard, it is recommended that written records be kept of all discussions and inspections and that these records be dated and signed. All photographs should be uniquely referenced and the date, time, location and orientation of the photographs should be recorded. Any samples taken should be photographed in situ (prior to sampling), provided with a unique sample number and records kept of the tests undertaken and the results obtained.

3.2 Agreement between Parties Forensic geotechnical investigations may either be carried out by an independent investigator or by different investigators employed by the various parties (employer, contractor, engineer, insurer, etc). Wherever possible, all parties should be afforded the opportunity to witness critical stages of the investigation such as exhuming of foundations, taking of samples, removing collapsed structures, etc and observations made during such crucial stages of the investigation should be shared with all parties. This is to avoid the situation where crucial evidence is ruled to be inadmissible during the legal process as the accuracy of the observations, location of the samples or photographs and other such details can only be vouched for by one party. In investigations of this nature, it is preferable that as much agreement as possible is obtained during the investigation stage as it is likely that much of the evidence will have been removed when the time arrives for resolving disputes or legal argument,.

3.3 Reporting and data storage It goes without saying that all data obtained during the course of the investigation must be adequately documented and stored in a well referenced and easily retrievable format. Even the simplest of failure investigations can drag on for a number of years. Given the mobility in the job market, it is likely that the data will be analyzed by a different team of engineers to those responsible for its collection.

In the case of small investigations, it is preferable that a report be compiled which details the extent and findings of the investigation and appends any relevant supporting documents. In the case of larger investigations, the report is more likely to be a summarised version of

events and an index of the various supporting documents. On major projects, the number of documents may run into the hundreds and possibly thousands. Under these circumstances, electronic storage of documents should be considered due to the ease of reproduction and portability of electronic records. 4 CONCLUDING REMARKS The gathering of data for a forensic geotechnical investigation should be conducted with an open mind. Collection of selective data intended to support a particular hypothesis may be counterproductive in that the data so collected will be inadequate to test the veracity of alternative theories. It is as important to collect supporting evidence as it is to note evidence that is inconsistent with postulated failure mechanisms. The investigator should be mindful of the fact that the data gathered may be subject to scrutiny in subsequent legal proceedings. Obvious bias in the collection and reporting of data will discredit the findings of the investigation. Where the investigator has access to all parties (e.g. in the case of a joint appointment), consultations should be held with all parties involved. More often than not, follow-up consultations will be required as information obtained from one party may be queried or refuted by another. Wherever possible, an attempt should be made to obtain agreement between the parties on important issues at the time as this reduces the amount evidence to be lead in any future legal proceedings. 4 REFERENCES Rao, V.V.S. (2005) –TC40 Terms of Reference. Report submitted to ISSMGE, December 2005.

Chapter 5

COMPILATION OF DATA

Dr.V.V.S.Rao, Nagadi Consultants Pvt. Ltd., New Delhi, INDIA

dr_rao@nagadi.co.in INTRODUCTION: Although some geotechnical failures such as landslides occur in the absence of any human

intervention, most geotechnical problems involve both the ground (soil, rock, and ground water)

and the works ( man made structure or intervention). The works may be a structure that imposes

loads on the ground, the alteration of the surface geometry (cuts or fills), alteration of drainage

patterns or the creation of underground openings. For a forensic analysis of such problems the

first step is to compile all available data about the project. This compilation begins with the

detailed description of the distress or malfunction of the structure which has been attributed to

geotechnical causes. The distress or malfunction can be a complete failure, or excessive

deformations, or unacceptable responses to vibrations, or excessive seepage in case of water

retaining structures.

TYPES OF DISTRESS:

a. Complete failure

In the case of a complete failure the stress conditions in the supporting or participating soil have

reached their ultimate resistance level and the structure has collapsed. Typical examples are

foundation failures, collapse of retaining structures, land slides, and slope failures in cuts and

embankments. Generally, water seeping through the slopes is the culprit. In case of natural

slopes like in hills, the water seeping through the porous seams and rock joints play havoc.

Piping action due to poor compaction in some pockets of an earthen dam body can even lead to

its sudden collapse.

b. Excessive deformations

A soil-structure system is always designed to tolerate certain magnitude of deformations. These

deformations include both total deformations occurring in the system as a whole and differential

deformations between elements of the structural system. We are mainly concerned with the

deformations caused by the supporting soil medium. These deformations are combinations of

immediate ones which occur within a short time after the forces are transferred to the soil and the

ones which occur slowly over a period of time after the commissioning of the project. The soil

strata which has cohesionless soil within the influence zone undergoes mainly immediate

settlements while in the cohesive strata the deformations are time dependent. In both cases the

major causation factor is the water. This water may be from the ground water table fluctuation

or/and ingression from external sources. Normally, immediate settlements which can be

estimated and provided for in the designs are acceptable. Differential settlements may lead to a

condition when the structure can not be utilized fully for the designed purpose.

c. Excessive vibrations

In case of industrial as well as laboratory structures having testing facilities with sensitive

instrumentation, vibrations of individual foundations due to forced excitation on them and also

due to vibrations transmitted through the soil may not be acceptable. Similarly, in case of

structures built in seismic sensitive zones, the effect of earthquakes can lead to damage to

foundations including excessive tilt/collapse of the structure due to liquefaction. In all these

cases, the response of the soil-structure should be within tolerable limits.

d. Seepage problems.

In case of water retaining structures like dams and water/sewage treatment plants at ground level,

the water seeping through the soil strata may cause erosion due to piping, flow of soil, or

excessive seepage in spite of foundation treatments. The seepage problems are particularly

common during rains.

CAUSES OF DISTRESS

During the data collection stage it will not be possible to identify the primary cause of distress.

However, as the data collected should be comprehensive and sufficient to analyze the problem

from all angles. A study of several published case histories suggest the primary causes can be

grouped under five headings:

a. Underestimation of forces

A design engineer normally assumes certain magnitudes of different types of forces that may act

on the soil-structure system, depending upon relevant standard codes and also upon his

experience. As “Engineering Judgment” is involved, there can be difference of opinions between

designers at this stage. It is advisable to highlight the forces considered in the design.

b. Meteorological data

Factors like high flood level, maximum intensity of rains/snow, direction and velocity of winds,

and variations in daily as well as seasonal temperatures, etc. are used while selection of design

forces. Proper selection of their magnitudes is important for design.

C .Inadequacy of geotechnical investigations

Trying to economize on investigations leads to lack of sufficient data for scientific evaluation of

subsoil properties.

d. Design soil parameters

This is one of the critical stages in design. Improper values may bring disaster.

e. Improper design criteria

At this stage important decisions regarding tolerable deformations and allowable stresses on the

materials to be used are made. As described above, the designers’ engineering judgment plays a

very important role. The “efficiency” of the structure depends on this stage.

f. Inappropriate/inadequate design

The design theories to be adopted depend upon the type of sub-soil strata and substructures.

Adopting design methods irrelevant to the type and nature of the subsoil will lead to inadequacy

of the system response. Soil-structure interaction is another design aspect which needs to be

considered in depth. Short-cut methods like empirical designs should be strictly avoided.

g. Improper construction methodology

Construction methods and equipment needed for the same depend directly upon the subsoil

conditions. Contractors having suitably skilled labor should have been selected for the

construction.

DATA TO BE COLLECTED

A. Predesign stage

This is mainly desk work. The history of the project, from the concept stage up to the selection of

the site form the first part of the study. The factors considered while selecting the site as also the

persons involved in the process should be identified. In the second part, the data collected for the

design including their sources should be compiled. The data to be compiled include:

a. Detailed topographical survey,

b. Metrological including hydrological data,

c. Results of preliminary soil survey data from trial pits , shallow boreholes, etc.,

d. Sources of energy and water, and,

e. Details about equipment and machinery to be installed.

B. Design stage

As this is the crucial stage, all data, even though on the outset looks superfluous, should be

collected. The important ones are:

a. Data regarding the structural aspects including the details of the machinery, if any,

b. Detailed report of subsoil investigations:

1. Locations of points of investigation w.r.t. structures

2. Type of exploration like, boreholes, penetration tests, load tests, etc,

3. Depth to which explorations are done,

4. Results of field and laboratory tests: their adequacy and accuracy w.r.t. the

importance of the structures,

5. Interpretation and analysis of all results, theories used for analysis,

6. Selection of strength as well as consolidation parameters of soil strata

c. Designs including assumptions regarding loads, tolerances in deformations, and strengths

of all building materials used, along with calculations.

d. Drawings, whether all information needed for execution, including the precautions to be

taken, etc. are included,

e. Construction sequence to be followed.

f. Proof checking of all designs

g. Approval of the project authority

C. Construction stage.

a. Selection of suitably qualified and experienced executing agency

b. Contract agreement

c. Project organization, field and office. The responsibilities and liabilities of all involved

persons should be clearly identified.

d. Monitoring the construction, verification of soil conditions

e. Works as designed vis-à-vis as constructed:

Information on the works as designed is generally contained in the construction drawings

and project specifications. These documents specify the work to be carried out by the

contractor and provide a basis from which any deviations may be assessed. The works as

constructed may differ from the works as described on the construction drawings and

project specifications for a number of reasons such as on-site design modifications,

substandard materials or workmanship, geometric deviations (both construction tolerance

and setting out errors) and concessions granted to remedy non- compliance.

f. Quality control during construction:

Details regarding the personnel who did the quality control tests and the methods adopted

need to be documented. Detailed documentation should have been maintained.

g. Metrological conditions during the constructionperiod

h. Details regarding the instrumentation used for maintaining the accuracy of construction

and also for monitoring the behavior of the structure along with the readings and their

interpretation should have been maintained.

i. Construction sequence and progress reports: detailed documentation is necessary

j. Completion report:

The design engineers should conduct detailed and thorough checking of the construction

to verify whether all the design details have been properly adhered to in the construction.

k. Photos and video recordings.

D. Post construction stage:

a. Trial runs (first loadings), instrumentation, observations and conclusions

b. Final approvals

c. First observation of distress/ malfunction

d. Observation of progress of distress, instrumentation and records

e. Emergency remedial measures taken.

CHAPTER 6

LABORATORY TESTS IN FORENSIC INVESTIGATIONS

Robinson R. G. Associate Professor, Dept. of Civil Engineering, IIT Madras, Chennai-600 036

robinson@iitm.ac.in ABSTRACT: Forensic geotechnical engineering involves systematic scientific investigations to detect the causes of failure or distress in a structure. Scientific method, involving the postulation of probable hypothesis and proving the hypothesis, is generally adopted in forensic investigations. In order to prove the hypothesis very often field and laboratory testing is needed. Conventional laboratory testing are also often very important. In this paper some of the aspects to be considered while analysing the experimental data obtained from laboratory studies are outlined. 1. INTRODUCTION

Field testing and laboratory testing form one of the major components of forensic investigations. The objective of the testing program is to determine the in-situ properties of the soil and the cause for distress. The field tests can be divided into two categories: non-destructive testing and destructive testing. Nondestructive testing such as geophysical surveys can be used to evaluate the ground conditions without causing damages to the site. The other category of non-destructive testing involves the determination of elevation of the ground and structure that will be used to estimate any settlement or heave.

Destructive testing involves removing a limited section of the building or undergoing

subsurface explorations. Details of destructive testing are given in Day (1998). The geotechnical forensic engineer has various choices among the following:

• Borehole investigations including SPT • Vane shear • Pressuremeter test • Flat dilatometer • Cone penetration tests like static cone, dynamic cone, piezocone, seismic cone,

nuclear density cone, Video cone, etc. • Plate load tests • Insitu density measurements, etc.

It is very important to note that the engineering properties derived from the field tests

depends strongly on the type of test and method of analysis. Therefore, they should be used to test the hypothesis with caution. Also the following aspects should be considered during interpretations.

• Variations due to seasonal change • Variation in ground water table • Variation due to the new stress path the soil has undergone due the construction of the

structure in question

• Variation due to any environmental change like chemical contamination, temperature change, etc.

2. LABORATORY TESTS

The common laboratory tests in geotechnical investigations are: (i) Classification Tests such as Specific gravity, Atterberg limits, Grain size distribution, etc. (ii) Strength Tests such as Direct shear, Triaxial shear, plane strain and simple shear (iii) Consolidation tests such as anisotropic consolidation, isotropic consolidation and can be either stress controlled or strain controlled. (iv) Permeability tests (v) Compaction tests, etc.

The in-situ properties may differ from properties that were used in the design. It is essential to evaluate whether the difference is caused due to the construction of the structure in question or due to other factors. Some aspects to be considered while interpreting the results are discussed below. 2.1 Classification tests

It is very important to evaluate the pre-test conditions adopted while doing the classification tests. For example the classification test results strongly depend on the degree and method of drying. It is the usual practice to oven dry the soil sample before conducting classification tests, though the code of practice suggests appropriate drying techniques. Liquid limit, plastic limit and shrinkage limit of some soils are severely affected by drying as can be seen in Table. 1. The grain size distribution is also significantly affected. Therefore, if any correlations were used in the design based on the index properties, the pre-test condition should be properly analysed. In addition, the soil classification also may get changed. For example, as per Indian Standard Classification system, the air-dried sample of Parur clay, in Table 1, falls as CH in the plasticity chart. However, the oven dried sample at 105 oC is classified as MH.

Table 1. Index properties of Parur Clay (data from Pandian et al. 1993)

Grain size Distribution Condition wL (%)

wP (%)

PI (%) Clay size (%) Silt size (%) Sand size (%)

Natural 106 47 59 51 42 7 Partially air dried 91 40 51 32 48 20 Air-dried (25-30oC) 85 34 51 21 52 27 Dried at 60oC 70 32 38 16 43 41 Dried at 105oC 60 32 28 15 39 46

The values of liquid limit also depend on the type of apparatus used for the evaluation. Sridharan and Prakash (2000) observed that the percussion method gives higher liquid limit values for the montmorillonitic soils than the cone method and that the cone method gives higher liquid limit values than the percussion method for kaolinitic soils. It may

be noted that the plasticity chart is based on the liquid limit value obtained from Casagrande’s apparatus. One may end up with different classification depending on the type of test adopted. 2.2 Compaction Tests

Very similar to the index properties, compaction characteristics also are affected by the pre-test conditions. If the compaction tests were conducted in the laboratory on oven-dried soil samples, the results cannot always be expected in the field that were initially wet but allowed to dry up to the optimum moisture content. Wesley (1973) reports such examples on allophone clays. It is also to be noted that over-compaction reduces the strength.

Fig. 1 Standard compaction curves for allophone clay (Wesley 1973) 2.3 Strength Tests Soils in general are not elastic materials and their behaviour in the field depends on

many factors including the magnitude of the imposed stress changes; the way in which they change; drainage conditions and the previous history of loading, etc. When a load is applied or removed from a mass of soil in the ground by a foundation or excavation, respectively, each element of soil experiences changes in its state of stress. A stress path gives a continuous representation of the relationship between the components of stress at a given point as they change (Lambe 1964, Lambe and Whitman 1969, Lambe and Marr 1979). Use of a stress path provides a geotechnical engineer with an easy recognisable pattern which assists him in identifying the mechanism of soil behaviour. It also provides a means of selecting and specifying the sequences of stresses to be applied to a sample in a test for a particular purpose. In geotechnical engineering practice, if the complete stress path of the problem is understood one is well along the way towards the solution of that problem (Holtz and Kovacs 1981). It is essential to consider the drainage conditions, rate of loading, stress

controlled or strain controlled tests, degree of saturation, etc. It is also well established in the literature that the type of test also has influence on the shear strength values (Leonards 1982) and the tests appropriate to the situation is to be selected. In practice, the soil is subjected to stress increments. But the standard tests are performed in a strain controlled apparatus. The effect of the type of loading is illustrated in the following fig.

The other important factor that affects the engineering properties is the sampling disturbance. Method of boring and sampling and their influence on the parameters are also to be analysed.

2.4 Volume change and Permeability The volume change behaviour of the soil depends on the type of clay mineral present in the soil and the pore medium. It is well known that the swell shrink behaviour of soils is due to the presence of the clay mineral montmorillonite. However, the non-swelling soils under normal conditions can exhibit heave when exposed to chemical solutions. For example, Rao and Rao (1994) reported a case study where the kaolinitic soils, commonly classified as non-swelling, showed excessive have and damage to structures due to caustic soda leakages on the ground.

The consolidation and permeability characteristics also depend to a large extent on the pore medium chemistry. Therefore, the possibility of volume change due to chemical alteration due to contamination of chemicals cannot be ruled out and due consideration be given in laboratory evaluations in such situations. 3. EFFECT OF WATER TABLE FLUCTUATION A single story building, founded on a silty stratum experienced severe distress with many cracks on the walls (Fig. 2a). Investigations indicated that the foundations are adequate in terms of bearing capacity, considering the worst case of water table on the surface. Consolidation tests indicated that the soil is overconsolidated and the settlement is within the permissible limits. It was observed that the water table fluctuates between the ground level during wet season (Fig. 2b) to a depth of 3 m during summer. This was confirmed by observing the water table levels in the nearby wells. The effect of water table fluctuation on the distress needs proper evaluation through experimental study.

Fig. 2a Crack in the building Fig. 2b View of the test pit

4. EXPANSIVE SOILS 4.1 Identification: Index properties and differential free swell tests are commonly used to identify expansive soils. It is often observed that the sites with soil that were classified as soils of low potential for swelling cause severe distress (Fig. 3). The basic properties of the soil are shown in Table. 2. From the table, the soil may be classified as Soil of low to marginal potential for swelling. However, the damage to the floors is severe. Probably, the sample needs to be examined by subjecting it to repeated cycles of wetting and drying.

Fig. 3 Distress due to swelling soil

Table 2. Index properties of soil

Property Value Atterberg Limits Liquid limit (%) Plastic limit (%) Plasticity Index (%) Shrinkage limit (%) Grain Size Analysis Sand fraction (%) Silt size (%) Clay size (%) Compaction Characteristics Maximum dry unit weight (kN/m3) Optimum moisture content (%) Swelling Characteristics Differential Free Swell (%) Swelling Protential (%)

39 20 19 10

40% 35% 25%

17.0 21

20 0.6

Potential for swelling Low to marginal

4.2 Swelling Pressure

An important variable required in the prediction of heave in swelling soils is the swelling pressure, which is the pressure required to hold the soil at constant volume when water is added. Knowledge of swelling pressure is essential for the design of a variety of geotechnical structures on expansive soils. The swelling pressure is evaluated in the laboratory by a number of testing methods which include oedometer testing of samples, suction measurements, triaxial methods, etc.

Out of all the methods, laboratory oedometer testing method is extensively used to

determine the swelling pressure due to its simplicity and operational ease. Brackley (1973)

lists three different oedometer methods for the determination of swelling pressure as explained below.

Method A - The sample is inundated and allowed to swell vertically at a small seating pressure until primary swell is completed. The sample is then loaded in intervals similar to the procedure of conventional consolidation testing until the specimen reaches its initial thickness. The pressure required to bring back the sample to its initial thickness is regarded as the swelling pressure. This method is also often termed as Swell-consolidation method. Method B - Three identical samples are loaded with different pressures near the expected swelling pressure and submerged in water. The vertical movements were plotted against the applied pressure and the pressure corresponding to zero volume change is taken as swelling pressure. While only one sample is enough to determine the swelling pressure in method A, at least three identical samples are needed in method B. This method is also often called as Different pressure method. Method C – In this method, also called as Constant volume method, a specimen is maintained at constant height by adjusting the vertical pressure after the specimen is inundated in free water. The pressure required to maintain constant volume is the swelling pressure.

Each of the methods is equally sensible, but gives entirely different swelling pressure values for the same placement conditions of the soil. A number of investigators have attempted to study the cause for the variation of the swelling pressure values by these methods. Johnson and Snethen (1978) compared the swelling pressure values by different oedometer methods and found that the magnitude of swelling pressure depends on the degree of confinement. Ali and Elturabi (1984) conducted methods A and C for the measurement of swelling pressure of expansive soils. Results obtained show that method A gives higher swelling pressure values than method C. Sridharan et al. (1986) compared the results from the three oedometer methods (Methods A, B and C) to determine the swelling pressure and concluded that method A gives an upper bound value, method B gives the least value and method C gives intermediate values. They also found no definite relation between the three methods. Soundara and Robinson (2009) also observed that the swelling pressure depend on the test method, as can be seen in Table. 2. The reason for this difference was attributed to the structure change that occurs during the tests. Therefore, it is essential to identify which method that was used to evaluate the swelling pressure during an investigation.

Table 3 - Comparison of Swelling Pressure values by different methods

(Soundara and Robinson, 2009)

Swelling pressure (kN/m2) by Method Sample A B C

A1 325 160 185 A2 270 150 175 B1 240 160 175 B2 210 115 140 B3 60 - 50

5. INTERFACIAL FRICTION ANGLE The interface friction between soils and construction materials is a very important input parameter required for the design of a variety of geotechnical structure. Studies indicated that the mode of shear significantly affect the magnitude of interface friction (Subba Rao et al. 1996). In the direct shear mode, if the test material is placed above the solid material, the limiting maximum value of interface friction is the critical state angle of internal friction of the soil. However, if the test material is placed below the soil, the limiting maximum value is the peak angle of internal friction of the sand. 6. SUMMARY Laboratory testing and field studies are important components of forensic investigations. Many of the parameters obtained in the laboratory depend on the pre-test conditions and also method of testing. The influence of pre-test conditions and the type of tests adopted needs to be carefully analysed in the forensic investigations. 7. REFERENCES Ali, E.F.M., and Elturabi, M.A.D. (1984). “Comparison of two methods for the measurement

of swelling pressure.” Proc. of 5th Int. Conf. on Expansive soils, Adelaide, Australia, 84/3, 72-74.

Brackley, J.J.A. (1973). “Swell pressure and free swell in compacted clay.” Proc. of 3rd Int. Conf. on Expansive soils, Haifa, 1, 169-176.

Chandrasekaran, V. S. (2003). Centrifuge Modelling: A useful technique for forensic geotechnical Engineering. Proc. Workshop on Forensic geotechnical Engineering, Chennai, pp. 57-60.

Day, R. W. (1998). Forensic Geotechnical and Foundation Engineering. McGraww-Hill. Holtz, R. D. and Kovacs, W. D. (1981). An Introduction to Geotechnical Engineering,

Prentice-Hall, 733 pp. Johnson, L.D., and Snethen, D.R. (1978). “Prediction of potential heave of swelling soils”.

Geotech. Test. J. ASTM, 1, 117-124. Lambe, T. W. (1964). Stress path method. Jl. Soil Mech. and Found. Div., ASCE, 93(SM6),

309-331. Lambe, T. W. and Whitman, R. V. (1969). Soil Mechanics, John Wiley and Sons, New York. Lambe, T. W. and Marr, W. A. (1979). Stress Path Method: Second Edition. Jl. Geotech.

Engg. Div., ASCE, 105(GT6), 727-738. Leonards, G. A. (1982). Investigation of failures. Jl. Of Geotechnical Engg. Div., ASCE, Vol.

108, GT2, 187-246. Pandian, N. S., Nagaraj, T. S. and Babu, G. L. S. (1993). Tropical Clays. I: Index Properties

and Microstructural Aspects. Jl. of Geotech. Engg. ASCE, Vol. 119, No. 5, 826-839. Rao, S. M. and Rao, K. S. S. (1994). Ground heave from caustic soda leakages- a case study.

Soils and Foundations, Vol. 34, 13-18. Sridharan, A., Rao, S.A. and Sivapullaiah, V. (1986). “Swelling pressure of clays”. Geotech.

Test.J., ASTM, 9(1), 24-33. Sridharan, A. and Prakash, K. (2000). Percussion and Cone Methods of Determining the

liquid Limit of Soils. Geotechnical Testing Journal, Vol. 23, No. 2, pp. 242–250.

Soundara, B. and Robinson, R. G. (2009). Influence of test method on swelling pressure of compacted clay. International Journal of Geotechnical Engineering, Vol. 3, No. 3, pp. 439-444.

Subba Rao, K. S., Allam, M. M. and Robinson, R. G. (1996). A note on the choice of interfacial friction angle. Geotechnical Engineering, ICE, London, 119(2), 123-128.

Wesley, L. D. (1973). Some basic engineering properties of halloysite and allophone clays in Java, Indonesia. Geotechnique, Vol. 23, No. 4, 471-494.

CHAPTER 7

BACK ANALYSES IN GEOTECHNICAL ENGINEERING

G L Sivakumar Babu and Vikas Pratap Singh Department of civil Engineering

Indian Institute of Science Bangalore 560012

1. Introduction

Back-analysis is an approach commonly used in geotechnical engineering to estimate operable material parameters in situ (Deschamps and Yankey 2006). Further, back analyses are required to provide technical evidences to prove or to disprove the hypotheses made on the cause of failures and to establish scenarios of failure (Hwang 2008). The approach of back analyses is popular because there are significant limitations in the use of laboratory and in-situ test results to accurately characterize a soil profile. Back analyses have been commonly used to study the causes of failures in geotechnical engineering applications such as slope stability, landslides, earth retaining structures, dams, highways and foundations. Back analyses are being used extensively in the geotechnical engineering practice, several studies including Leroueil and Tavenas (1981), Leonards (1982), Stark and Eid (1998), and Tang et al. (1999), describe various applications and the limitations of back-analyses. 2. Basic Considerations

In accordance with Hwang (2008), following points shall be considered while performing back analyses for the given problem. 1. Back analyses should be performed for ‘as-built” conditions because many of the

assumptions made in design are either non-existing or different from reality. 2. Reconnaissance of the site must be done preferably jointly with all the parties

involved so as to sort out differences in opinions, if any. Site observations may be captured in the form of photographs/videos that plays a vital role in decision making and provides support to the conclusions drawn.

3. To ensure that the results of the analyses are reliable, the data available must be

carefully verified. This includes an appropriate appraisal of local geology of the site of interest to help in understanding actual ground conditions and related historical events.

4. Design drawings and calculations, if available, must be checked to ensure that the works have been executed in accordance with appropriate design.

5. Depending upon the complexity of the problem, analyses can be performed by using

one or more of the following: (a) rules of thumb that includes indices such as stability number and overload factor, (b) empirical relationships, (c) closed from solutions, (d) simple numerical models, and (e) sophisticated numerical models.

6. While using numerical methods (generally using commercial software packages), an

understanding of the suitability, capability and the limitations of the method for the particular problem must be developed and the judicious interpretation of the output shall be made by the expert analyst. The software package adopted for analyses must have sufficient technical backup.

7. Depending on the algorithm and the numerical scheme adopted, different software

packages may lead to drastically different results. Therefore, it is necessary to conduct a few test runs so as to calibrate the parameters to be adopted by comparing the results of analyses with observations or with known solutions.

8. Following important source of errors in the numerical analyses shall be given due

importance while drawings conclusions and final recommendations based on the interpretations from the computed output: (a) implications of modeling of a 3D system as a 2D model, (b) modeling time dependency of soil behaviour (e.g., rate of dissipation of pore water pressures), and (c) modeling of the nonlinearity of the soil behaviour (i.e. choice of an appropriate constitutive soil model).

3. Scope

Follwing are some the areas that can be identified to have scope for back analyses in geotechnical engineering applications: 1. Back analyses may be used to study the settlement response of different types of

foundations and structures, classification of cracking damage, backfill settlement, to identify causes of settlements such as limestone cavities or sinkholes, consolidation of soft and /or organic soils, underground mines and tunnels, extraction of ground water or oil, landfills and decomposition of organic matter.

2. Back analyses may used to study failures/behaviour of structures founded on

expansive soil by analyzing factors such as lateral and vertical movements of expansive soils, special considerations in design of foundations and pavements on expansive soils and treatment measures.

3. Back analyses may be used to study potential causes of lateral movement in

applications such as rock falls, surficial slope failures, landslides, slope softening and creep, and dam failures.

4. Back analyses may be used to earthquakes induced phenomenons such as surface faulting and ground rupture, liquefaction, slope movement and settlement, and foundation behaviour.

5. Back analyses may be used to study bearing capacity failure of buildings, roads,

retaining walls and historical structures. 6. Back analyses may be used to study problems in structures such as slopes and

pavement due to the ground water and presence of moisture. 4. Examples of Potential Errors

Deschamps and Yankey (2006) provided a few project examples to illustrate the challenge and potential errors that can be present in back-analysis for material strengths, rupture (slip) surface, pore pressures, and three-dimensional or “End Effects”. A brief overview of the examples related to material strength and 3D effects from the study of Deschamps and Yankey (2006) is presented below. 4.1. Material strengths This example was drawn from a case history related to the Grandview Lake Dam located in Bartholomew County, central Indiana to illustrate the dependence that the back-calculated strength along a weak seam has on assumptions of strength in other zones. It was desired to perform back-analysis for estimating the strength along the planer slip surface. An assumed rupture surface was developed from inclinometer measurements. The dam was constructed primarily of glacial till and residual soils weathered from claystones. The first challenge was to select the operable strength of the dam materials. There was no distinct zonation of materials in the dam, and therefore, no basis for subdividing the dam into discrete materials.

To select the operable strength of the dam materials, results of the consolidated undrained (CU) triaxial tests were used. Four different characterizations of embankment strength were considered for back analyses purposes to characterize weak seam strength. A summary of the back-calculated friction angles in the weak seam is shown in Table 1.

The strength along the weak seam was characterized as having a zero cohesive

intercept because it was rationalized that this material was at or near its residual strength because the deformations along the very thin seam were significant, at least several inches. Moreover, there is little tendency for volume change at residual conditions such that shear-induced pore pressure changes were considered negligible

Table 1. Back-calculated strength (Deschamps and Yankey 2006)

Embankment strength Back calculated friction angle (degree)

Lower bound 22-24 Upper bound (high friction angle) 16 Upper bound (high cohesion) 11

Average 18 Table 1 illustrates the range in back-calculated friction angles is from 11°

to 24°, with the average strength providing 18°. Although this range can be viewed as extreme, it is apparent that even if a narrower range of strengths were used to characterize the dam, the back-calculated strength would still vary over an appreciable range. A conservative (low) estimate of embankment strength leads to a relatively high interpretation of strength along the weak seam. Note also the significant difference in back-calculated strengths for the two upper bound cases, wherein the case with primarily cohesive strength leads to a much lower back-calculated friction angle for the geometry considered because of the higher shear strength of the compacted materials. 4.2. Three-dimensional or “end effects” This example was drawn from a case history related to the ‘Lock and Dam 10’ on the Kentucky River to demonstrate the importance of understanding of three-dimensional effects when back-calculating strengths. ‘Lock and Dam 10’ is a relatively small concrete gravity dam owned by the Commonwealth of Kentucky and built circa 1905. The dam is a spillway over its complete length of 240 ft; it has a height of 34 ft; a width of 32 ft; and is made up of ten monoliths 24 ft long. Follwing are key features of the study: Problem: Stability analyses are conventionally performed on idealized two-dimensional cross sections, which are based on plane strain conditions. At ‘Lock and Dam 10’, coring through the dam indicated that the construction joints between concrete monoliths were essentially rubble, and could not be relied upon as shear connections between monoliths. Although it was considered imprudent to rely on the shear resistance between monoliths as a design consideration, it was recognized that some resistance was likely to be present. Back Analyses: An attempt was made to estimate the magnitude of this resistance in order to understand the inconsistency between required strength and interpreted strength. Accordingly, numerical modeling with the program FLAC was used to model the entire dam as a beam. The assumption was made that only a nominal frictional resistance (35°) was available between monoliths (no tensile or cohesive strength) and that the ends were fixed at the ends. Figure 1 illustrates the idealized model of the dam taken in plan view. The dam is attached to an abutment and training wall on the left, and the lock river wall on the right, both assumed to be stable. The distributed load on the beam was progressively increased to represent increasing the net hydrostatic pressures from higher pools.

Figure 1. Plan view of Kentucky River Dam No. 10, modeled as a beam (Deschamps and

Yankey 2006) Interpretation: The modeling effort produced a surprising result in which a zone within the dam formed a compressive arch that developed significant flexural resistance. Based on this analysis, the compressive arch that develops has sufficient capacity to carry the complete hydraulic load acting on the dam during the maximum design flood, independent of any frictional resistance at the base, and with only frictional resistance between monoliths. This example clearly illustrates how difficult it would be to back-calculate strengths if there is a significant, but uncertain, three-dimensional influence. Although, the three-dimensional effects are extreme in the present case, influences of 5 to 30% are generally believed to be expected. 5. Concluding Remarks

Back analyses provide a means to analyze the failures that frequently occurs in various technical fields including geotechnical engineering. It is of prime importance that back analyses must represent the in situ conditions to the extent possible. Choice of back analysis methodology must be based on the technical inputs and data available from the failure site, detailed laboratory investigations, and other factors such as complexity of problem, availability of experts, and the cost of analyses. Back analyses also form the key step in the field of forensic studies, and provides basis for the techno-legal aspects in geotechnical engineering. 6. References

Deschamps, R. and Yankey, G. (2006). Limitations in the back-analysis of strength from failures. ASCE Journal of Geotechnical and Geoenvironmental Engineering, 132(4), 532-536.

Hwang, R. N. (2008). Back Analyses in Forensic Geotechnical Engineering. ISSMGE/TC40 - Forensic Geotechnical Engineering. (Draft: 2008/10/20), 1-8.

Leonards, G. A. (1982). Investigation of failures. ASCE Journal of Geotechnical Engineering Division, 108(2), 185-246.

Leroueil, S., and Tavenas, F. (1981). Pitfalls of back-analysis. Proceedings of the 10th International Conference on Soil Mechanics and Foundation Engineering, Balkema, Rotterdam, 1, 185–190.

Stark, T. D. and Eid, H. T. (1998). Performance of three-dimensional slope stability methods in practice. ASCE Journal of Geotechnical and Geoenvironmental Engineering, 124(11), 1049-1060.

Tang, W. H., Stark, T. D., and Angulo, M. (1999). Reliability in back analysis of slope failures. Soils and Foundations, 39(5), 73–80.

1 INTRODUCTION

Correction of an existing landslide or the prevention of a pending landslide is a function of a reduction in the driving forces or an increase in the available re-sisting forces. Any remedial measure used must in-volve one or both of the above parameters. The IUGS Working Group on Landslides (Popescu, 2001) has prepared a short checklist of landslide re-medial measures arranged in four practical groups, namely: modification of slope geometry, drainage, retaining structures and internal slope reinforcement, as shown in Table 1. As many of the geological fea-tures, such as sheared discontinuities, are not well known in advance, it is better to put remedial meas-ures in hand on a “design as you go basis”. That is the design has to be flexible enough for changes dur-ing or subsequent construction of remedial works.

Although slope instability processes are generally seen to be “engineering problems” requiring “engi-neering solutions” involving correction by the use of structural techniques, non-structural solutions in-cluding classical methods such as drainage and mod-ification of slope geometry, as well as some novel methods such as lime/cement stabilization, grouting or soil nailing, are increasingly being used (Popescu,

1996). The cost of non-structural remedial measures is considerably lower when compared with the cost of structural solutions.

Terzaghi (1950) stated that, “if a slope has started to move, the means for stopping movement must be adapted to the processes which started the slide”. For example, if erosion is a causal process of the slide, an efficient remediation technique would in-volve armoring the slope against erosion, or remov-ing the source of erosion. An erosive spring can be made non-erosive by either blanketing with filter materials or drying up the spring with horizontal drains, etc.

Morgenstern (1992) followed this theme when he noted that post-failure analyses can be used to pro-vide a consistent explanation for landslide causal events. The back-analyses can then be used as a ba-sis for design of the stabilizing measures if engineer-ing works are required. This approach has the added appeal that the remedial design is normalized in terms of the post-failure analytical model.

Most landslides must usually be dealt with sooner or later. How they are handled depends on the proc-esses that prepared and precipitated the movement, the landslide type, the kinds of materials involved, the size and location of the landslide, the place or

Chapter 8

Back Analysis of Slope Failures to Design Landslide Stabilizing Piles

M. E. Popescu, Ph.D., P.E., Eur.Ing. Parsons Brinckerhoff Americas Inc. / Illinois Institute of Technology, Chicago, Illinois, U.S.A.

V.R. Schaefer, Ph.D., P.E. Iowa State University, Ames, Iowa, U.S.A.

ABSTRACT: It is generally accepted that shear strength parameters obtained by back analysis of slope

failures ensure more reliability than those obtained by laboratory or in-situ testing when used to design reme-dial measures. In many cases, back analysis is an effective tool, and sometimes the only tool, for investigating the strength features of a soil deposit. The fundamental problem involved is always one of data quality and consequently the back analysis approach must be applied with care and the results interpreted with caution. Procedures to determine the magnitude of both shear strength parameters (c' and φ') or the relationship be-tween them by considering the position of the actual slip surface within the failed slope are discussed Using the concept of limit equilibrium the effect of any remedial measure (drainage, modification of slope geome-try, restraining structures) can easily be evaluated by considering the intercepts of the c'-tan φ' lines for the failed slope (c0', tan φ0') and for the same slope after installing some remedial works (c'nec, tan φ'nec), respec-tively. This procedure is illustrated to design piles to stabilize landslides taking into account both driving and resisting force acting on each pile in a row as a function of the non-dimensional pile interval ratio B/D. The accurate estimation of the lateral force on pile is an important parameter for the stability analysis because its effects on both the pile-and slope stability are conflicting. That is, safe assumptions for the stability of slope are unsafe assumptions for the pile stability, and vice-versa. Consequently in order to obtain an economic and safe design it is necessary to avoid excessive safety factors.

components affected by or the situation created as a result of the landslide, available resources, etc. The technical solution must be in harmony with the natu-ral system, otherwise the remedial work will be ei-ther short lived or excessively expensive. In fact, landslides are so varied in type and size, and in most instances, so dependent upon special local circum-stances, that for a given landslide problem there is more than one method of prevention or correction that can be successfully applied. The success of each measure depends, to a large extent, on the degree to which the specific soil and groundwater conditions are prudently recognized in an investigation and in-corporated in design.

In this paper a methodology involving back anal-ysis of the slope failure and the use of piles to reme-diate the landslide are presented.

2 BACK ANALYSIS OF FAILED SLOPES TO DESIGN REMDIAL MEASURES

2.1 Failure envelope parameters A slope failure can reasonably be considered as a full scale shear test capable to give a measure of the strength mobilized at failure along the slip surface. The back calculated shear strength parameters, which are intended to be closely matched with the observed real-life performance of the slope, can then be used in further limit equilibrium analyses to de-sign remedial works. The limit equilibrium methods forming the framework of slope stability/instability analysis generally accept the Mohr-Coulomb failure criterion: τf = c'+ σ' tan φ' (1) where τf and σ' are the shear stress and effective normal stress respectively on the failure surface and c' and φ' are parameters assumed approximately constant for a particular soil.

A significant limitation in the use of this criterion is that the constant of proportionality is not really a constant when wide range of stress is under consid-eration. There is now considerable experimental evi-dence to show that the Mohr failure envelope exhib-its significant curvature for many different types of soil and compacted rockfill. Therefore, if the as-sumption of a linear failure envelope is adopted, it is important to know what range of stress is appropri-ate to a particular slope instability problem. To avoid this difficulty a curved failure envelope can be approximated by the following power law equation: τf = A (σ')b (2) which was initially suggested by De Mello (1977) for compacted rockfills and subsequently found ap-propriate for soils (Atkinson and Farrar, 1985).

Table 1. A brief list of landslide remedial measures

1. MODIFICATION OF SLOPE GEOMETRY 1.1. Removing material from the area driving the land-

slide (with possible substitution by lightweight fill)

1.2. Adding material to the area maintaining stability (counterweight berm or fill)

1.3. Reducing general slope angle 2. DRAINAGE

2.1. Surface drains to divert water from flowing onto the slide area (collecting ditches and pipes)

2.2. Shallow or deep trench drains filled with free-draining geomaterials (coarse granular fills and geosynthetics)

2.3. Buttress counterforts of coarse-grained materials (hydrological effect)

2.4. Vertical (small diameter) boreholes with pumping or self draining

2.5. Vertical (large diameter) wells with gravity draining

2.6. Subhorizontal or subvertical boreholes 2.7. Drainage tunnels, galleries or adits 2.8. Vacuum dewatering 2.9. Drainage by siphoning 2.10. Electroosmotic dewatering 2.11. Vegetation planting (hydrological effect)

3. RETAINING STRUCTURES 3.1. Gravity retaining walls 3.2. Crib-block walls 3.3. Gabion walls 3.4. Passive piles, piers and caissons 3.5. Cast-in situ reinforced concrete walls 3.6. Reinforced earth retaining structures with strip/ sheet

- polymer/metallic reinforcement elements 3.7. Buttress counterforts of coarse-grained material (me-

chanical effect) 3.8. Retention nets for rock slope faces 3.9. Rockfall attenuation or stopping systems (rocktrap

ditches, benches, fences and walls) 3.10. Protective rock/concrete blocks against erosion

4. INTERNAL SLOPE REINFORCEMENT 4.1. Rock bolts 4.2. Micropiles 4.3. Soil nailing 4.4. Anchors (prestressed or not) 4.5. Grouting 4.6. Stone or lime/cement columns 4.7. Heat treatment 4.8. Freezing 4.9. Electroosmotic anchors 4.10. Vegetation planting (root strength

mechanical effect)

2.2 Procedures for back analysis of slope failures Shear strength parameters obtained by back anal-

ysis ensure more reliability than those obtained by laboratory or in-situ testing when used to design re-

medial measures. In many cases, back analysis is an effective tool, and sometimes the only tool, for in-vestigating the strength features of a soil deposit (Duncan, 1999). However one has to be aware of the many pitfalls of the back analysis approach that in-volves a number of basic assumptions regarding soil homogeneity, slope and slip surface geometry and pore pressure conditions along the failure surface (e.g. Leroueil & Tavenas 1981). A position of total confidence in all these assumptions is rarely if ever achieved.

While the topographical profile can generally be determined with enough accuracy, the slip surface is almost always known in only few points and inter-polations with a considerable degree of subjectivity are necessary. Errors in the position of the slip sur-face result in errors in back calculated shear strength parameters. If the slip surface used in back analysis is deeper than the actual one, c' is overestimated and φ' is underestimated and vice-versa.

The data concerning the pore pressure on the slip surface are generally few and imprecise. More ex-actly, the pore pressure at failure is almost always unknown. If the assumed pore pressures are higher than the actual ones, the shear strength is overesti-mated. As a consequence, a conservative assessment of the shear strength is obtainable only by underes-timating the pore pressures.

Procedures to determine the magnitude of both shear strength parameters or the relationship be-tween them by considering the position of the actual slip surface within a slope are discussed by Popescu and Yamagami (1994). The two unknowns - i.e. the shear strength parameters c' and φ' - can be simulta-neously determined from the following two re-quirements:

(a) F = 1 for the given failure surface. That means

the back calculated strength parameters have to satisfy the c'-tan φ' limit equilibrium relation-ship;

(b) F = minimum for the given failure surface and the slope under consideration. That means the factors of safety for slip surfaces slightly inside and slightly outside the actual slip surface should be greater than one (Fig.1a).

Based on the above mentioned requirements, Sai-

to (1980) developed a semi-graphical procedure us-ing trial and error to determine unique values of c' and tan φ' by back analysis (Fig.1b). An envelope of the limit equilibrium lines c' - tan φ', corresponding to different trial sliding surfaces, is drawn and the unique values c' and tan φ' are found as the coordi-nates of the contact point held in common by the en-velope and the limit equilibrium line corresponding to the actual failure surface. A more systematic pro-cedure to find the very narrow range of back calcu-

lated shear strength parameters based on the same requirements is illustrated in Fig.1c.

Figure 1. Shear strength back analysis methods. The procedures discussed above to back calculate

the linear strength envelope parameters, c' and φ' in equation (1) can be equally applied to back calculate the nonlinear strength envelope parameters, A and b in equation (2) (Popescu et al., 1995).

The fundamental problem involved is always one of data quality and consequently the back analysis approach must be applied with care and the results interpreted with caution. Back analysis is of use only if the soil conditions at failure are unaffected by the failure. For example back calculated parameters for a first-time slide in a stiff overconsolidated clay could not be used to predict subsequent stability of the sliding mass, since the shear strength parameters will have been reduced to their residual values by the failure. In such cases an assumption of c' = 0 and the use of a residual friction angle, φr is warranted (Bromhead 1992). If the three-dimensional geomet-

TSS 1′

c′

tan φ′

F

TSS 1

FS

1.0

1 0

1′ Trial Sliding Surfaces, Failure

Surface, (a

0 1′

tan φ′

c′1

Envelope

Region II

Region I

Contact Points

(b)

Range of shear strength

(c

rical effects are important for the failed slope under consideration and a two-dimensional back analysis is performed, the back calculated shear strength will be too high and thus unsafe.

2.3 Design of remedial measures based on back analysis results

In order to avoid the questionable problem of the representativeness of the back calculated unique set of shear strength parameters a method for designing remedial works based on the limit equilibrium rela-tionship c' - φ' rather than a unique set of shear strength parameters can be used (Popescu, 1991).

The method principle is shown in Fig. 2. It is considered that a slope failure provides a single piece of information which results in a linear limit equilibrium relationship between shear strength pa-rameters. That piece of information is that the factor of safety is equal to unity (F=1) or the horizontal force at the slope toe is equal to zero (E=0) for the conditions prevailing at failure. Each of the two conditions (F=1 or E=0) results in the same relation-ship c'-tan φ' which for any practical purpose might be considered linear.

The linear relationship c'-tan φ' can be obtained using standard computer software for slope stability limit equilibrium analysis by manipulations of trial values of c' and tan φ' and corresponding factor of safety value. It is simple to show that in an analysis using arbitrary φ' alone (c'=0) to yield a non-unity factor of safety, Fφ*, the intercept of the c'-tan φ' line (corresponding to F=1) on the tan φ' axis results as:

tan φ0' = tan φ' / Fφ* (3) Similarly the intercept of the c'-tan φ' line (corre-

sponding to F=1) on the c' axis can be found assum-ing φ'=0 and an arbitrary c' value which yield to a non-unity factor of safety, Fc*:

c0'=c' / Fc* (4) Using the concept of limit equilibrium linear rela-

tionship c'-tan φ', the effect of any remedial measure (drainage, modification of slope geometry, restrain-ing structures) can easily be evaluated by consider-ing the intercepts of the c'-tan φ' lines for the failed slope (c0', tan φ0') and for the same slope after in-stalling some remedial works (c'nec, tan φ'nec), respec-tively (Figure 2). The safety factor of the stabilized slope is:

⎟⎟⎠

⎞⎜⎜⎝

⎛===

nec

0

nec

0c 'tan

'tanF,'c

'cFminFφφ

φ (5)

Errors included in back calculation of a given slope failure will be offset by applying the same re-sults, in the form of c' - tan φ' relationship, to the de-sign of remedial measures.

(a)

(b)

(c)

(d) Figure 2. Limit equilibrium relationship and design

of slope remedial measures.

LIMIT EQUILIBRIUM RELATIONSHIP FOR THE FAILED SLOPE: Fs=1 or En=0

LIMIT EQUILIBRIUM RELATIONSHIP FOR THE STABILIZED SLOPE

c′nec c′0

tan φ′nec

tan φ′0

tan φ′

c′

INITIAL GEOMETRY

LOADING PASSIVE PARTS

tan φ′

c′

UNLOADING ACTIVE

Load Unload

BEFORE DRAINAGE (WT 2)

AFTER DRAINAGE (WT 2)

tan φ′

c′

WT 1 WT 2

NO PILE

PILE INTERVAL (B/D)1

PILE INTERVAL (B/D)2<(B/D)1

tan φ′

c′

d d B D

PILE ROW

The above outlined procedure was used to design piles to stabilize landslides (Popescu, 1991) taking into account both driving and resisting force. The principle of the proposed approach is illustrated in Figure 3 which gives the driving and resisting force acting on each pile in a row as a function of the non-dimensional pile interval ratio B/D. The driving force, FD, is the total horizontal force exerted by the sliding mass corresponding to a prescribed increase in the safety factor along the given failure surface. The resisting force, FR, is the lateral force corre-sponding to soil yield, adjacent to piles, in the hatched area shown in Figure 3. FD increases with the pile interval while FR decreases with the same interval. The intersection point of the two curves which represent the two forces gives the pile interval ratio satisfying the equality between driving and re-sisting force.

Figure 3. Driving vs. resisting force for stabilizing piles

The accurate estimation of the lateral force on pile is an important parameter for the stability analy-sis because its effects on both the pile-and slope sta-bility are conflicting. That is, safe assumptions for the stability of slope are unsafe assumptions for the pile stability, and vice-versa. Consequently in order to obtain an economic and safe design it is necessary to avoid excessive safety factors.

The problem is clearly three-dimensional and some simplification must be accepted in order to de-velop a two-dimensional analysis method based on the principles outlined above. However the only simplicity to be accepted and trusted is the simplic-ity that lies beyond the problem complexity and

makes all details and difficulties simple by a sound and profound understanding.

3 APPLICATION

3.1 Site Conditions The described methodology is applied to a landslide in Ohio in the United States. The site is located along the Ohio River in south-central Ohio. A re-placement bridge was proposed at the site and site preparations reactivated an ancient slide. The cross section is shown in Figure 4. The slope consists of shale bedrock overlain by shale weathered to a re-sidual clay. Overlying the residual clay is alluvial silts and clays. Construction activities at the site led to the reactivation of an ancient slide. The slip plane discerned from surface scarps and inclinometer data is shown in Figure 4. It can be seen that the failure surface is planar in nature and occurs just above the shale bedrock in the weathered residual clay.

To accommodate the new bridge a fill was pro-posed on the existing slope, which was now moving and would have exacerbated the instability. Hence the use of piles to stabilize the slope was proposed.

Figure 4. Cross section of slope.

3.2 Back analysis Back analyses were conducted of the slope failure using limit equilibrium techniques as described pre-viously (Popescu, 1991). The back analyzed rela-tionship between friction angle and cohesion for the residual clay and the failure surface are shown in Figure 5. The resulting strength parameters vary depending upon the water level in the Ohio River. The relationship between the friction angle and the cohesion are shown in Figure 5 for cases of no pile and pile B/D ratios of 0.75 and 0.5. The back ana-lyzed friction angle of about 13º for the no pile case compares favorably with residual shear test results.

The numerical results for the B/D ratios in Figure 5 were obtained using the methodology proposed by Liang (2002) and coded into an Excel spreadsheet.

46

48

50

52

54

56

58

60

62

0 10 20 30 40 50 60

Length (ft)

Shale

Failure Surface Residual Clay

Alluvial Silts & Clays

New Fill

Ohio River

Existing Fill

Hei

ght

PILE ROW

d B d D

DOWNHILL

UPHILL

FD FR

B/D

FD, FR

narrowly spaced piles

largely spaced piles

FD – Driving Force

FR – Resisting Force

Figure 5. Back analyzed relationship between fric-

tion angle and cohesion.

3.3 Driving and resisting forces The driving forces were determined using limit

equilibrium analyses utilizing the program XSTABL (Interactive Software Designs, Inc. 1994) and spreadsheet analyses. The driving forces are shown in Figure 6 for various B/D ratios. The resisting forces were determined using the Ito and Matsui (1975) method as outlined by Popescu (1995). The resisting forces are shown in Figure 6 for various B/D ratios.

Figure 6. Driving and resisting forces as a function of

B/D ratio. From the results in Figure 6 it can be seen that the

resisting force and driving force cross at a B/D ratio slightly larger than 0.5 with a required resisting force of about 1800 kips. A shear force of this mag-nitude could be obtained using six-foot diameter shafts; however, eight-foot diameter shafts were se-

lected to provide a margin of safety for the drilled shafts.

4 CONCLUSIONS

This paper has outlined an approach to back analyz-ing the strength parameters in a slope failure and de-termining the force required to stabilize a slope us-ing piles considering the back analysis results. The use of the technique has been demonstrated through application to a case history.

5 REFERENCES

Atkinson, J.H., Farrar, D.M. 1985. Stress path tests to measure soils strength parameters for shallow landslips. Proc.11th Int. Conf. Soil Mech. Foundation Eng., San Francisco, 2: 983-986.

Bromhead, E.N. 1992. Slope Stability. 2nd Edition, Blackie Academic & Professional, London, 411 pp.

De Mello, V.F.B. (1977). Reflections on design decisions of practical significance to embankment dams. Géotechnique, 27(3):281-354.

Duncan, J.M. 1999. The use of back analysis to reduce slope failure risk. J. Boston Soc. Civil Eng. 14:1:75-91.

Interactive Software Designs, Inc. 1994. XSTABL An Inte-grated Slope Stability Analysis Program for Personal Com-puters. Reference Manual, Version 5, Moscow, ID.

Ito, T. and Matsui, T. 1975. Methods to estimate landslide forces acting on stabilizing piles. Soils and Foundations, 15:4:43-59.

Liang, R.Y. 2002. Drilled shaft foundations for noise barrier walls and slope stabilization. University of Akron, Akron, Ohio, Report prepared for the Ohio DOT and FHWA.

Leroueil, S. and Tavenas, F. 1981. Pitfalls of back-analyses. Proc. 10th Int. Conf. Soil Mech. Found. Eng. 1:185-190.

Popescu, M.E. 1991. Landslide control by means of a row of piles. Keynote paper. Proc. Int. Conf. on Slope Stability En-gineering, Isle of Wight, Thomas Telford, 389-394.

Popescu, M.E. 1995. Keynote Lecture: Back analysis of slope failures to design stabilizing piles. 2nd Turkish Symposium on Landslides, 15-26 October, Adapazari, Turkey.

Popescu, M.E. 1996. From Landslide Causes to Landslide Re-mediation, Special Lecture. Proc. 7th Int. Symp. on Land-slides, Trondheim, 1:75-96.

Popescu M.E. 2001. A Suggested Method for Reporting Land-slide Remedial Measures. IAEG Bulletin, 60, 1:69-74.

Popescu M.E., Yamagami T. 1994. Back analysis of slope fail-ures - a possibility or a challenge? Proc.7th Congress Int. As-soc. Eng. Geology, Lisbon, (6), p.4737-4744.

Popescu M.E., Yamagami T., Stefanescu S. 1995. Nonlinear strength envelope parameters from slope failures. Proc.11th ECSMFE, Copenhagen,. (1), p. 211-216.

Morgenstern, N.R. 1992. Keynote Paper: The role of analysis in the evaluation of slope stability. Proc. 6th International Symposium on Landslides, Christchurch, 3:1615-1629.

0

3

6

9

12

15

0 0.5 1Cohesion, ksf

Fric

tion

Ang

le, d

eg

No

B/D = 0.75

B/D = 0.50

0

1000

2000

3000

4000

5000

6000

0.30 0.40 0.50 0.60 0.70 0.80B/D Ratio

Forc

e, D

rivin

g &

Res

istin

g, k

ips

Driving Force

Resisting Force

Saito M. 1980. Reverse calculation method to obtain c and φ on a slip surface. Proc. Int. Symp. Landslides, New Delhi, 1:281-284.

Terzaghi, K. 1950. Mechanisms of Landslides, Geological So-ciety of America, Berkley, 83-123.

1

Chapter 9

BACK ANALYSES IN FORENSIC GEOTECHNICAL ENGINEERING

Dr. Richard N. Hwang richard.hwang@maaconsultants.com

Abstract: Back analyses are required to provide technical evidences to prove or to disprove the hypotheses made on the causes of failures and to establish scenarios of the failures. Analyses must be performed by experienced engineers who are familiar with the analytical tools to be adopted for analyses. The analytical tools adopted must be suitable for the cases to be investigated and the constitutive laws to be used in the analyses must be representative of the materials to be simulated. It is important to realize the fact that there are limitations associated with analyses and the results obtained must be interpreted by experienced engineers who have sufficient practical experience. Keywords: forensic, geotechnical, failure, analyses 1 INTRODUCTION

Failures seldom occur for a single reason and for cases in which litigation is involved, the causes of failures are inevitably difficult to ascertain. It is necessary to make various assumptions regarding why and how failures happen and perform analyses to prove or disprove these assumptions. Concession among the parties involved is often required in reaching conclusions. Sometimes, it has to be left to moderators to make final judgments.

Discussed herein are the general principles of back analyses. A case history is presented to illustrate the complexity of back analyses and the difficulties associated with the interpretation of the results obtained.

2 COMMON CAUSES OF FAILURES

Defective design of temporary works will likely lead to failures during construction and such

failures usually occur rather suddenly. The collapse of Nicoll Highway during the construction of the Circle Line of the Singapore Mass Rapid Transit System in April 2004 is a good example. The retaining system of a cut-and-cover tunnel collapsed suddenly while the excavation was about to reach the formation level and a section of the 6-lane expressway fell into the sinkhole. Failure started as the waling on the northern wall buckled at 9am on 20 April 2004 and by 3pm all the struts in a 100m section totally failed. The incident led to the death of 3 construction workers and 1 supervisory staff. The expressway was closed for seven and a half months. A Committee of Inquiry was immediately appointed by the Ministry of Manpower to investigate the causes of failure. It was concluded that the failure was mainly caused by the inadequate design of the retaining system (COI, 2005; Moh and Hwang, 2007; Yong and Lee, 2007). Effective stress parameters were adopted for marine clay which is essentially an undrained material, resulting in much overestimates of soil strengths and underestimates of wall deflections and bending moments. However, the failure was triggered by buckling of the stiffeners of walings and propagated to all other members of the retaining system.

Defective design of permanent works, on the other hand, will lead to problems which may last for a very long time. The leaning Tower of Pisa is a notable example. The construction of the tower started in 1173. The incline of the tower was already apparent when the construction was halted in 1178. The tower was only 10.6m tall and consisted of three stories then. The tower, 58m in height, was completed in 1319 and the bell-chamber was not finally added till 1372. Geographical surveys show that the tower is founded on loose sediments accumulated in a river course. The tilt was obviously caused by differential settlements. There have been several attempts made for preventing the tilting from worsening but resulted in only adverse consequences. The incline reached 5.5 degrees and the government officially closed the tower in 1990. The tilt was finally halted by extracting soil beneath the foundation on the north side and the tower was re-opened at the end of 2001. It is necessary to study the history of the tower and the measures taken in all these attempts to establish the entire scenario.

2

Failures are not always related to geotechnical problems. If there are no apparent ground movements, failures are most likely due to defective design of structures. A 15-year old building, Lian Yak Building (commonly referred to as “Hotel New World”), in Singapore collapsed in 1986 and claimed 33 lives. An intensive investigation was conducted by the government and failure of the foundation was one of the possible causes studied. Of 33 columns in the building, the foundations of 5 of them were selected for detailed examination and serious defects were observed in piles and in pile caps (Hulme, et al., 1993). However, it was found that the basement walls and slabs were virtually undamaged and there was no evidence of differential settlements. Checking of design calculations revealed that the dead loads of the structures were either grossly underestimated, as in the case of brickwalls, or completely missed, as in some case of self-weight of slabs. It was thus concluded that the failure was due to under-design of the structures and was not related to geotechnical problems.

Many failures could be attributed to conditions, such as adverse soil conditions, abnormal changes of groundwater level, heavy storms, or unexpected loading, etc., which were not accounted for in designs. Some of these conditions are unforeseeable but most of them are merely unforeseen because of the limited data available. Groundwater rushed into an arrival shaft as a portal was made on diaphragm wall for receiving the shield machine during the construction of the Taipei Metro. As a result, a section of twin tunnels was seriously damaged and 39 rings in the up-track and 34 rings in the down-track tunnels had to be replaced (Ju, Duann and Tsai, 1998). An PVC pipe was found at the invert of the up-track tunnel after the shaft was drained and the shield machine exposed. It could be onetime used for pumping water from the underlying water-bearing gravelly layer for irrigation or fish farming. The soils surrounding the tunnel had been solidified by jet grouting previously. A large piece of timber was found next to the pvc pipe and it is suspected that the treated ground was much disturbed, resulting in cracks which became water paths, as the shield machine forced its way out. Groundwater was then able to rush from the water bearing gravelly layer into the shaft from the abandoned pvc pipe. Drift woods were frequently encountered during constructions in the Taipei Basin, however, there is no way to ascertain their locations beforehand. It was also a common practice to pumping groundwater for irrigation and fish farming in the area, however, pumping wells are equally difficult to locate.

Failures due to defective constructions are not uncommon. Of the five disastrous events in metro constructions in Asia Pacific in the period of 2001 to 2007, one was caused by defective ground freezing work, one was caused by defective diaphragm wall and two were caused by defective ground treatment by jet grouting (Moh and Hwang, 2007). In all these 4 cases, water was the major source of the problem. In fact, water was responsible for a majority of failures in underground works carried out in soft ground.

3 BACK ANALYSES

A timeline showing all the events chronologically will be very helpful in understanding what has

happened. It will even be better if all major events can be presented in a Gantt chart with their durations clearly identified.

Analyses should be performed for “as-built” conditions because many of the assumptions made in design are either non-existing or different from reality. Therefore, reconnaissance of the site is a must and may have to be conducted for more than once. It is preferably to be conducted jointly with all the parties involved in the case so differences in opinions can be sorted out at the site. Photos and videos will be very useful in helping one’s memory and will provide vital supports to the conclusions to be made.

In many cases, failures involve soil-structural interaction and advices from structural engineers are desirable. For cases involving failure of structures, analyses should be performed jointly with structural engineers. 3.1 Data Collection and Verification A checklist should be prepared to document the data available in hand and the data which are still missing. The data needed will depend on the modes of failures. Although the cause of failure may appear to be obvious right at the beginning, it is still necessary to investigate all the potential modes of failure before they can be eliminated.

To ensure that the results of analyses are reliable, the data available must be carefully verified. For example, ground conditions may be mis-interpreted, instrument readings may be erroneous, construction activities may be wrongly logged, etc. Some of the data may be misleading without being realized, and some of data may be contradictory and have to be sorted out. It is natural for patties

3

involved to hind facts, intentionally or unintentionally, which are not in their favor, therefore, data must be critically reviewed and judgments must be applied whenever data appear to be dubious.

An appropriate appraisal of local geology sometimes is helpful in understanding ground conditions and historical events related to the site of interest should be investigated.

3.2 Checking Design and Calculations

If design drawings are available, it is necessary to check whether or not the works were

constructed as designed. Design calculations, if available, should be firstly checked to see if they are appropriate. However, inconformity with specifications and/or codes of practice will not necessarily lead to failures and the redundancy demanded by specifications and codes may not be required at the time of failure.

It should be noted that ground conditions might have been altered once a failure occurs and it is necessary to figure out what they were prior to the event. This sometimes may not always be possible and guesswork may sometimes be required.

3.3 Analytical Tools

Depending on the complexity of the problem, analyses can be performed by using: a) rules of thumb using indices such as stability number, overload factor, etc. b) empirical relationship c) closed form solutions d) simple numerical models e) sophisticated numerical models

Simple models are available for many failure modes and many of them can be solved by hand calculations. Complicated problems can be solved by numerical analyses. In any case, the tools adopted for analyses must be suitable for the type of problem to be solved and must have sufficient technical backup.

The analytical tools adopted for analyses must have been validated in accordance with stringent quality assurance program and well documented. For this reason commercial software packages are preferred to in-house programs because the former are usually well tested and improved based on the feedback from users.

Numerical methods have their limitations and the results can not be trusted blindly. Depending on the algorithm and the numerical scheme adopted, different software packages many lead to drastically different results. Therefore, a few test runs are necessary to calibrate the parameters to be adopted by comparing the results of analyses with observations or with known solutions.

3.4 Limitations

With the rapid advancement of computer technology, finite element method and finite difference method have become important tools for design and they are very useful tools for back analyses as well. They have become so user-friendly that even fresh graduates from colleges can perform the analyses with little guidance. This, however, leads to the danger of mis-handling of problems and mis-interpretation of the results of analyses. After all, geotechnical engineering is an art rather than science. The results of analyses provide important evidences for judgments to be based on and experience should prevail at the end.

It is important to realize the limitations associated with analyses so the results can be correctly interpreted. First of all, although tools are available, three-dimensional analyses are extremely labor intensive and time consuming. Therefore, unless the cases are critical enough and resources are available, one-dimensional and two-dimensional analyses are performed and the results are usually appropriate for practical purposes. However, simulation of a 3D system by a 2D model inevitably alters the nature of the problem and introduces errors in the results of analyses.

Secondly, soils are usually classified as either drained materials, i.e., pure sand, or undrained materials, i.e., pure clay, in analyses. However, in reality, most of soils are neither pure sand nor pure clay and are mixtures of sand and clay. Their behavior will depend on how fast porewater pressures dissipate. In other words, problems are time-dependent. The rate of dissipation of porewater pressure will not only depend on the rate of loading but will also depend on the permeability of soil and the

4

length of drainage path. The reliability of the results of analyses will depend on how well this time-dependency of soil behavior is handled.

Thirdly, it is a well-known fact that soil behavior is highly non-linear and there are many soil models to handle the nonlinearity. The results of analyses may vary considerably if different constitutive laws, which describe the stress-strain relationship of soils, are selected.

4 CASE STUDY

The collapse of Nicoll Highway, refer to Fig. 1, is an ideal case for illustrating the points

mentioned above. The site is located in a piece of land reclaimed in the 80’s. As shown in Fig. 2, the subsoils at this site contain mainly two thick layers of marine deposits (namely, the upper marine clay and the lower marine clay) and are underlain by the Old Alluvium which is a competent base stratum. Figure 3 is a plot of the results of a cone penetration test carried out in the vicinity of the site. The excavation was supposed to be carried out to a depth of 33.5m and diaphragm walls with a thickness of 800mm (locally, 1000mm) were used.

200m0mReference: COI, 2005Fig. 3

Golden MileComplexGolden Mile

Tower

Areaaffected

N

Fig. 1 Collapse of Nicoll Highway

資料來源： COI, 2005

I-65

I-10466kV Cable Crossing

M2

S325 S3

26 S327 S328

S329 S330

S331

S332

S333

S334

S335

S33 6

S337

S338

S339

S340

Waling buckled at9am, 20 April, 2004

M3

Fig. 3 Fig. 2 Retaining system in Section M3

5

Reference： COI, 2005

Old Alluvium

Lower MarineClay

F2 lower

F2 upper

UpperMarine

Clay

Fill

JGP

RL 102.90-1.00-4.00-7.50

-11.00-14.50-17.50-20.50-23.80-26.80-29.80-33.50

South

Diaphragm Walls800/1000mm

Estuarine

Estuarine

123456789

10

North

4/20441mm

I-104 I-65

3/26202mm

3/26282mm

4/20175mm

105

55

50

60

65

70

75

80

85

90

95

100

Ele

vatio

n, m

105

55

50

60

65

70

75

80

85

90

95

100

Reference： COI, 2005

Old Alluvium

Lower MarineClay

F2 lower

F2 upper

UpperMarine

Clay

Fill

JGP

RL 102.90-1.00-4.00-7.50

-11.00-14.50-17.50-20.50-23.80-26.80-29.80-33.50

South

Diaphragm Walls800/1000mm

Estuarine

Estuarine

123456789

10

123456789

10

North

4/20441mm

I-104 I-65

3/26202mm

3/26282mm

4/20175mm

105

55

50

60

65

70

75

80

85

90

95

100

105

55

50

60

65

70

75

80

85

90

95

100

Ele

vatio

n, m

105

55

50

60

65

70

75

80

85

90

95

100

105

55

50

60

65

70

75

80

85

90

95

100

Fig. 3 Ground conditions and excavation scheme in Section M3

4.1 The Incident Excavation for constructing the cut-and-cover tunnels was carried out by using the bottom-up

method of construction. The collapse occurred in Section M3 on 20 April 2004 while the 10th dig was completed and excavation reached a depth of 30.5m four days earlier. There were two inclinometers, i.e., I-65 and I-104, available for monitoring wall deflections. Figure 4 shows the wall deflection paths, which are the plots of maximum wall deflections versus depth of excavation in a log-log scale, for these two inclinometers (Hwang, Moh and Wong, 2007). Wall deflections on the two sides of the excavation were about the same till 9 March, 2004, when excavation reached a depth of 25m, and deflections of 198mm and 215mm were recorded by Inclinometers I-65 and I-104, respectively. Subsequently, there was a period in which I-104 was not read because it was damaged. When monitoring resumed on 26 March, the deflection of the southern wall was found to have increased by 67mm to 282mm while the readings for Inclinometer I65 on the north were fairly steady in this period. The readings for Inclinometer I-104 kept on increasing while those for I-65 remained to be steady subsequently, presumably, because of the asymmetry of ground conditions. In fact, I-65 appeared to move outward by 27mm and the maximum deflection reduced from 202mm on 26 March to 175mm on 20 April, as depicted in Figs. 2 and 4. On the other hand, Inclinometer I-104 moved inward by 90mm to 441mm in the 3-day period from 17 April to 20 April.

0

10

20

30

40

50

0 5 10 0.00 0.05 0.10

Tip Resistance, qpMPa

Shaft Friction, fsMPa

Dep

th, m

0

10

20

30

40

50

0 5 10 0.00 0.05 0.10

Tip Resistance, qpMPa

Shaft Friction, fsMPa

Dep

th, m

Fig. 4 Results of a cone penetration test

Failure started as the waling at the northern end of Strut S338 at the 9th level, refer to Fig. 2, buckled at 9am on 20 April and failure propagated to other struts. By 3pm of the day, all the struts for a 100m section totally failed. As Nicoll Highway sank, gas, water and electricity cables ruptured,

6

causing power to go out for about 15,000 people and 700 businesses in the Marina and Suntec City area. Tremors were felt at Golden Mile Complex. Tenants and residents in the building were also evacuated.

1

10

100

1 10 100 1000 10000

Maximum Wall Deflection, mm

Dep

th o

f Exc

avat

ion,

m

I-65

I-104

1

10

100

1 10 100 1000 10000

Maximum Wall Deflection, mm

Dep

th o

f Exc

avat

ion,

m

I-65

I-104

Fig. 5 Deflection paths for diaphragm walls 4.2 The Design

The contract was awarded as a design-and-build contract and temporary works were designed by

the contractor. The computer program PLAXIS, which was first developed in 1987 at the Technical University of Delft and made commercially available by PLAXIS bv of Delft, Netherlands, was adopted in the design of retaining system. PLAXIS allows users to select different material types (drained, undrained and non-porous) and different material modes (Mohr-Coulomb, soft soil, etc.) Mohr-Coulomb Model was adopted for defining failure of soils for the case of interest. When using the Mohr-Coulomb model, it is possible to input either effective stress parameters (c’ and φ’) or the undrained strength parameters (c’ = cu and φ’ = 0). The use of a Mohr-Coulomb soil model with effective strength parameter has been referred to as Method A in the report by the Committee of Inquiry (COI, 2005). Method B refers to the use of Mohr Coulomb soil model with undrained strength parameters in combination with undrained material type. Method A was adopted in the analyses and effective strength parameters, c’ and φ’, instead of the undrained strength parameters, c’ = cu and φ’ = 0, were used for marine clays.

This, as depicted in Fig. 6, drastically over-estimated the undrained strength of clays and analyses indicated that wall deflections and the associated bending moments were grossly under-estimated by a factor of, roughly, 2. However, except Levels 6 and 9, the strut loads computed by using Model A were larger than those computed by using Model B. The strut loads computed by using Model B were 107% of those computed by using Model A at Level 6 and 110% at Level 9.

Kf

2cu

Confining stress p'. p

Total stress path

φ'effective stress path

q

(a) Model A

7

Kf

2cu

Confining stress p'. p

Total stress path

φ' effective stress path

q

(b) Model B Fig. 6 Comparison of undrained strength of clay in Models A and B 4.3 Back Analyses

Back analyses were carried out by reducing the stiffness of soils and JGP slabs to match the

performance of the wall when the reading of I-65 reached 159mm, exceeding the alert level of 105mm, as the excavation reached a depth of 18m in the 6th stage of excavation on February 23, 2004. The results of analyses predicted a maximum wall deflection of 253mm in the subsequent stages of excavation. On April 1, the reading of I-104 reached 302.9mm, back analyses were again carried out by further reducing the stiffness of materials and the maximum deflection was revised to 359mm. Model A was continuously adopted in these back analyses.

4.3 Forensic Studies

Subsequent to occurrence of failure, the Singapore Government immediately formed an

independent Committee of Inquiry (COI), headed by a Senior District Judge, to look into the incident. After thorough investigations, in which 173 witnesses were interviewed and 20 experts offered their professional opinions, an Interim Report was released on 13 September 2004 and a very comprehensive Final Report was made available to the public on 13 May 2005 (COI 2005).

The Committee identified critical design and construction errors, particularly the design of stiffeners on the walings at the connections between the diaphragm walls and the struts, that led to the failure of the earth retaining system. The Committee also found deficiencies in the project management that perpetuated and aggravated the design errors, including inadequate instrumentation and monitoring of works, improper management of instrumentation data, and lack of competency of persons carrying out specialized work.

At the request of the Land Transport Authority, the Engineering Advisory Panel (EAP) conducted an extensive study to review the case and the design of other parts of the Circle Line. Three-dimensional analyses were carried out using the commercial software ABAQUS to analyze the performance of the retaining system. Each of the first nine layers of excavation was modeled as a uniform lift with all the elements removed at the same time. The 10th excavation step was simulated in three sub-steps, starting with the removal of the elements at the eastern end of the excavation and proceeding westwards (Yong and Lee, 2007) .

4.4 Causes of Failure

As mentioned in Section 4.2, wrong strength parameters were adopted for marine clay, leading to

under-design of wall deflections and bending moments. Although back analyses were performed during the course of construction, they were performed by engineers who did not have experience on the program PLAXIS and the results were not reviewed by experienced engineers. In this regard, however, regarding soil parameters, the experts did differ in opinions in the Inquiry. It is amusing to note the statement made by the representative of the Contractor in the Inquiry that:

(Quoted) “….. despite more than six months of intensive work by the six teams of experts who have been reviewing the collapse at M3, the experts still can not reach any agreement on the correct input parameters to be adopted in a back analysis. Very significant difference still remain,

8

particularly respect to the parameters to be adopted for JGP. It is also noted that because of the stiffness characteristics of the ground, diaphragm wall and JGP are all highly non linear, it is virtually impossible to obtain agreement between the monitored strut loads and wall displacements throughout all the stages of excavation sequence using a single set of linear elastic stiffness value, as adopted in Plaxis analyses.” (Unquoted)

Although these statements may be exaggerating, they nevertheless illustrate the difficulty associated with analyses.

StrutLoad

Plate stiffener

diaphragmwall

(a) Original design

StrutLoad

waling

C-channel

diaphragmwall

(b) Revised design

Fig. 7 Stiffeners of walings Model A was adopted by the contractor in analyses for all the cut-and-cover sections in this

contract. Although considerable efforts have been made to deal with excessive wall deflections, the excavation in other sections was able to be completed. This indicates that the under-design of diaphragm walls is not solely responsible for the incident. The failure must be due to a combination of adverse factors.

As depicted in Fig. 3, the soft deposits on the two sides of the excavation differed in thickness while analyses were performed for only the southern half of the mesh on the assumption that the results would be conservative. This assumption may not be valid reality because the deflections of the southern wall tended to be larger due to the imbalance of earthpressures on the two sides of the excavation. The 3D finite element analyses conducted by EAP also indicated the significance of the curvature of the alignment. The wall panels tended to split at the joints as the wall moved inward as the excavation proceeded. This is particularly true for the south wall where wall deflections tend to widen the gaps.

The failure in fact was triggered by the buckling of walings. As mentioned in Section 4.2, except Levels 6 and 9, the strut loads computed by using Model A were larger than those computed by using Model B. Even for Levels 6 and 9, the strut loads computed by using Model B were 7% to 10% larger. Therefore the buckling was not a result of the use of wrong material model for clay.

The Inquiry revealed that the loads at some of the connections between struts and walings were under-estimated either because a wrong type of joint was installed or because slays were omitted. Secondly, around February 2004, several instances of stiffener plate buckling as well as the buckling of a waling were reported at the Nicoll Highway Station, the contractor proposed replacing the stiffener plates by C-channels in an attempt to improve the performance of the connection. The replacement of

9

double stiffener plates with C- channels, refer to Fig. 7, provided only minor improvement in terms of axial load bearing capacity for the waling connections, but this came at the expense of ductility. The change rendered it more susceptible to the brittle “sway” failure mode. This is clearly seen in results of the post collapse finite element analyses and physical tests. The poor detailing is exacerbated by the discontinuity of walings, as pointed out in the EAP study, which is particularly important at the curved south wall where deflection due to excavation caused separation of the joint between diaphragm wall panels.

The instrumentation and monitoring system was not properly executed and some of the crucial readings were found erroneous and misleading. This made it impossible to correlate the strut loads with the performance of the retaining system.

In addition to all the above-mentioned shortcomings in the design, the performance of the two grouted slabs was also questioned and investigated in the Inquiry and in the EAP study. The removal of sacrificial JGP layer above the formation level was not followed with timely installation and preloading of struts to compensate for the loss of reaction load sustained in the JGP layer. As a result, it caused significant lateral movement to the temporary walls and increase in load to the struts above. This triggered the local failure of connections at the joint between struts and walings at critical locations and contributed to the chain of collapse of diaphragm walls.

5 CONCLUSIONS

Failures frequently occur and back analyses enable their causes to be identified. The experience learned can then be passed on from generation to generation. For the results of analyses to be reliable, (1) the analytical tool used must be suitable for the type of problem to be analyzed (2) the analyses must be conducted by experienced engineers who have sufficient background on the

algorithm of analyses (3) the input parameters must be representative of the materials to be studied (4) the construction sequence must be reliable (5) instrument readings must be faithful (6) the results must be interpreted by experienced engineers who have sufficient practical experience

on the performance of soils and structures. It is however important to realize that there are limitations associated with analyses and the results obtained only play a role of providing supporting evidences and engineering judgments should always prevail. REFERENCES COI 2005. Final Report of the Committee of Inquiry into the Incident of the MRT Circle Line Worksite that Led

to the Collapse of Nicoll Highway on 20 April 2004, Presented by Committee of Inquiry to Minister for Manpower on 10 May 2005, Singapore

Hulme, T. W., Parmar, H. S., Hou, K. H. and Sripathy, P. (1993) The collapse of the Hotel New World, Singapore, a technical inquiry, The Structural Engineer, v71, no. 6/16, March, pp. 91~98

Hwang, R. N. and Moh, Z. C. Wong, K. S. 2007. Reference envelopes for deflections of diaphragm walls in Singapore Marine Clay, Proc., 16th Southeast Asian Geotechnical Conference, Kuala Lumpur, Malaysia, 8 ~ 11 May

Ju, D. H., Duann, S. W. and Tsai, H. H. (1998) Ground freezing for restoration of damaged tunnel, Proc., 13th Southeast Asian Geotechnical Conf., November 16~20, Taipei, Taiwan, pp. 615~620

Moh, Z. C. and Hwang, R. N. (2007) Lessons learned from recent MRT construction failures in Asia Pacific, Opening Keynote Address, Proc., 16th Southeast Asian Geotechnical Conf., 8~11 May, pp.3~20, also, Geotechnical Engineering, Special Issue, v38, no. 3, December, Bangkok, Thailand, pp. 121~137

Yong, K. Y. and Lee, S. L. (2007) Collapse of Nicoll Highway – A global failure at the curved section of a cut-and-cover tunnel construction, Chin Fun Kee Lecture, Proc., 16th Southeast Asian Geotechnical Conf., 8~11 May, pp.23~36

Chapter 10

GEOTECHNICAL FAILURE: THE CAUSE IS NOT ALWAYS OBVIOUS AND MAY BE COMPLEX

Jan E Hellings MBA, PhD, DIC, MSc, CEng, FICE,

Dr Jan Hellings & Associates, Headley, UK, RG19 8AB jan@janhellings.com

ABSTRACT: This article considers the failure of four projects, three in the UK and one in the Netherlands. In many cases, the failure is not straightforward and may be due to a number of causes rather than to a single cause. 1. INTRODUCTION Designers and constructors clearly do not anticipate that failure of their project will occur; the principle risks and concerns of any job are usually identified and effort is channelled into controlling, monitoring and providing for behaviour scenarios arising out of these. But, perhaps because of too focused a view, less critically perceived features can be overlooked. In concentrating on elements – efficiently and often very cleverly dealing with such elements – the overall ‘scheme’ view can be forgotten. The situation is often exacerbated by interfaces, whether it be by environment, by discipline, by techniques, by participants, by cultures or by trading practices. (Fernie 2007). This paper will describe four projects where failure has occurred, in some cases spectacularly and at great cost. The projects are: collapse of rail tunnels at London’s Heathrow Airport in 1994, the collapse of a section waste water tunnel in Kingston Upon Hull, UK in 1999, a failure during construction of the Tramtunnel in The Hague, The Netherlands in 1996, and finally the failure of foundations to a house in central London. 2. HEATHROW AIRPORT TUNNELS COLLAPSE One of the worst civil engineering disasters in the United Kingdom in the last quarter of a century occurred during the night of the 20-21st October 1994 when tunnels in the course of construction beneath Heathrow Airport’s central terminal area collapsed. They continued to collapse over the following days. The public and those engaged in the construction work were exposed to grave risk of injury. Although there were no injuries, many people were put at risk and the consequential costs was significant. A number of the lessons arising from this collapse can be applied to engineering projects generally.

The tunnel was constructed using the “NATM” (New Austrian Tunnelling Method) form of construction, or spayed concrete lining (SCL). The safety of this approach depends upon close monitoring as the work progresses; it was the first time it had been used in London clay. In view of this, an extensive full scale trial was undertaken at the site prior to initiating the actual works. The method involves the use of SCL, applied in stages in such a manner that the ‘stand up’ time of exposed clay is not compromised. The collapse is well documented in the UK Health and Safety Executive publication (HSE 2000). The HSE labelled this failure as having all the hallmarks of an “organisational accident”. The report states that “hazards were not identified by all the parties and risks were not controlled, during construction, through the “defensive” systems (i.e. preventative management systems) used by the parties”. The investigations’ findings were as follows: 2.1 Contractual Arrangements and Culture • A lack of awareness of risks: the risk

evaluation at the outset was not sufficiently comprehensive.

• The quality of workmanship was according to the “self certification” procedure which followed the competitive tendering arrangement. This, in the author’s view, is not as robust as an independent certification procedure.

• The separation of permanent and temporary NATM works design: this is often a problem in construction generally.

• Separation of compensation grouting and tunnelling monitoring processes.

2.2 Design

• There was a lack of appreciation of the differences between hard rock and soft clay behaviour.

• The design was not considered sufficiently robust.

• The invert of the tunnel was of a form which made construction tolerances more critical.

• The monitoring regime was unsatisfactory. • The ground conditions were as expected.

2.3 Construction Quality • The tunnel wall profiles were not correct. • There was defective invert construction. • There was defective joint construction. • The invert was “over flat”. 2.4 Construction Management • There was insufficient specialised staffing. • There was poor communication between

parties involved. • There was poor sequence of tunnel

construction. • There was bad timing of invert repairs. • There was no integration in planning

construction activities. • There was a lack of awareness of

instrumentation data warning of impending failure.

It can therefore be seen from the Heathrow experience that failures are likely to arise from multiple causes. Designs need to take account of the ease of construction. Monitoring should be continuous and autoprocessed and preferably available on the web. Supervision must be informed of the design objectives and there needs to be integration and communication of the team. The cost of recovery of the project is estimated to have been of the order of 250 million U.S dollars; there was no loss of life but a successful HSE prosecution. 3. HULL WASTE WATER FLOW TRANSFER TUNNEL COLLAPSE In November 1999, tunnel construction was progressing normally with 80% of the roof successfully completed (Grose and Benton 2005). About two weeks earlier the tunnel boring machine (TBM) heading westwards had passed through the previously constructed 7.5m diameter access shaft having stopped in the shaft for essential maintenance. The centre of the collapse was within

a few metres of the shaft which at that time was about 200m behind the face. The tunnel ring comprised 6 concrete segments. The rings were assembled with temporary bolts which were removed once there was access beyond the rear of the TBM train. The removal of these bolts is common in mainland Europe; the alternative would have been expensive stainless steel bolts and the bolts contribution to the rings strength and stiffness was ignored in the permanent design. The geology at the collapse location was different from that elsewhere along the route; in this location there was a deep infilled former channel of the River Hull the tunnel arisings here being in alluvium, whereas elsewhere the tunnel is in glacial deposits. The first sign of the impending collapse was a small leak not considered significant at the time between segments at the knee joints. The progress of the collapse, at least in its early stages, was seen by a number of witnesses. Sand, silt and clear water was observed bubbling through the joint. The invert segments began to twist and at this time leaks from the shoulder joints worsened. This was approximately 6 hours after the leak had first been noticed. As the situation worsened significant volumes of sand had flowed into the tunnel, enough to derail a train which remained in the collapsed zone until recovery some months later. Segments were spalling, and the extent of damaged segments were spreading away from the shaft. Just before the tunnel was abandoned, spalled concrete was observed to be dropping from the crown segments. Movement continued at the surface until a crater 60m in diameter and 2.5m deep had been formed, the centre of which was around 3m east of the shaft. Fortunately the area around the shaft was being used as a surface car park. A relatively unusual feature of the collapse was that there was no immediately obvious cause or contributing factors. Despite a very thorough investigation, the team did not find an incontrovertible single root cause for the collapse. Although in some ways the situation was not the desired outcome, pragmatic balance had to be struck between the cost of investigations desire to provide answers.

It was established that the two primary factors leading to the collapse were: (i) The presence of substantial volumes of fine sand, under considerable water pressure, adjacent to the tunnel; and (ii) A leak through the lining large enough to allow sand to wash through. All the credible causation factors were investigated to the extent that their likelihood of occurrence could be judged, and in this way steps could be taken to ensure safe completion and commissioning of the tunnel. 4. THE HAGUE TRAMTUNNEL The Tramtunnel was constructed using the ‘cut and cover’ system; the walls of the tunnel were constructed variously of diaphragm and sheetpile walls. After construction of the roof slab, excavation took place beneath. An arched jet grout layer was installed between the toes of the diaphragm and sheetpile walls which was intended to fulfil three functions: • Lateral support for the walls. • Water resistant layer to minimise water

seepage into the building pit. • Resist upward water pressure The jet grout-arch consists of interlocking jet grout columns with varying heights of 1.5m at the crown to 2.5m at the side of the arch. The design relied on a completely watertight arch. In 1998, during the last stage of excavation, a serious sandboil occurred. As a result 60m 3 of soil flowed into the tunnel, leaving a hole in the street next to the tunnel. As this was the first part of the tunnel to be constructed, all parties involved lost confidence in this construction method. Research into the collapse (Van Toll and Sellmeijer 2003) found that there had in the years before the construction of the Tramtunnel been several near catastrophic events involving these jet grouts screens in Cairo, Berlin and Bilbao. The research focused on two aspects: the reliability of jet grout screens and the consequences of imperfections in jet grout screens. It was realised that all injection layers will show some imperfections. This would allow ground water from beneath the gel layer to percolate through the gel layer along the ‘imperfection’; for a

sandboil not to occur therefore, the percolating water should be discharged in a controlled manner using a drainage system in the soil above the jet grout later, without the risk of erosion. A covering layer of sand of certain thickness is therefore necessary and was absent to a sufficient extent in the case of the Tramtunnel. Compressed air was introduced in the tunnel while imperfections in the jet grout layer were repaired by additional grouting. 5. LONDON HOUSE FOUNDATION FAILURE The house is a two storey building constructed over a hundred years ago. The house was originally constructed with a shallow strip footing on London clay. The present owner purchased the property in 1994 and in 1995 noticed cracking appearing in the walls of the property. There are six very large lime trees in close proximity to the house. In 1997 an investigation was undertaken and it was decided that the property needed underpinning as the roots of the lime trees were drying out the underlying London clay and leading to subsidence. Continuous mass concrete footings were designed to be constructed to 3m below ground level. A contractor was appointed and the work completed in 1998. However in 2003 further cracking started to appear in the walls of the property, despite the underpinning. In recent years the cracking has become considerably more serious and at one stage there were concerns for the structural stability of the walls of the house. It was thought that the roots of the trees had surprisingly extended below the underside of the 3m deep footings and were therefore drying out the London clay at some considerable depth. However a further investigation revealed that the footings had not in fact been constructed to 3m as they should have but in some cases were around 1.5m or less. Clearly the contractor had not undertaken the work in the manner that he should have. CONCLUSIONS The four case studies have all involved failures related to the geotechnical situation of a variety of structures. In general it can be seen that the failures were rather surprising in the manner that they occurred; often several factors combined in

order to make the failure occur. It is not always obvious what has led to failure. Experience tells us that often it is at interfaces and in areas where risk is perceived to be low that problems occur. In forensic investigation of failure in civil engineering it is almost a reverse of “Occam’s razor” that needs to be applied (Fernie 2007); Occam’s razor is a principle that states that the explanation of any phenomenon should make as few assumptions as possible eliminating those that make no difference in the observed predictions of a hypothesis or theory. Clearly all possible risks should be investigated thoroughly before and during any construction. REFERENCES Fernie, R (2007) – Personal communication with the author Grose, W.J. and Benton, L (2005) – Proceedings of the Institution of Civil Engineers, Geotechnical Engineering 158, October, issue, GE4, 179-185 HSE (2000) – The Collapse of NATM Tunnels at Heathrow Airport Toll, A.S. and Sellmeijer, J.B. (2003) – Souterrain The Hague: Imperfections in Jet Grout Layers, Reclaiming the Underground Space, Saveur (ed.), 381-386

Chapter 11

Instrumentation and Monitoring for Forensic Geotechnical Engineering

Y.Iwasaki

Geo-Research Institute, Osaka, Japan

ABSTRACT

Instrumentation plays an important role in forensic geotechnical engineering. In soil mechanics, there are at least several reasons or mechanisms to explain a problematic phenomenon. The scientific facts that were recorded by instrumentation during the construction and/or post construction give a direction to select one of several possible hypotheses. Two case histories are presented to show the effects of instrumentation upon the forensic investigation of the failure of geotechnical constructions. One is Teton Dam, Idaho in the U.S.A. that was earthen dam and failed during the first filling stage in 1976. Someconclusions were given by Independent Panel to review the failure; however, Teton Dam has been providing endless long debatebecause of not enough monitored data to rely on. Another one is Nicoll Highway Collapse that was caused by failure of retaining wall during subway construction in Singapore in 2004. In this second case, monitored records of the deflection of retaining wall gave very important information. From the seventh excavation step, deflection of wall at south side became much larger than at north side. One sided deformation was due to the inclined dense gravel layer towards to south below the soft layer and deeper soft soil that differed from the initial design of horizontal layer condition. After the failure, it was confirmed thicker soft soil layer in the south compared to north side by boring study. If the retaining wall had failed without knowing different performance between these walls, it might be much works needed to reach the right conclusion. Instruments and sensors in geotechnical engineering are briefly reviewed. In recent decades, new technologies were introduced to sensor and instrumentation in geotechnical engineering. Among these innovative instruments, two techniques are introduced. Carrier Phase Tracking GPS for measuring displacement between reference GPS receiver and target GPS receiver with a high accuracy of displacement of 2-5mm. BOTDR (Brillouin Optical Fiber Reflectometer) is another cutting edge technology that provides strain and temperature. Technical basic knowledge of these instruments is explained to understand the principle that is useful not only to follow but also to take advantage of the next generation of instrumentation becoming to prevail at present and in near future.

Keywords: forensic, Teton dam, Nicoll Highway Collapse, Carrier Phase Tracking GPS, BOTDR 1 INTRODUCTION

In the preface to second edition of the textbook on Soil mechanics by Terzaghi and Peck in 1967, Prof. Ralph B. Peck added a new chapter on Performance Observation. This chapter 12 was intended to aid the engineer in the use of observational method which is at the very heart of successful application of soil mechanics. Among several good reasons to adapt observation, Dr. Peck indicates field observations for providing evidence in lawsuits. The text book indicates that lawsuits frequently arise from conflicts between the owner and contractor because of the completed structures or between the contractor and neighbour because of damage to the latter’s property. Firstly, two cases are presented where instrumentation was key factor in providing fact process of failure. Secondly, instrumentations and sensors are reviewed including some of the cutting edge technology that provides innovative instrumentation in geotechnical engineering at present and in near future.

2 LESSONS IN THE PAST

In recent cases, big disasters in geotechnical engineering show the need of the instrumentation and monitoring. Two examples are shown in the following section. One is Teton Dam that failed on June 5, 1976 and another is Nicoll Highway Collapse on April 21, 2004.

2.1 Teton Dam Failure

The Teton Dam was a federally built earthen dam on the Teton River in southeastern Idaho, USA which when filling for the first time suffered a catastrophic failure on June 5, 1976. The collapse of the dam resulted in the deaths of 11 people and 13,000 head of cattle. The dam cost about USD $100 million to build, and the federal government paid over $300 million in claims related to the dam failure. Total damage estimates have ranged up to $2 billion. The dam was never rebuilt. (Wikipedia) Not enough instrumentation was installed to monitor pore pressure in the dam.

Photo-1 Teton Dam, Idaho, June 5, 1976 from <http://en.wikipedia.org/wiki/Teton_Dam>

Hydraulic fracture in the core was considered as most likely the key mechanism estimated by desk work with laboratory aids.

Independent Panel to Review Cause of Teton Dam Failure (1976) reached a conclusion as follows,

The Panel had quickly identified piping as the most probable cause of the failure, and then focused its efforts on determining how the piping started. Two mechanisms were possible. The first was the flow of water under highly erodible and unprotected fill, through joints in unsealed rock beneath the grout cap, and development of an erosion tunnel. The second was “cracking caused by differential strains or hydraulic fracturing of the core material.” The Panel was unable to determine whether one or the other mechanism occurred, or a combination.

If pore pressure sensors had been installed in the dam section, process of increasing pore pressures during filling water might have provided vital information of the failure process.

Fig.1 Geotechnical condition for design of the failed section

Fig.2 Deflection of retaining wall of the failed section

The Panel referred on monitoring as “There were not enough instruments in the dam to provide

adequate information about changing conditions of the embankment and abutments.”

The debate on the cause of the failure still continues until recently (Solava and Delatte, 2003) and at present. If pore pressure monitored data had been provided, we could have reached the valuable understanding.

2.2 Collapse of Nicoll Highway

The Collapse of Nicoll Highway （ The Committee of Inquiry(2005)）that took place on 21 April 2004 was caused by an accident of subway construction in Singapore. The Failed section in the subway construction near Nicoll Highway was a cut and cover excavation shown in Fig.1. (Number with underline is displacement of JGP slab) The excavation is planned in a very deep soft marine clay layer

and some special design of JGP (Jet Grouting Pile) was considered. The slab near the bottom of the excavation was to take strong horizontal load from the soft clay layer. Among various instrumentations, the most important information was the inclination of the temporal retaining wall. Fig.3 Maximum deflection of walls of the south and north side Though under emergency, the measurement was not made daily. However, the obtained data gives clear tendency of the deformation as shown in fig.2. Two things should be noted.

One is the deflection of retaining wall. The other is the horizontal compression of JGP slab. It should be noticed not only the increament of the maximum deflection itself but also the difference between south and north walls. As shown in Fig.2 and 3, the maximum deflections at sixth and seventh excavation level show the same order of 160mm and 200mm for north and south sides. However, after the seventh excavation, the maximum deflection of the south side wall increased with the excavation depth. At the same time, the deflection of the north side was kept constant about 20cm and even decreasing. However, this difference of deflection needs some reasons of either weaker soil condition or smaller stiffness of retaining wall in south side. Fig.4 Deflection of walls at the final step and the confirmed geotechnical condition by boring after the failure

The original design condition was assumed as the same soil layers at both sides. Boring study after the failure have revealed that the dense sand gravel layer below the soft marine clay is inclined toward south and the soft marine layer is deposited deeper than the original design as shown in Fig. 4. If the site engineer could evaluate the different response of the walls of south and north, he might proceed to perform boring study and could prevent the failure. The site engineer just agreed with the proposal of modifying design parameter by the contractor and allowed to increase the allowable design level without any geotechnical consideration as shown in Fig.3. The JGP slab is rather stiff elements compared to the soft soil at the site. The horizontal compression displacement allowable in the slab might be controlled by failure strain level under compression of the order less than possibly a few percents. Since the horizontal length of the JGP slab is about 20m, 20 cm of the compression at 8th strut stage (Fig.2) corresponds to 1% of compression strain, which was the design allowalble strain. The instrumentation and monitored data really had provided not only before the failure but also after the failure to give constrains of the possible mechanism discussed in forensic engineering process.

3 INSTRUMENTATION FOR MONITORING

Recent technologies of monitoring displacements or strains have much advanced within the past decade. Among these several technologies, only two methods of GPS (Global Positioning System) and BOTDR (Brillouin Optical-fiber Time Domain Refletometer) are introduced here.

However, conventional methods are still important and used as daily practice.

Due to the limitation of the number of paper, the author will review only a trend of instrumentation for monitoring in geotechnical engineering in some aspect.

Reader should refer an introductory book as well as containing professional level by John Dunnicliff (1993). 3.1 Mechanical Sensor

Photo.2 shows a simple and yet accurate monitoring device to measure gap movement for foundation of Hudson-Athens Lighthouse (U.S. National Register of Historic Places) in the Hudson River in the state of New York in the United States.

Photo.2 Mechanical type of gap sensor

(http://en.wikipedia.org/wiki/Strain_gauge)

The two halves of the device are rigidly attached to the foundation wall on opposite sides of the crack. The red reference lines are on the transparent half and the grid is on the opaque white half. Both vertical and horizontal movement can

9/1 10/1 11/1 12/1 1/1 2/1 3/1 4/10

100

200

300

400

500

20042003

Wall Deflection(mm)

Time

South.I104

North.I65

Design Level

60

70

80

90

100

9/1 10/1 11/1 12/1 1/1 2/1 3/1 4/1GL=102.9m

Heigh

t(m)

lower JGP

upper JGP

Strut and Excavation Stages

Maximum Deflection of Retaining Walls

5/1

5/1

9th8th

6th7th

4th 5th

2nd1st strut

3rd

Earthpresure by soft claysouthside Earthpresure by soft claynorthside

be monitored over time. In this picture, the crack can be seen to have widened by approximately 0.3mm (and no vertical movement) since the gauge was installed.

3.2 Strain Gage Sensor

Electric sensor is also common for automatic monitoring. Photo.3 is strain gage based displacement sensor. The range of measurement is 5 to 10 mm with accuracy of 0.5-1%/Full Scale. Photo-3 Strain gage based displacement sensor (Tokyo Sokki)

Strain gage for steel plate was also developed based upon vibrating wire sensor as shown Photo-4.

Photo.4 VW spot-welded strain Gauge (Durham Geo Slope indicator Co.)

Vibrating wire transducer measures change of natural

frequency of a wire installed in a very small pipe on a steel plate. It is designed to set the sensor by spot-welding to steel surface.

The accuracy of the vibrating wire sensor is about 0.05% of full scale of 2500µstrain.

3.3 Displacement, Load cell and Pore pressure sensor

Conventional strain gauge and vibrating wire are basic transducer for not only strain but also other application of displacement, load, and pore pressure sensors. Other types of transducers like LVDT (Linear Voltage Differential Transducer) are also used for various types of sensors. The accuracy of LVDT sensor is around 0.1-.0.5%/F.S.

3.4 Inclinometer

Inclinometer sensor system that had been developed in 1970s was based upon force balance sensor. The force balance principle that is to measure how much force needed to get back the original zero position of the displacement sensor have resulted in the wider range to measure more than 30 degrees with the accuracy of 0.01-0.02%/FS. Recently, Micro-Electro-Mechanical Systems (MEMS) have realized much cost effective with the same performance

becomes available in the market. 4 RECENT DEVELOPMENT OF SENSORS

In the past decade, sensor technology had entered another era of innovative principles like MEMS (Micro Electro Mechanical Systems). In the chapter, among these cutting edge technology, GPS and fiber optic sensors are introduced.

4.1 Carrier-phase tracking of GPS Carrier-phase tracking signal of GPS has resulted in a revolution in land surveying. A line of sight along the ground is no longer necessary for precise positioning. Positions can be measured up to about 30 km from reference point without intermediate points. This use of GPS requires specially equipped carrier tracking receivers. Displacements of multiple points may be monitored by Carrier Phase Tracking GPS method that provides relative displacements of target monitoring points to the reference point.

(Difference in travel distance=n19+δ (cm))

Fig.5 Basic Concept of the Carrier Phase Tracking modified from (http://www.shamen-net.com/word/img/)

Photo 5 GPS receiver

(http://www.furuno.co.jp/product/gps/terrain/dana2000.html) As shown in Fig.5, distance between the monitoring station and

GPS is divided into two segments of equi-distance and the difference. The difference in travel distance is obtained as to detect the time difference between the receiving times of the same phase of the signal carrier signals. The L1 and/or L2 carrier signals are used in carrier phase surveying. The wave length used for the system is about 19cm of the L1 carrier band. The difference in travel distance = n19 + δcm) <where n is an integer number and called integer bias. With accuracy of 1% of wave length in detecting the leading edge, the error might be as low as 2 millimeters.

Fig.6 Determining wave number of integer bias for Carrier-phase tracking of GPS method

Fig.7 Scattering of monitored position by GPS carrier phase tracking (Iwasaki, 2008)

Table-1 Errors of displacement obtained by GPS Errors are caused by several factors. One of the common factors is the variation of wave velocity near the surface among the monitoring points. Wave velocity may be affected by such

space conditions like humidity, temperature, and air pressure. To avoid such errors, it is better to set monitoring position with relative position within some limited range of several hundred meters. The most time consuming process is to estimate the wave number n. Satellite is moving over the receivers from L1 to L2 along the known orbit. Using four GPS satellite signals, the phase difference as δ is obtained for every satellite. As shown in Fig.6, point P is the crossing point between ray path from satellite to the monitoring receiver and the equi-distance circle (Circle 1 in Fig.6). As the satellite moves from L1 to L2, the crossing point P1 moves to P2. Since the radius of the circle is very long compared with base line, the arc from P1 and P2 to the reference receiver (RR) is considered as a straight line. The locus of the points P1 and P2 forms an elliptical orbit. The wave number n is obtained as to give the minimum error to show the orbit as elliptical shape in 3D space.

Fig.8 Response of GPS in vertical component Fig.8 shows response of the vertical component against forced displacement that is shown in red bold line. It is found that the response of vertical component is not sharp to small displacement but becomes very clear when the input displacement is larger than the 5mm of the level of standard deviation.

4.2 Optical Fiber Sensors

Fiber optics, though used extensively in the modern world, is a fairly simple and old technology. Guiding of light by refraction, the principle that makes fiber optics possible, was first demonstrated in Paris in the early 1840s. Prof. Jun-ichi Nishizawa, a Japanese scientist at Tohoku University, was the first to propose the use of optical fiber for communications in 1963. Nishizawa invented other technologies that contributed to the development of optical fiber communications as well.

Photo.6 Optical fiber

Component Standard Deviation(mm)UD 5.02NS 2.06EW 1.86

5/10 5/15 5/20 5/25 5/30 6/4-10

0

10

20

30

40

Vertical displacement(mm)

forced displacement

-10

-5

0

5

10

0.0 10.0 20.0 30.0

-20

-10

0

10

20

0.0 10.0 20.0 30.0

-10

-5

0

5

10

0.0 10.0 20.0 30.0

2/1 2/8 2/15 2/22 2/29 3/7 3/14 3/21 3/28-20

-10

0

10

20 2008/2/1-3/26

2008/2/1-3/26

Displacement(mm)

NS

UD

2008/2/1-3/26

GPS Monitoring at G-3, Mikage

EW

2/1 2/8 2/15 2/22 2/29 3/7 3/14 3/21 3/28-10

-5

0

5

10

2/1 2/8 2/15 2/22 2/29 3/7 3/14 3/21 3/28-10

-5

0

5

10

(%)

(%)

(%)

Number in total 1252

EWav=0.31mm

SD=1.86mm

NSav=0.43mm

SD=2.06mm

UDav=0.017mm

SD=5.02mm

http:/ / en.wikipedia.org/ wiki/ Optical_fiber

Fig.9 Structure of optical fiber

Figure 10. Refractive Index (A and B: refraction, C: full reflection)

Optical fiber has been developed for sending images of

medical use and for optical signal as communication cable as shown in Photo. 9.

Fig.9 shows a typical structure of the optical fiber that composes from core and clad. An optical fiber consists of fiber of quartz or plastics in a cylindrical shape that works as waveguide for transmitting light along its axis, by the process of total internal reflection. The fiber consists of a core surrounded by a cladding layer. The refractive index (or index of refraction, see Fig.10) of a medium is a measure of how much the speed of light (or other waves such as sound waves) is reduced in the medium compared to a reference medium. In treating light, the reference medium is usually vacuum air.

Fig.12 Basic Concept on BOTDR�

The RI of 1.0 means light travel in the medium with the speed in vacuum air. The RI greater than 1.0 means light travel in the medium with lower speed than in vacuum air. To confine the optical signal in the core, the refractive index of the core must be greater than that of the cladding. If the wave length of the light is longer than compared to diameter of the core cable, the light is confined within the core and can travel very long distance with little scattering. The boundary between the core and cladding may either be abrupt change of refractive index, in step-index fiber, or gradual, in graded-index fiber. Single mode optical cable(SM) is designed as abrupt change to show little loss with capability of travelling long distance of 10km.

4.3 BOTDR (Brillouin Optical-fiber Time Domain Refletometer)

When a sharp and strong optical pulse is sent into an optical fiber cable, some weak optical signal was recognized as returned to the entered gate. The signals are originated from the scattering of the light at the propagating front. Three different kind of scattering are identified as Rayleigh, Raman, and Brillouin. The scattered signals of Brillouin are found to show special characteristics. The frequency of the scattered pulse is shifted about 10GHz from that of the input pulse and changes proportional to strain and temperature in the fiber as shown in Fig.11

�

<http://park.itc.u-tokyo.ac.jp/hotalab/research/bocda-r> Fig.11 Brillouin shift vs. Strains and Temperature

As shown in Fig.12, when a signal is sent to an optical fiber, backward travelling scattered light continuously arrives at the front gate. The point where the scattered signal originated is estimated by the delayed arrival time from the time of sending pulse. Since the relationship between the strain and Brillouin frequency shift is known, the analyzed frequency shift is easily transformed to strains or temperature. When temperature effect should be known independently, another optical fiber cable with free strain may be installed. The accuracy of the measurement depends upon accuracy to obtain the number of the frequency of Brillouin shift. At present, the measurement is usually given for 1m segment of the cable. The error range of strain is about 10-20µstrain and that of temperature is 0.5-1 degree in Celsius.

4.4 BOTDR as Pipe Strain Gauge

Nakano et al. (2003) reported a case history of the application of the BOTDR to identify sliding surface by inserting a pipe equipped with optical fiber in a slope as shown in Fig.13.

+

Fig.13 Monitoring slope instability by vertical pipe with

optic cable

Fig.14 Lateral deformation estimated from equivalent curvature with no axial strain (Mohamad (2007, 2009))

An optical fiber was fixed along a plastic pipe. The fiber

cable was given some pretension strain. The pipe strain gauge of optical fiber gives continuous change of the strain compared to conventional pipe strain measurement that gives only discrete points. As shown in Fig.13, the strain shows continuous change from compression to extension that corresponds to bending deformation at the depth of 10 to 11m. This depth was considered as a sliding surface.

((a) strains at pipe edges) ((b) strain in a loosed tube cable)

Mohamad et al.( 2009) Fig.15 Measured strains along edges and in a loosed tube BOTDR vs. conventional inclinometer Fig.16 Comparison of Lateral Displacement (Mohamad (2007, 2009))

-2 0 2 4 6 8 10 12 14-16

-14

-12

-10

-8

-6

-4

-2

0

2

4

6

Displacement (mm)

Ele

vation (m

)

Exc.Lvl.

BaseSlab Floor

1st Prop

2nd Prop3rd Prop

(26 Oct)

BOTDR8 Nov 05Incl 3 Nov 05Incl 21 Nov 05Incl Averaged12 Nov 05

(b)

-0.04 -0.03 -0.02 -0.01 0 0.01 0.02-16

-14

-12

-10

-8

-6

-4

-2

0

2

4

Strain (%)

Ele

vati

on

(m

)

Pile 126: BOTDR Strain Measurementbetween 12/08/05 and 8/11/05

Exc.Lvl.

Outersideof ExcavationNearsideof ExcavationAxial strain

(a)

∆T [°C]

4.5 BOTDR as Inclinometerequiped with Secant Pile Wall

Mohamad et al.(2007, 2009) show some BOTDR application

to monitor axial strain of bored piles and lateral deflection of retaining wall of secant pile wall of a series of intersecting reinforced concrete piles.

The diameter of the secant wall was 450mm and a pair of optical fiber cable was installed along the reinforced steel was installed along steel cage giving a pretension of 2000µstrain. Another single optic fiber loose tubed cable was equipped along the monitored pile. Fig.14 shows how to evaluate lateral deformation from the edge strain along a pile. Since the edge strains are induced by mix modes of axial compression and bending deformation. To obtain bending strains, axial compression should be separated. These axial components should be subtracted from the apparent edge strains as shown in Fig.14.

Temperature is another factor that affects upon the apparent strain values.

Fig.15 shows an example of the measured strains of BOTDR for retaining wall after excavation. The excavated level was about -4.5m. Fig.15 (a) shows two edge strains of outside and nearside of the pipe as well as axial strain of the averaged values.

Fig.15 (b) shows rather constant temperature in the soil up to a level of -6.5m and gradually decreases to -13 degree at the top of the pipe.

The air temperature was lower than the in the ground, temperature decreases from near the ground surface to the top of the pile.

Fig.16 shows comparison between a lateral deformation curve obtained from BOTDR and two curves by conventional inclinometer measurements. It is noticed that the averaged curve of the two deflection curves from the conventional inclinometer corresponds well to the one from BOTDR.

Compare to the conventional inclinometer of manual handling, inclinometer system by BOTDR provides automatic and realtime evaluation of deformation of retaining wall.

Inclinometer measures inclination angle relative to gravity. BOTDR measures a pair of the edge strains under bending mode. It should be notice that the distance between two edges must be wide enough to keep acceptable accuracy of bending deformation.

At present, the facility to send and receive the signal as well as analyzing the BOTDR is rather expensive, In the near future, the cost effective system is expected available and become useful especially for big and deep excavation works. 5 CONCLUSION

In the first part, two comparative case studies are overviewed. Teton Dam that was failed by piping. No further detailed mechanism is clarified due to not enough instrumentation. In the collapse of Nicoll Highway, though the data were not collected as it should be the available data of inclination of the subway construction nearby the Highway show clear fact of one sided deformation of the wall from which the geotechnical condition for the design was anticipated different from the design condition. If the construction had been modified, the collapse might well be avoided. After the failure, the results of instrumentation contributed to identify the process of to the failure from the geotechnical point of view. In the second part, review of sensors is made including some edge cutting technologies of two sensors of Carrier-phase tracking of GPS method and BOTDR. These sensing technologies in addition with other sensor like MEM (Micro-machine Electronic Machine) sensor are still on a way of development. Forensic engineers, especially in the field of geotechnical field

where most of the phenomena takes place in dark underground, should use and take advantages of installation and monitored data to reach the most likely mechanism among several possibilities.

6 REFERENCES

The Committee of Inquiry, 2005,”Report on the Incident at the MRT Circleline Worksite that led to the Collapse of the Nicoll Highway on 20 April 2004,” Ministry of Manpower, Government of Singapore

Dunnicliff J. 1988. “Geotechnical Instrumentation for Monitoring Field Performance,” John Wiley & Sons, New York (USA)

Iwasaki, Y. 2008, “Basic Study of Carrier-phase tracking of GPS method at Mikage site,” GRI report

Independent Panel to Review Cause of Teton Dam Failure, 1976, “Report to the U.S. Department of the Interior and State of Idaho on Failure of Teton Dam. Idaho Falls,” Idaho. December 1976

Nakano, M., Yamazaki, H., and Okuno, M., 2003, “The earthquake disaster prevention monitoring system using optical fiber technology,” JSCE Journal of Earthquake Engineering, Vol.27, pp.1-4

Mohamad, H., Bennett, P.J., Soga, K., Klar, A. & Pellow, A.2007.” Distributed Optical Fibre Strain Sensing in a Secant Piled Wall,” Proceedings of the Seventh International Symposium on Field Measurements in Geomechanics (FMGM2007), ASCE Geotechnical Special Publication, No. 175

Mohamad, H., Soga,K., & Bennett, P. J.,2009, “Fibre optic installation techniques for pile instrumentation,” Proc.,17th ICSMGE(in print)

Solava and Delatte, 2003, “Lessons from the Failure of the Teton Dam,” Proceedings of the 3rd ASCE Forensics Congress, October 19 - 21, 2003, San Diego, California

Chapter 12

LEGAL PROCESS AND JURISPRUDENCE

Dhirendra S. Saxena (Sax), P.E., F NAFE Senior Principal and Chief Consultant, ASC geosciences, inc., Lakeland, Florida, USA

9.1 INTRODUCTION For geotechnical engineers it is only a myth to believe that practicing perfect engineering, or conforming to normal standards of care, will provide immunity from civil liability. Unfortunately, when problems or failures occur, all parties including engineers get named in the lawsuit regardless of their innocence. In USA, attorneys sue everyone involved with a damaged project to secure compensation or claim to their liability insurance limit. A practicing geotechnical engineer cannot provide services without the fear of a lawsuit. Strategies for limiting liability range from assessing risk to securing professional liability insurance and including limitation of liability classes in consulting contracts between engineer and the client. Although experts retained by opposing parties generally disagree on issues resulting from differences in professional judgment, they are invaluable to the jurisprudence system in America. A new discipline known as forensic geotechnical engineering (FGE) has been created to deal with investigation of soil-interaction related failures of engineered facilities or structures. Services of forensic geotechnical engineers, experienced in the jurisprudence system, are generally commissioned to investigate such failures. They also prepare reports, and sometimes provide expert witness testimony, in an attempt to assist their client in understanding the cause and liability for the failure. Because of the adversarial nature of most failure investigations many forensic engineers are generally under pressure to give an opinion that benefits their client’s position. Accordingly, the technical facts are not presented, yielding to this kind of pressure sometime leads to inappropriate occurrences, practicing outside one’s competence, manipulation of facts, crafted testimony, inadequate or defective investigations, misrepresentation of standards of practice, or other ethical violations. A forensic geotechnical engineer looks at a pile of rubble as a problem that needs to be solved. Each broken piece in the pile tells a story about the forces that acted upon it and somewhere in the twisted mass of intertwined building components is the first facture surface that resulted in the progressive collapse of the structure. At the end of the failure analysis process, which may include laboratory testing of critical items and hours of analysis in the office, a clear picture of what happened begins to develop, and the forensic engineer must then be able to explain the mode of failure and the forces, action, or process that caused the failure. The explanation may be a simple letter, a detailed report, or expert testimony in the courtroom (Bell, 2007). “What happened?”, and “Why did it happen?” are usually the first two questions asked of the forensic engineer. Of course, these are often followed by, “How can it be fixed?”, and all too often, “Who’s fault is it?”, and “Who is going to pay?” (Bell, 2007) When something breaks, falls down, or otherwise fails to perform as it was intended, our society demands to know why. This inevitably becomes the job of the forensic engineer, who must be equally comfortable

2

in the field, the office, and the courtroom while providing an objective analysis of the failure, whether the facts are helpful or harmful to the client’s interests. FGE prepares civil engineers to read, think, speak, and analyze like a lawyer. In addition, it familiarizes him with jurisprudence system so that he is much able to understand and deal with legal issues since he has to work closely with statutes and regulations may become involved in litigation, or who may serve as an expert witness. 9.1.1 Who Wants to Know What A forensic geotechnical engineer needs to answer questions, such as:

• The building owner who wants to know why the building fell down and who is

going to pay for it.

• The insurance adjuster who needs to determine if the cause of the collapse is covered under the building owner’s insurance policy, and if it was the result of a defect that someone else might be liable for.

• The contractor and subcontractors, who built the building and their liability

carriers, want to know if the defect is construction-related and if so, which trade or trades are responsible.

• The design professionals and their liability carriers want to know if the

defect is design-related. Each of these entities may retain their own forensic engineers to assist them with understanding how they may or may not be responsible. There will be attorneys for each entity who will evaluate and argue how the law should be applied to the insurance policies and contracts. The forensic engineer’s opinion of what happened may be explored during a deposition in which the attorneys for all sides have the opportunity to probe the engineer as to the process used and the basis for the opinion. If the parties cannot reach a settlement, the forensic engineer may be called to provide expert testimony in the courtroom, with the goal of helping the judge and/or jury understand the facts and how they relate to the particular incident. Opposing attorneys normally attempt to discredit the engineer through cross-examination with the hopes of revealing some flaw in the forensic engineer’s thought process. The forensic engineer must be able to apply the art and science of engineering during the entire process, from the extreme conditions encountered in the field, through the intense pressure experienced in the courtroom, and never lose sight that seeking the truth is all that matters. By understanding what went wrong, the forensic engineer is able to add to the body of knowledge that hopefully will prevent a similar occurrence from happening again. 9.2 TECHNICAL ISSUES Within the framework of dispute relating to construction, particularly within the jurisprudence system, forensic engineering may include investigation of the physical causes of accidents and other sources of claims and litigation, preparation of engineering reports, testimony at hearings

3

and trials in administrative or judicial proceedings, and the rendition of advisory opinions to assist in the resolution of disputes affecting life or property. Within the construction industry, disputes usually result from unfulfilled expectations of the project developers and purchasers, from errors or omissions of the design team, or from construction or material defects. Non-technical issues, principally time and cost, are frequently considerations and unanticipated influences such as weather and unforeseen conditions that may negatively influence the performance of construction. In the “perfectly” constructed project, the expectations of the owner/developer are completely understood by the design team and are converted to a clear, comprehensive set of drawings and specifications. After execution of an equitable contract outlining the terms and conditions, schedule, and compensation arrangements, the competent construction team executes the construction according to the design within budget and on schedule. And finally, the expectations of the owner/developer/occupant are met. Unfortunately, in our imperfect world, few projects are perfect. Expectations are not met, designs are inadequate, construction is improper, materials fail, and nature exerts its influence. Whether before, or after calling the lawyers, a forensic engineer is usually needed to provide the technical expertise to investigate and answer the questions. The process used by the forensic engineer is an adaptation of the scientific method, “the process by which scientists attempt to construct an accurate (that is, reliable, consistent and non-arbitrary) representation of the world.” The scientific method has four steps:

1. Observation and description of a phenomenon (an occurrence, observation, event, or trend) or group of phenomena i.e. identify the problem.

2. Formulation of a hypothesis (a theory or statement regarding cause and effect) to explain the phenomena, i.e. decide on the procedure.

3. Use of the hypothesis to predict quantitatively the results of new observations. 4. Performance of experimental tests of the predictions with properly performed

experiments, i.e. collect the data and perform the analysis. It is the fourth step that often is overlooked. Because the experimental tests may lend either to confirmation of the hypothesis or to the ruling out of the hypothesis, the scientific method requires that a hypothesis be ruled out or modified if its predictions are clearly and repeatedly incompatible with experimental tests. Further, no matter how elegant or desirable a hypothesis or theory is its predictions must agree with experimental results if it is a valid description of nature. The scientific method attempts to minimize the influence of the forensic engineer’s bias or prejudice on the outcome of an experiment. That is, when testing a theory, the engineer may have a preference for one outcome or another, and it is important that this preference not influence the results or their interpretation. Another common mistake is to ignore or rule out data (measurements or information) which do not support the engineer’s theory. Ideally, the engineer is open to the possibility that the theory is correct or incorrect. Sometimes, however, the engineer may have a strong belief that the hypothesis is true (or false), or feels internal or external pressure to get a specific result. In that case, there may be a psychological tendency to find “something wrong” with data which do not support the engineer’s expectations, while data which do agree with those expectations may not be checked as carefully. 9.3 ETHICAL ISSUES

4

Ethics is defined as a discipline dealing with good and bad, right and wrong, moral duty and obligation. Why does a group need a code of ethics? One does not have to look any further than the daily newspaper to find that unethical behavior exists in the highest levels of government, in the clergy and in virtually every segment of society. Social scientists have studied and found that in any group a certain percentage will deviate from the norm. Therefore, there is the need to codify and regulate or influence the behavior of various professions and groups. But, when we speak of a code of ethics applicable to engineers do we mean merely the business practice with respect to another professional or a client, or are there higher goals and standards to which a professional engineer should aspire? The NSPE code of ethics states that in the fulfillment of their professional duties engineers shall:

“1. Hold paramount the safety, health and welfare of the public in the performance of their professional duties; 2. Perform services only in the areas of their competence; 3. Issue public statements only in an objective and truthful manner; 4. Act in professional matters for each employer or client as faithful agents or trustees; 5. Avoid deceptive acts in the solicitation of professional employment.”

Upon close examination of the Canons, it is seen that none are in conflict with the practice of forensic engineering. The forensic engineer should adhere to the SPE Cannons not only as a member of NSPE but also as an engineer (Dixon, 1992). Engineers are recognized as the most ethical professionals above doctors, accountants, and lawyers in a survey of 200 CEO’s of national companies. While ethical core values for engineers are generally identified as integrity, honesty, faithfulness, charity, responsibility, and self discipline the classical seven deadly sins are described as pride, greed, anger, envy, gluttony, lechery, and sloth. It is clear that there are problems in ensuring the integrity of expert testimony. There are no minimum qualifications for becoming an expert witness, nor any means of regulating them. Code of ethics are voluntary and not often enforced (Grover et al., 2003). In summary, the middle of the road rule for an ethical engineer is to be honest, tell the truth no matter how bad it may seem, and let your conscience be your guide (Baker, 1997). 9.4 LEGAL ISSUES The legal considerations of these forensic geotechnical engineering services illustrate the reality that the engineering investigation of a failure incident is a fact-finding mission that results in uncovering the probable causes of that failure. It concentrates on the identification of hidden clues. The procedures adopted for the analysis, testing, opinions, and written reports should be able to satisfy even legal scrutiny of their validity. 9.4.1 Role of Expert Witness Experts are vitally important to universal jurisprudence. Expert witness can assist retained counsel in fact gathering, request for documents, and evaluation of evidence; fashioning precise questions for opposing experts and through candid discussions of strengths and weaknesses of the case; retained counsel to properly qualify and question an expert witness to maximize his

5

impact on the jury; and expansion of the traditional testifying roles in the litigation process. An expert in forensic engineering is not automatically dubbed an “expert witness”. An individual is not granted the privileges of an expert witness until a legal forum confers that recognition upon a properly qualified professional. Experts must possess specialized knowledge of science, profession, business or occupation that is beyond the experience of the lay juror and can assist the fact finder (Grover, 2003). Additionally, expert witnesses in the field of engineering are governed by their own code of ethics. Well over 90% of civil cases settle prior to expert witnesses being called for a trial. As such experts can expect that the majority of testimony that they give will be given at deposition. Accordingly, the experts need to excel during their depositions as they are key element of the discovery process. If the expert witness is a litigation consultant, he/she is not subject to discovery by the opposing counsel. If disclosed as an expert, then he/she is subject to discovery by the opposing counsel. Engineers should perform services only in areas of their competence and they should undertake assignments when qualified by education and/or experience in the specific fields involved. Direct examination is an expert’s opportunity to persuade the jury that his/her opinion should be believed. The goal during direct testimony is to persuade the jury to find one’s opinion creditable. To connect with the jury an expert needs to understand what jurors want and employ the best ways to communicate with them. Experienced expert prepares a well written, persuasive expert based upon reliable accepted methodology. Expert witness can express his/her opinion in an unambiguous and legally sufficient manner. Expert witness can dramatically improve their performance by being good teachers, both to the lawyers who hire them, and to jurors and other fact- finders. They must keep it simple with illustrations, explanation and non-technical language. The four key elements of a successful expert witness assignment are (SEAK, 2006): 9.4.1.1 Part I: Prevention The most valuable experts isolate themselves from attacks and deny opposing counsel ammunition to attack their credentials and credibility. It helps to identify a detailed checklist of potential areas of attack that experts may be subject to regarding their credentials and credibility including: every word on their CVs, past testimony, their image, controversial or political associations, missing credentials, fee schedules, fee agreements, marketing materials, web page, speeches, work on past case, apparent and actual conflicts of interest, non-related litigation, hobbies, professional complaints or discipline, presentations and writings. The best experts form and express airtight opinions and opinions that hold up under the most rigorous scrutiny and cross-examination. There are many ways in which opposing counsel is able to poke holes in an expert’s opinion. The expert must provide specific actions to bullet-proof his opinion including, proper case and client selection, avoiding time crunches, using careful and confident language, not overstating or understand facts or opinions, consistency, dealing with the opinions of other experts, knowing exactly what needs to be proved, testing alternative theories, taking careful and precise measurements, being well-trained and well-versed in any computer program used, verifying computer results, leaving no stone unturned, taking photographs, verifying factual assumption, gaining as much first hand

6

knowledge as possible, thoroughly researching the issues at hand, obtaining and carefully reviewing all relevant documents, not sharing draft reports with counsel, and avoiding “junk science”. 9.4.1.2 Part II: Preparation Peak performance requires proper disciplined preparation done correctly prior to deposition and trial. Well-prepared experts are able to deliver confident testimony, deal with cross-examination far more effectively, and are in much better position to articulate and defend their opinions. It should include advanced techniques that can and should be used to prepare for depositions, direct examination and cross-examination. 9.4.1.3 Part III: Performance The best experts recognize that most cases are won and lost in the discovery phase and that the expert’s deposition is a crucial, often outcome determinative, component of the case. In order to excel at the highest level during a deposition, experts need to be able to recognize and defeat opposing counsel’s deposition tactics and recognize how these tactics differ from those used during trial. The best experts deliver powerful and understandable direct testimony. They also learn to explain and demonstrate numerous advanced techniques for delivering captivating, memorable and persuasive direct expert testimony. These advance techniques include, showing – not telling, getting to the point up front and explaining later, being well-prepared and well-organized, making the complex simple, entertaining, being likeable, highlighting your most relevant qualifications, working on a smooth flow and style, getting out of the jury box early and often, using visual aids that work, aggressively self-editing, employing powerful, memorable analogies, showing your human side and bonding with the jury, using precise language, using confident language, employing short preview and review summaries, using numbered lists, citing references, speaking conversationally, conforming your testimony to the theme of the case, and reading and reacting to the jury. 9.4.1.4 Part IV: Practice The best experts skillfully answer cross-examination questions and explain how the question could have been avoided, how they could have and should have prepared to answer the question, identifying the tactic that counsel is using and delivering a response that defeats the tactic and/or allows the expert to go on the offensive. 9.4.2 Legal Considerations In order to fully evaluate the forensic expert’s role in the legal process, the expert should understand the definitions and concepts of a dispute resolution. The professional services provided in the investigative process and the duties expected of the expert witness are generally the same no matter what method of formal dispute resolution is selected by the client. Following is an overview of the various methods of dispute resolution that may be encountered (Task Committee on Guidelines for Failure Investigation, 1989). 9.4.2.1 Civil Litigation This process occurs in a court of justice having jurisdiction in the dispute. The court is presided over by a judge who decides on the applications of the law and procedure. These

7

decisions usually are determined by a jury of lay-persons or by the judge in the absence of a jury. The decision of this court and jury may be subjected to appeals and retrials. 9.4.2.2 Arbitration Instead of a judge and/or jury, an arbitrator or a panel of arbitrators are used in a private proceeding. Unlike litigation, arbitration eliminates or significantly relaxes the discovery process and is usually limited to the production of documents to an extent determined by the arbitrators. 9.4.2.3 Administrative Hearings or Special Investigations Hearings are held by governmental authorities to resolve disputes or to gather information of public opinion prior to making an administrative decision, or for inquiries into the integrity of the structure or the cause of a failure of an engineering facility. 9.4.2.4 Private Litigation This form of dispute resolution requires both parties to agree to resolve their differences in a private court that is created, and paid for, by both parties. A mutually respected individual, such as a retired jurist, is chosen as a judge, and civil litigation rules are followed. 9.4.2.5 Mediation This method of “non binding” dispute resolution is used to resolve certain construction-oriented disputes. A mediator acceptable to both sides is selected and probes the disputants to determine their position relative to “minimum” settlement. He or she then strives to attain a middle ground that both can accept comfortably. 9.5 OVERVIEW A well organized, methodical approach to failure investigation is needed because of the complexities of modern engineered facilities and the vast variety of possible failure causes. The investigative process should not employ a rigid “cookbook” approach, but there are certain steps that are common to an effective failure investigation. An emphasis on an open-minded acceptance of all pertinent data, while recognizing the relevant information, is the fundamental basis of a failure investigation. As various hypotheses develop and change, continuous review of previously rejected data is required. There is seldom a single cause of a failure but rather a complex interaction of components and forces. As a result, the outcome of a failure investigation seldom leads to absolutely irrefutable results but rather to a most probably cause of failure. While the collection of data and facts surrounding the failure are specific, expert opinions arising from these data may differ widely. The most accepted failure investigation findings will be the one employing a qualified investigation team that presents the most plausible failure scenario, based on well documented support data. (Task Committee on Guidelines for Failure Investigation, 1989). A close coordination between the forensic geotechnical engineer, his client, and client’s

8

attorney is important in order to understand the facts as they relate to a particular incident. This teamwork normally lends credence to their case and is instrumental in making it a winning combination. Two case histories that include all elements of forensic geotechnical engineering as well as legal issues of jurisprudence system are included with this chapter. They include situations where forensic geotechnical engineering was effectively utilized to identify, investigate, remediate, litigate, and resolute legal issues. 9.6 REFERENCES ASC geosciences, inc. (2002) – Report of Field Investigation, Data Evaluation, and Engineering Consultation Services, ASC geosciences, inc., Report No. 02L2016. Baker, P.E., John A. (1997) – Professional Engineering Ethics versus Expert Witness Requirements, Los Angeles Section ASCE, Forensic Engineering Technical Group. Bell, P.E., John T. (2007) – What is Forensic Engineering?, Florida Engineering Society Journal, May 2007, pp.8-9. Dixon, E. Joyce (1992) – The NSPE Code of Ethics Applied to Forensic Engineering, Journal of the National Academy of Forensic Engineers, Vol. IX No. 1. Grover, J.D., P.E., J.L. (2003) Ethical Considerations for Expert Witnesses in Forensic Engineering, Ethical Dilemmas of Technical Forensic Practice, pp. 441-52. Saxena, D.S., P.E. (2005) – "Forensic (Geo-technical and Foundation) Engineering Case History", National Academy of Forensic Engineers (NAFE) Seminar, Chicago, Illinois, 10 July 2005. Saxena, D.S., P.E. (2007) – "Case Studies in Forensic Geotechnical and Foundation Engineering", 6th International Conference on Case Histories in Geotechnical Engineering, Arlington, VA, Paper No. OSP-2. SEAK inc. (2006) – Advanced Testifying Skills for Experts – The Master’s Program, pp. 4-11. Task Committee on Guidelines for Failure Investigation (1989) – Guidelines for Failure Investigation, American Society of Civil Engineer (ASCE). B. PART 2 9.7 Appendices 9.7.1 Case History One Project specifics and forensic facts summary;

• The backyard slope of a one-story dwelling subsided abruptly, failed, and

9

slid into the lake along with the rip-rap from lake edge. It also extended into the neighboring property on the south side.

• The lakefront backyard of a residence was damaged and stability of the

structure was threatened, and removal/replacement of the completed structure was considered a viable yet costly option.

• The owner filed a claim against the developer/builder for negligence and for

not informing them of a potentially unstable pre-existing condition.

• The owner secured the services of a Forensic Geotechnical Engineer (FGE) to investigate, remediate, and assist in litigation as well as serve as an expert witness during the resolution of dispute and legal issues.

• The developer/builder offered a band- aid solution that the owner then

rejected upon advice of the FGE. Figures 1-4 illustrate the salient features of this case history.

Fig. 1 Close-up view of the subsided backyard

10

Fig. 2 View of subsided slope of adjoining house

Fig. 3 View of sodded and finished backyard

11

Fig. 4 Reoccurrence of slope failure in finished

backyard two weeks after completion 9.7.1.1 Forensic Field Exploration and Subsurface Condition Evaluation Subsurface stratigraphy beneath the site, as illustrated in Figure 7, consisted of very soft, loose low strength subsoils to 15.0 ft depth. These low strength subsoils consisting of highly plastic clays (mc 113%, LL-107, PI-89) which were unstable under the weight of the additional fill material. Slope stability analyses were performed for the original subsurface profile as well as the contractor/developer proposed restructured slope that yielded Factors of Safety (F.S.) of 0.87 and 0.93, respectively, for the two conditions. 9.7.1.2 Observations and Findings The expert’s investigation at the project development site revealed that;

• Slope failure was triggered by the presence of unconsolidated sediment layers under the site fill.

• Unstable sediments to continue to consolidate resulting in continued

movement of the back of the house and the backyard.

• A retaining wall should have been constructed prior to the house being built to contain and stabilize the soils beneath the house foundation.

12

Fig. 7 Subsurface stratigraphy identifying delineation of compressible phosphatic clay

Upon the advice of their legal counsel the owner put the developer on notice, and elected to proceed with the proper fix pending resolution of litigation and the claim.

9.7.1.3 Remediation Based on findings and observations from the FGE investigation, a two part remediation program was recommended and consisted of:

1. Installation of helical anchor piers under the exterior wall footing, along the back of

the house, for the underpinning stabilization. 2. Installation of a helical anchor bulkhead to stabilize the backyard along the lake. 3. Based upon field monitoring of pier installation operations, anchored timber

bulkhead, post stabilization walk through, and review/evaluation of the field stabilization data, it was concluded that the methods and procedures utilized and installed by the specialty contractor were effective, satisfactory, and acceptable. Slope stability analyses of this remediated structure indicated a Factor of Safety (F.S.) of 1.57.

Figures 6-9 illustrate various sequences of the underpinning and bulkhead stabilization program.

13

Fig. 6 View of support cap and bracket assembly over

helical support anchor under the footing

Fig. 7 Close-up of geotextile lay down and wrap up

at helical anchor tie back on timber bulkhead

14

Fig. 8 Additional fill being dozed into the anchored geotextile

Fig. 9 View of completed and restored backyard

9.7.1.4 Concluding Remarks

1. Following the close coordination between the forensic geotechnical engineer, the owner, and the remediation contractor and based upon results of remediation monitoring it was determined that the repair of the residence structure had been achieved satisfactorily.

2. The final repair resulted in restoration of the residential structure and backyard to its

originally planned and constructed stage, as illustrated in Figure 9. 3. As a result of the monitored and satisfactory remediation program, the consultant

recommended acceptance of the restored structure. 4. The property owner retained an attorney who sent a notice of claim, and remediation

15

cost summary, to the developer/builder. Following litigation, arbitration the property owner, through their legal counsel, was able to recover these remediation costs along with attorney fees.

9.7.2 Case History Two This case history identifies a request received from a building owner for a forensic engineering review to investigate a situation where a severe site specific soil-structure deficiency occurred and caused post construction damage to an office building in west central Florida. Evaluations were also made to see if repairing or replacement of the structure was, in fact, necessary. 9.7.2.1 Project Description Construction of an office building was completed in June 2006 as per design plans, permitting, and pre-construction technical support from the civil engineer. Foundation support included installation of 12 inch diameter timber piling, driven to 24 tons capacity (as per the pile driving records prepared by a geotechnical engineering representative at the site) along the exterior perimeter grade beam supporting the load bearing walls. The interior slab was a 4 inch thick fiber mesh concrete supported on compacted soil for the entire building footprint. Soon after occupying the building, the owner(s) observed certain impacts on their property. Concerns identified by the owner(s) had included:

• noticeable settlement and deflection of the interior soil supported floor slab.

• visible cracks throughout the interior of the building. • differential settlement of building foundations (exterior piles and interior

floor slab) resulting in misalignment of the doors. • possible damage to the below grade utility lines.

• substantial deflection distress and development of cracks throughout the

interior and foundation floor slab within the office building as a consequence of differential settlement due to the consolidation of existing peat material within the building footprint area. Additionally, total and differential settlement of the foundation support system and resulting distress appeared to be of a continuing and progressive nature.

Figures 1 – 8 illustrate the salient features of the case history and owner’s justifiable concern.

16

Fig. 1 Two close up views of sinking floor at conference room door

Fig. 2 Close up view of interior partition wall separation from sinking floor slab

Fig. 4 Close up view of wall crack showing new crack and additional movement

17

Fig. 8 Separation and sinking of slab along building exterior

A 2-stage program was carried out by the forensic geotechnical consultant in conjunction with filing of a claim by the building owner. 9.7.2.2 Stage 1 Investigation Stage 1 consisted of a review of the field inspection of the structure, review of the project foundation drawings, and pre-design geotechnical exploration report. The findings, comments, and conclusions derived from the Stage 1 investigation at the project site revealed that:

• there were wide spread separations between the walls and the floors, and between partition walls and ceilings. The separations appeared to be greatest toward the center of the building.

• a floor elevation survey of the building floor slab conducted in March

2007 indicated partial settlement of 7.5 cms (3 inches) at the center of the building, as illustrated in Fig. 9.

• subsoil test borings performed by the pre-design geotechnical engineer

were 12.0 m (40 ft) deep and consisted of organic compressive soils. Pressure treated timber piles with an embedment depth of 20 to 25 ft (for a 12 inch diameter) were utilized and a net allowable capacity of 12 tons was recommended.

• during construction pile lengths of 16.0 m to 18.0 m (50 to 55 ft) were

used. Furthermore, 30.0 cm (12 inch) square prestressed precast concrete (PCC) piles were substituted in lieu of pressure treated timber piles. In addition, the contractor elected to use an unusually high energy hammer in lieu of the recommended hammer of low to moderate energy for the timber pile.

18

• All three of the borings had significant weight-of-rod (WOR) and weight-

of-hammer (WOH) zones throughout the in filled soils. In those cases where underlying sandy soils were encountered, it appeared that the deep sands were in a more stable condition. This profile indicated an ongoing raveling associated with the relic sinkhole, or an indication that the in filled organic soils were weak and very loose.

• the interior floor slab was only soil supported and not structural (with

interior grade beam supported). • proper pile capacity determination, its installation, and structural grade

beam supported floor slab was not performed. 9.7.2.3 Stage 2 Investigation In an effort to properly address, investigate, and evaluate the alleged subsoil deficiency a detailed subsoil exploration program was undertaken by advancing test borings to depths ranging from 40.0 m (130 ft) to 60.0 m (190 ft). Test boring stratigraphy showing internal erosion and solution features is illustrated in Fig. 5. It consisted of elements listed below.

• the site was underlain by very poor soil conditions not capable of providing support for any type of construction that assumes bearing to be developed from the underlying soils.

• upper fill soils were mixed with debris underlain by predominately

organic soils with organic content ranging from 5% (at 80 ft depth) to 18.5% (at 60 ft depth) and 85% (at 135 ft depth).

• no limestone was encountered to the termination depth of 43.0 m (135 ft)

to 60.0 m (190 ft). Project site stratigraphy showing internal erosion and solution features is illustrated in Fig. 5.

19

Fig. 5 Project site stratigraphy showing internal erosion and solution features

• damage to the building was the result of factors including consolidation

of significant factors including consolidation of significant depth of very loose organic soils (located within a large relic sinkhole) along with the consolidation of buried debris located within or just below the develop fill soils and loading street zone.

• the old aerial photographs from 1941 through 2005 showed the wetland

feature and confirmed that the building was constructed in a wetland drainage basin area over deposits of very deep, very loose organic soils.

9.7.2.4 Concluding Remarks

• the owner reported that the damage had been ongoing since the building was constructed, which was consistent with a concrete slab supported by very loose, compressible, debris laden organic soils. The pile supported perimeter walls supported by the timber pile showed very little damage, indicating that no significant differential movement of the perimeter wall had yet occurred.

• it was determined that deleterious soil condition was known to exist prior

to construction of the building and is the reason that a pile foundation was recommended by the project geotechnical engineer. However, the foundation design only accounted for support of the perimeter walls and did not address the need for support of the slab. It was concluded that this design flaw/oversight was the primary reason that the building had sustained substantial interior damage.

• it was also concluded that no feasible and effective options were

20

available to remedy the relic sinkhole condition and that the wetland site over deposit of very deep organic soil was not a suitable site for this building construction.

• additionally, as a direct result of structural and geotechnical engineer’s

severe design deficiency, as well as contractor’s failure to exercise, due care the building had sustained substantial irreversible damage.

• the property owner’s claim to the developer/builder was successful and

the owner was able to recover a major portion of the damage cost resulting from design oversight and relic sinkhole related damage.

1

Chapter 13.

POTENTIAL ROLE OF RELIABILITY IN FORENSIC GEOTECHNICAL ENGINEERING

K K Phoon Department of Civil Engineering, National University of Singapore, Singapore 117576

E-mail cvepkk@nus.edu.sg

G L Sivakumar Babu Department of Civil Engineering Indian Institute of Science, Bangalore, India, 560012

E-mail: gls@civil.iisc.ernet.in

M Uzielli Georisk Engineering S.r.l., Florence, Italy

E-mail : muz@georisk.eu

ABSTRACT This chapter discusses some possible roles for reliability and risk in forensic geotechnical engineering. A preliminary statistical framework is presented to quantify the difference between expected and observed performance in the presence of unavoidable and potentially significant geotechnical variabilities. The probable ranges of geotechnical variabilities are reviewed. Other potentially useful results in the recent reliability and risk literature are highlighted. The intention of this chapter is to stimulate further discussions and research in this important but somewhat overlooked area.

2

1. INTRODUCTION Forensic geotechnical engineering (FGE) is a relatively new field. ISSMGE TC40 was formed only in 2005 under Dr VVS Rao as chair. Subcommittee 6 was convened under Dr KK Phoon to explore the potential role of reliability in FGE. A cursory Google search using keywords “reliability, forensic, and geotechnical” shows that this concept is entirely new. This chapter seeks to explore if reliability/risk concepts are potentially useful to forensic geotechnical engineering. This study is very preliminary given the dearth of previous literature. Brown (2006) noted that: “The concept of reliability as ‘the likelihood that a system will perform in an acceptable manner’ (Bea, 2006) is important in forensic geotechnical engineering.” No details were discussed. Engineering with natural materials such as soils and rocks is difficult and challenging owing to the variability and the uncertainties associated with engineering decisions. In fact, one key aspect that distinguishes geotechnical engineering from structural engineering is the large variability (natural/intrinsic variability, testing errors, and transformation uncertainties introduced when measured parameters are converted to engineering parameters) related to naturally occurring geomaterials. In the case of quality of concrete, variability (in terms of coefficient of variation or COV) can be classified as follows (Rétháti 1988 citing 1965 specification of the American Concrete Institute): COV < 10% excellent COV = 10 – 15% good COV = 15 – 20% satisfactory COV > 20% bad The COVs related to the natural variability of geomaterials as well as those pertaining to measurements errors and transformation uncertainties can be much larger and do not fall within a narrow range. Phoon & Kulhawy (1999a; 1999b) compiled an extensive database of soil statistics for calibration of simplified reliability-based foundation design equations. Details are given in Section 3. One key observation that emerged from their calibration studies is that geotechnical variability of common design parameters can be broadly grouped using the following qualitative categories:

Geotechnical parameter Property variability COV (%) Undrained shear strength Low 10 – 30 Medium 30 – 50 High 50 – 70 Effective stress friction angle Low 5 – 10 Medium 10 – 15 High 15 – 20 Horizontal stress coefficient Low 30 - 50 Medium 50 – 70 High 70 - 90

These observations are not entirely surprising given that the volume of geo-materials investigated by direct or indirect means is extremely small in comparison to the volume of interest. Chiles and Delfiner (1999) cited volume fractions investigated at Brent field, North sea, to be 10-9 for cores and cuttings and 10-6 for logging. The immediate impact of these observations is that Leonards (1982) definition of “failure” - unacceptable difference between expected and observed performance – cannot be evaluated in a meaningful way using deterministic methods. To elaborate, “expected performance” must vary given the backdrop of potentially significant geotechnical variability. We can argue that this variability is “inherent”, in the sense of being irreducible, in geotechnical engineering based on the prevailing standards of best practice. Some reduction in some of the components of variability is always possible by carrying out more tests, but it is generally not realistic to expect orders of magnitudes reduction. In view of this widely acknowledged and widely established reality in geotechnical engineering, it is more realistic and perhaps more credible to quantify “unacceptable difference” in a statistical sense. In broad terms, forensic engineering is related to the investigation of failures with the view of rendering an opinion regarding responsibility. Hence, an objective statistical measure of “unacceptable difference” (specifically, a difference not explainable by underlying variability) should provide useful additional information in the formulation of such an opinion. It should be emphasized, however, that the statistical criteria tentatively proposed in the following sections should never replace – nor downplay the importance of - preliminary qualitative assessments of the quality of design based on geotechnical knowledge

3

and experience. Rather, they should serve as a complementary tool that could improve the degree of objectivity in forensic geotechnical assessments. 2. STATISTICAL FRAMEWORK 2.1 Rejection criteria for factor of safety A preliminary statistical framework can be constructed by using the widely used global factor of safety (FS) as an indicator of performance. It is plausible to imagine FS as a lognormal random variable with mean = μ, standard deviation = σ and coefficient of variation = σ/μ = θ. Silva et al. (2008) provided some empirical evidence to support this lognormal assumption for slope problems. The lognormal distribution satisfies the most basic constraint that most load and strength parameters are non-negative. It is implicit in this model that one speaks of the “likelihood” of failure, rather than failure in our usual absolute deterministic sense. Fig. 1 illustrates a lognormal distribution for μ = 3 and θ = 0.3. It can be shown that ln(FS) is normally distributed with mean (λ) and standard deviation (ξ) given by:

λ = ln(μ) – 0.5ξ2 and ξ2 = ln(1+θ2). (1)

The equivalent normal distribution for Fig. 1 is shown in Fig. 2 with λ = 1.056 and ξ = 0.29. Note that the standard deviation of ln(FS) (ξ) is approximately equal to the COV of FS (θ) up to θ of about 0.5.

0

0.2

0.4

0.6

0 2 4 6

FS

Prob

abili

ty d

ensit

y Fail

Fig.1 Log-normal distribution of FS

4

0.0

0.5

1.0

1.5

-1 0 1 2 3

ln(FS)

Prob

abili

ty d

ensit

y Fail

Fig. 2 Equivalent normal distribution The observed performance, FS , can be viewed as a sample mean (of sample size 1) from a population of

factors of safety such as that shown in Fig. 1. It is obvious that a low value of FS near to 1 is not impossible, although it is “unlikely” if the population follows a lognormal distribution with μ = 3 and θ = 0.3. A naïve interpretation is that “failure” (as in FS < 1) is always possible and geotechnical variability is responsible, rather than human errors. Sowers (1993) noted that the majority of foundation failures were due to human shortcomings. This naïve interpretation essentially misses the key principle of hypothesis testing described below. Our proposed approach is to assume that: (a) “expected performance” can be described by the reliability index and (b) a conventional hypothesis test can be performed to ascertain if the target reliability index had been achieved in the original design based on the observed sample mean, FS . To do this, it is necessary to define “expected performance”. If FS follows a log-normal population, the reliability index is given by:

ξλ

=σ

=β)]FS[ln()]FS[ln(E

(2)

Numerous reliability calibration studies (e.g., Phoon et al. 1995) have shown that existing foundations are typically designed to achieve a target reliability index (β) of about 3 (corresponding to a probability of failure of about one in a thousand). Target reliability indices for other types of geotechnical design are not as well established presently. In Fig. 2, the reliability index is β = 3.64 or probability of failure = Φ-1(-β) = 0.000136, where Φ is the standard normal cumulative distribution function. In EXCEL, Φ(.) and Φ-1(.) are computed using the NORMSDIST(.) and NORMSINV(.) functions, respectively. Based on the definition of the reliability index and β = 3, the following null and alternate hypotheses on the population mean [actually, the mean of ln(FS)] can be formulated as: H0: λ = 3ξ H1: λ < 3ξ

5

Assuming that ξ is known and a sample size of 1, the null hypothesis is rejected at the customary 5% level of significance if:

( ) ( ) ( ) 645.105.03FSlnFSln 1 −=Φ<ξ

ξ−=

ξλ− − (3)

( )[ ]21ln355.1exp)355.1exp(FS θ+=ξ< (4)

The above rejection criterion provides an example of a simple numerical yardstick to evaluate “unacceptable difference between expected and observed performance” in the presence of potentially significant geotechnical variability. A rejection means that the observed factor of safety does not support the claim that β = 3 at the 5% level of significance. Different values of the threshold factor of safety can be computed for different significance levels. It is also possible for a rejection to arise because the underlying “true” geotechnical variability was grossly over-estimated. A “do not reject” scenario means that the observed factor of safety is not unreasonably “low” and failure may be caused by geologic “surprises”, limitations in the existing factor of safety, critical failure mechanism not identified, etc. Again, it is possible for “do not reject” to arise because we have grossly under-estimated the underlying geotechnical variability. The above framework is by no means perfect and comprehensive, given the diversity and complexities of actual failures. However, it does provide an example of an objective framework to perform an initial evaluation of the observed factor of safety, particularly to eliminate the more obvious claim that it is an “unfortunate” realization caused by geotechnical variability. Fig. 3 presents the critical values of factor of safety below which the difference between expected and observed performance is “unacceptable”. For example, if the expected performance corresponds to β = 3 and the underlying variability corresponds to θ = 0.3, an observed factor of safety of 1.4 is too “low” and not explainable as a random outcome from a population of FS with β = 3.0 and θ = 0.3. It will be useful to interpret actual case studies using this chart to evaluate its usefulness in forensic geotechnical engineering and to fine-tune it if necessary. In the general case where we observe n factors of safety, the rejection criterion for β will become:

( )( )[ ]n645.1 )1ln(exp

n645.1expFS2 −βθ+=

ξ−βξ< (5)

The rejection curves for β = 3.0 corresponding to various sample sizes (n) are shown in Fig. 4. If this statistical framework proves to be useful, more sophisticated rejection criteria can be developed based on a sample estimate of ξ (rather than the population version used in the above equations). It suffices to note here that they do not follow the t-distribution in standard statistical texts.

6

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Coefficient of variation

Cri

tical

fact

or o

f saf

ety

β = 3.5

β = 3.0

β = 2.5

Fig. 3 Rejection criteria for different expected performance levels - a possible role for risk in FGE based on one observed factor of safety

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Coefficient of variation

Cri

tical

fact

or o

f saf

ety

n = 5

n = 2

n = 1

Fig. 4 Rejection criteria for different sample sizes of observed factors of safety based an expected performance of β = 3. 2.2 Example An example of the use of Fig. 3 can be illustrated with reference to the results presented in Duncan (2000) wherein a case study of underwater slope failure was reported. The failure took place entirely within San Francisco Bay mud, a normally consolidated, slightly organic clayey silt or silty clay of marine origin. Previous experience in the area indicated that 1(H):1(V) slopes with a factor of safety of 1.25 were satisfactory. To reduce the quantity of excavation, slopes of 0.875(H): 1(V) having a factor of safety of 1.17 were excavated which failed subsequently. A risk-based back-analysis indicated a probability of failure of 18% which is

7

unacceptable though the factor of safety is 1.17. Duncan (2000) mentioned that the coefficient of variation of shear strength parameters was high and hence the probability of failure was high. This case study can be plotted in Fig. 3 as an open circle. It is clear that an observed FS = 1.17 cannot support the claim that β = 2.5 for any COV larger than about 0.2. Following the same argument, we can conclude that β = 3 is not supported for any COV larger than about 0.1. Because COV of 0.1 is the lower bound for most geotechnical problems, it is reasonable to say that β = 3 was not achieved in the original design with a fair degree of confidence. 3. GEOTECHNICAL UNCERTAINTIES The evaluation of geomaterial (soil and rock) parameters is one of the key design aspects that distinguish geotechnical from structural engineering. The basic premise underlying Section 2 is that the factor of safety (FS) is a lognormal random variable. It is possible to infer the underlying coefficients of variation of FS from the empirical evidence presented in Silva et al. (2008). This approach of lumping all sources of geotechnical uncertainties into a COV of FS is quite crude, but it does allow practitioners to apply the statistical framework in Section 2 rapidly to test the claim that failure is an unfortunate “roll of the dice” caused by geotechnical uncertainties. Some important COV thresholds for FS are given in Section 3.1. It is preferable to distinguish between two main sources of geotechnical uncertainties explicitly. The first arises from the evaluation of design soil properties, such as undrained shear strength and effective stress friction angle. This source of geotechnical uncertainty is complex and depends on inherent soil variability, degree of equipment and procedural control maintained during site investigation, and precision of the correlation model used to relate field measurement with design soil property. A total variability analysis that lumps all these components produces only site-specific statistics which should not be generalized to other sites. Realistic statistical estimates of the variability of design soil properties have been established by Phoon & Kulhawy (1999a, 1999b) and presented in Section 3.2. The second source arises from geotechnical calculation models. Although many geotechnical calculation models are “simple”, reasonable predictions of fairly complex soil-structure interaction behavior still can be achieved through empirical calibrations. Model factors, defined as the ratio of the measured response to the calculated response, usually are used to correct for simplifications in the calculation models. Recent literature includes estimation of model statistics for the calibration of deep foundation resistance factors for AASHTO [American Association of State Highway and Transportation Officials] (Paikowsky 2002). None of these studies addresses the applicability of model statistics beyond the conditions implied in the database. This question mirrors the same concern expressed previously on the possible site-specific nature of soil variabilities. More rigorous model statistics for both ultimate and serviceability limit states are presented in Section 3.3. It is crucial to highlight that the COV of FS is indeed exceedingly crude, because it lumps parametric and model uncertainties into a single number. 3.1 Coefficient of variation of factor of safety Silva et al. (2008) proposed several relationships between the annual probability of failure and the factor of safety based on 75 projects (zoned and homogeneous earth dams, tailings dams, natural and cut slopes, and some earth retaining structures) and expert judgment. As shown in Fig. 5, there are 4 categories of earth structures. They are defined by Silva et al. (2008) as: Category I—facilities designed, built, and operated with state-of-the-practice engineering. Generally these

facilities have high failure consequences; Category II—facilities designed, built, and operated using standard engineering practice. Many ordinary

facilities fall into this category; Category III—facilities without site-specific design and substandard construction or operation. Temporary

facilities and those with low failure consequences often fall into this category; Category IV—facilities with little or no engineering.

Silva et al. (2008) compared the empirical data in Fig. 5 with some theoretical curves, but did not provide any mathematical details. Their theoretical curves can be easily reproduced using the following procedure:

1. Assume that FS is lognormally distributed with parameters λ and ξ. 2. The horizontal axis of Fig. 5 is the mean factor of safety, μ. 3. If the COV of FS (θ) is sufficiently small, λ ≈ ln(μ) and ξ ≈ θ. 4. The vertical axis of Fig. 5 is the probability of failure, given by:

8

( ) [ ] ( )⎥⎦⎤

⎢⎣⎡

θ−

Φ=⎟⎟⎠

⎞⎜⎜⎝

⎛ξλ−

Φ=<=<=FSln00)FSln(obPr1FSobPrpf

Fig. 5 Annual probability of failure versus factor of safety for earth structures (Silva et al. 2008) The theoretical lognormal curves from Silva et al. (2008) are reproduced in Fig. 6 using the above procedure with θ = 0.072 (Category I), 0.109 (Category II), 0.174 (Category III), 0.316 (Category IV). Based on Silva et al. (2008), it would appear that the COV of FS for standard earth structures is between 0.1 and 0.2. These COV thresholds are important for FGE and they are summarized in Table 1. Lognormal probability curves for a more complete and more systematic range of COVs are shown in Fig. 7.

9

1.E-08

1.E-07

1.E-06

1.E-05

1.E-04

1.E-03

1.E-02

1.E-01

1.E+000.5 1.5 2.5 3.5 4.5

Mean factor of safety

Ann

ual p

roba

bilit

y of

failu

re

I (COV = 0.072)

II (COV = 0.109)

III (COV = 0.174)

IV (COV = 0.316)

Fig. 6 Lognormal probability curves back-calculated from Silva et al. (2008)

1.E-08

1.E-07

1.E-06

1.E-05

1.E-04

1.E-03

1.E-02

1.E-01

1.E+000.5 1.5 2.5 3.5 4.5

Mean factor of safety

Ann

ual p

roba

bilit

y of

failu

re

0.050.100.200.300.400.50

Coefficient of variation (COV)

Fig. 7 Lognormal probability curves for higher COVs of FS

10

Table 1. COV thresholds for factors of safety corresponding to different categories of earth structures (back-calculated from Silva et al., 2008)

Categories of earth structures COV(FS) I 0.072 II 0.109 III 0.174 IV 0.316

3.2 Parametric uncertainties There are three primary sources of geotechnical uncertainties as illustrated in Fig. 8: (a) inherent variability, (b) measurement error, and (c) transformation uncertainty (Phoon and Kulhawy 1999a). A comprehensive effort was undertaken to provide realistic geomaterial statistics of sufficient generality to underpin current and future developments of practical reliability-based design codes (Phoon and Kulhawy, 1999a, 1999b; Kulhawy et al., 2000). The results of this compilation effort are also useful for FGE and are summarized below. Others are reported by Jones et al. (2002) and Uzielli et al. (2006). 3.2.1 Inherent or natural variability Inherent variability results primarily from the “real” heterogeneity which is inherent to geomaterials. Such heterogeneity stems from natural geologic and physical/chemical/biological processes that produced and continually modify the soil/rock mass in-situ. Uzielli et al. (2006) provided a state-of-the-art review of methods for quantifying inherent variability in geotechnical engineering analyses and design. Approximate guidelines for inherent soil variability are given in Table 2. A more detailed presentation containing the mean, standard deviation (S.D.), and range of the COV are shown in Table 3 with the total number of data groups per test (m). The statistics for rock are given for comparison.

Fig. 8 Sources of uncertainties contributing to overall uncertainty in design soil parameter. Table 2. Approximate guidelines for ranges of mean values and COVs of inherent soil variability (Phoon and Kulhawy 1999a). Test type Parametera Soil type Mean COV(%) Lab strength su(UC) Clay 10-400 kN/m2 20-55

11

su(UU) Clay 10-350 kN/m2 10-30 su(CIUC) Clay 150-700 kN/m2 20-40 φ Clay & sand 20-40o 5-15 CPT qT Clay 0.5-2.5 MN/m2 < 20 qc Clay 0.5-2.0 MN/m2 20-40 Sand 0.5-30.0 MN/m2 20-60 VST su(VST) Clay 5-400 kN/m2 10-40 SPT N Clay & sand 10-70 blows/ft 25-50 DMT A reading Clay 100-450 kN/m2 10-35 Sand 60-1300 kN/m2 20-50 B reading Clay 500-880 kN/m2 10-35 Sand 350-2400 kN/m2 20-50 ID Sand 1-8 20-60 KD Sand 2-30 20-60 ED Sand 10-50 MN/m2 15-65 PMT pL Clay 400-2800 kN/m2 10-35 Sand 1600-3500 kN/m2 20-50 EPMT Sand 5-15 MN/m2 15-65 Lab index wn Clay & silt 13-100 % 8-30 wL Clay & silt 30-90 % 6-30 wP Clay & silt 15-25 % 6-30 PI Clay & silt 10-40 % b LI Clay & silt 10 % b γ, γd Clay & silt 13-20 kN/m3 < 10 Dr Sand 30-70 % 10-40c 50-70d a - su = undrained shear strength; UC = unconfined compression test; UU = unconsolidated-undrained

triaxial compression test; CIUC = consolidated isotropic undrained triaxial compression test; φ = effective stress friction angle; qT = corrected cone tip resistance; qc = cone tip resistance; VST = vane shear test; N = standard penetration test blow count; A & B readings, ID, KD, & ED = dilatometer A & B readings, material index, horizontal stress index, & modulus; pL & EPMT = pressuremeter limit stress & modulus; wn = natural water content; wL = liquid limit; wP = plastic limit; PI = plasticity index; LI = liquidity index; γ & γd = total & dry unit weights; Dr = relative density

b - COV = (3-12%) / mean c - total variability for direct method of determination d - total variability for indirect determination using SPT values Table 3. Inherent soil variability for soil and rock (Kulhawy et al. 2000).

Coefficient of variation (%) Test type

Parameter Material type m mean S.D. range

Index γ, γd fine-grained 14 7.8 5.8 2-20 wn fine-grained 40 18.1 7.9 7-46 wP fine-grained 23 15.7 6.0 6-34 wL fine-grained 38 18.1 7.1 7-39 PI - all data fine-grained 33 29.5 10.8 9-57 - ≤ 20% fine-grained 13 35.0 11.4 16-57 - > 20% fine-grained 20 26.0 9.0 9-40 γ, γd rock 42 0.9 0.7 0.1-3 n rock 25 25.9 19.4 3-71 Strength φ, tan φ sand, clay 48 13.9 10.4 4-50

12

sand 32 9.0 3.0 4-15 clay 16 23.5 13.0 10-50 su clay 100 31.5 14.2 6-80 qu rock 184 14.2 11.7 0.3-61 qt-brazilian rock 74 16.6 10.4 2-58 Stiffness Et-50 rock 32 30.7 15.0 7-63 CPT qc sand, clay 65 36.6 15.5 10-81 sand 54 38.2 16.3 10-81 clay 11 28.4 6.8 16-40 qT clay 9 7.9 4.9 2-17 VST su(VST) clay 26 25.3 6.5 13-36 SPT N sand, clay 23 38.0 10.8 19-62 DMT A,B sand, clay 56 27.9 11.9 12-59 sand 30 34.8 11.3 13-59 clay 26 19.9 6.2 12-38 ID - all data sand 30 40.7 21.6 8-130 - w/o outliers 29 37.7 14.2 8-66 KD - all data sand 31 41.2 19.2 15-99 - w/o outliers 29 37.6 13.5 15-67 ED - all data sand 31 42.7 19.6 7-92 - w/o outliers 30 41.1 17.6 7-69 3.2.2 Measurement error Inherent variability is caused primarily by the natural geologic processes that are involved in soil formation. Measurement error, on the other hand, arises from equipment, procedural/ operator, and random testing effects. A summary of total measurement error is given in Table 4 for laboratory tests and in Table 5 for field tests. Table 4. Summary of total measurement error of some laboratory tests (Phoon & Kulhawy 1999a). Parametera Soil

type No. data

group No. tests/group Property value

(unitsb) Property COV (%)

Mean Range Mean Range Mean Range su(TC) Clay, silt 11 - 13 7-407 125 8-38 19 su(DS) Clay, silt 2 13-17 15 108-130 119 19-20 20 su(LV) Clay 15 - - 4-123 29 5-37 13 φ(TC) Clay, silt 4 9-13 10 2-27 o 19.1o 7-56 24 φ(DS) Clay, silt 5 9-13 11 24-40o 33.3o 3-29 13 Sand 2 26 26 30-35o 32.7o 13-14 14 tan φ(TC) Sand, silt 6 - - - - 2-22 8 tan φ (DS) Clay 2 - - - - 6-22 14 wn Fine-grained 3 82-88 85 16-21 18 6-12 8 wL Fine-grained 26 41-89 64 17-113 36 3-11 7 wP Fine-grained 26 41-89 62 12-35 21 7-18 10 PI Fine-grained 10 41-89 61 4-44 23 5-51 24 γ Fine-grained 3 82-88 85 16-17 17.0 1-2 1 a - su = undrained shear strength; φ = effective stress friction angle; TC = triaxial compression test;

UC = unconfined compression test; DS = direct shear test; LV = laboratory vane shear test; wn = natural water content; wL = liquid limit; wP = plastic limit; PI = plasticity index; γ = total unit weight

b - units of su = kN/m2; units of wn, wL, wP, and PI = %; units of γ = kN/m3

13

Table 5. Summary of measurement error of common in-situ tests (Kulhawy & Trautmann 1996).

Coefficient of variation, COV (%) Test Equipment Procedure Random Totala Rangeb

Standard penetration test (SPT) 5c - 75d 5c - 75d 12 - 15 14c - 100d 15 - 45 Mechanical cone penetration test (MCPT) 5 10e -15f 10e -15f 15e -22f 15 - 25

Electric cone penetration test (ECPT) 3 5 5e -10f 7e - 12f 5 - 15 Vane shear test (VST) 5 8 10 14 10 - 20 Dilatometer test (DMT) 5 5 8 11 5 - 15 Pressuremeter test, pre-bored (PMT) 5 12 10 16 10 - 20g Self-boring pressuremeter test (SBPMT) 8 15 8 19 15 - 25g

a - COV(Total) = [COV(Equipment)2 + COV(Procedure)2 + COV(Random)2]0.5 b - Because of limited data and judgment involved in estimating COVs, ranges represent probable

magnitudes of field test measurement error c, d - Best to worst case scenarios, respectively, for SPT e, f - Tip and side resistances, respectively, for CPT g - It is likely that results may differ for po, pf, and pL, but the data are insufficient to clarify this issue 3.2.3 Transformation uncertainty The third component of uncertainty is introduced when field or laboratory measurements are transformed into design parameters using empirical or other correlation models (e.g., correlating the standard penetration test N value with the undrained shear strength) as shown in Fig. 9. Obviously, the relative contribution of these components to the overall uncertainty in the design parameter depends on the site conditions, degree of equipment and procedural control, and quality of the correlation model. Therefore, geomaterial statistics that are determined from total variability analyses only can be applied to the specific set of circumstances (site conditions, measurement techniques, correlation models) for which the design parameters were derived. In other words, the COV of geomaterials cannot be viewed as an intrinsic statistical property.

For each combination of soil type, measurement technique, and correlation model, the uncertainty in the design soil property is evaluated systematically by combining the appropriate component uncertainties using a simple second-moment probabilistic approach:

22

2e

22w

22d sTs

eTs

wT s εξ ⎟

⎠⎞

⎜⎝⎛∂ε∂

+⎟⎠⎞

⎜⎝⎛∂∂

+⎟⎠⎞

⎜⎝⎛∂∂

≈ (6)

in which ξd = T(ξm, ε), T(⋅) = correlation function between test measurement (ξm) and design parameter (ξd), ε = transformation uncertainty, w = inherent variability, e = measurement error, and s2 = variance.

14

Figure 9. Transformation uncertainty (Phoon & Kulhawy 1999b).

The above equation refers to the variance of the design parameter at a point in the soil mass. For foundation

design, it is not uncommon to evaluate the spatial average of the design parameter over some depth interval, rather than using the value of the design parameter at a point. The spatial average of ξd is defined as:

∫ ξξ b

t

z

z da

a dz )z(L

1 = (7)

in which ξa = spatial average, zt and zb = top and bottom coordinates of a depth interval, respectively, and La = zb - zt = averaging length.

If t and ∂T/∂w are constants, it can be shown that the variance of the spatial average (sξa2) is given by (Phoon

& Kulhawy 1999b):

( ) 22

2e

22wa

22

2a sTs

eTsL

wT s εξ ⎟

⎠⎞

⎜⎝⎛∂ε∂

+⎟⎠⎞

⎜⎝⎛∂∂

+Γ⎟⎠⎞

⎜⎝⎛∂∂

≈ (8)

in which Γ2(⋅) = variance reduction function, which depends on the length of the averaging interval, La. The following approximate variance reduction function is proposed for practical application (Vanmarcke 1983): Γ2(La) = 1 for La ≤ δv Γ2(La) = δv/La for La > δv

in which δv = vertical scale of fluctuation. It can be seen that the variance reduction function decreases as the length of the averaging interval increases. Therefore, the effect of averaging is to reduce the uncertainty associated with inherent variability (sw

2). The scale of fluctuation quantifies the spatial extension in which a property of interest can be considered

significantly autocorrelated. Within separation distances smaller than the scale of fluctuation, the deviations from the trend function are expected to show relatively strong correlation. When the separation distance between two sample points exceeds the scale of fluctuation, it can be assumed that little correlation exists between the fluctuations in the measurements. From a geological and geomorphological perspective, it is intuitive that soil formation and modification processes, as well as factors contributing to the definition of the

15

in-situ state (e.g. stress) would result in a greater heterogeneity of soil properties in the vertical direction and, hence, in a weaker spatial correlation. Hence, the scale of fluctuation of a given soil property in the vertical direction is generally much smaller than in the horizontal direction.

The scale of fluctuation is not an inherent property of a soil parameter. Numerical values of the scale of fluctuation depend at least on: (a) the spatial direction [e.g. horizontal, vertical]; (b) the measurement interval in the source data; (c) the type of trend which is removed during decomposition; (d) the method of estimation of the scale of fluctuation from residuals; and (e) modelling options from the specific estimation method. Uzielli et al. (2006) discussed these issues in detail.

Vanmarcke (1977) proposed the following approximate relationship for evaluating the scale of fluctuation:

2δ ≈ Δ

π (9)

where

cn

ii 1c

1n =

Δ = Δ∑ (10)

is the average distance between the intersections of the fluctuating component and the trend of a given profile (see Fig.10).

Figure 10. Simplified estimation of the scale of fluctuation as proposed by Vanmarcke (1983) Other methods to estimate the scale of fluctuation have been implemented in the geotechnical literature. Uzielli et al. (2006) provided an overview of such methods. Uzielli (2004) compared the estimates of the scale of fluctuation of profiles of normalized cone tip resistance using different estimation methods. None of the models were shown to rank consistently above or below the others, though however, in some cases the scatter among

16

estimates from different methods is significant. Ranges of �v for a number of geotechnical parameters are given in Table 6. Table 6. Literature values for the horizontal (�h) and vertical (�v) scale of fluctuation of geotechnical parameters (from Uzielli et al. 2006) Property* Soil type Testing method** �h (m) �v (m) su clay lab. testing - 0.8-8.6 su clay VST 46.0-60.0 2.0-6.2 qc sand, clay CPT 3.0-80.0 0.1-3.0 qc offshore soils CPT 14-38 0.3-0.4 1/qc alluvial deposits CPT - 0.1-2.6 qt clay CPTU 23.0-66.0 0.2-0.5 qc1N cohesive-behaviour soils CPT - 0.1-0.6 qc1N intermediate-behaviour soils CPT - 0.3-1.0 qc1N cohesionless-behaviour soils CPT - 0.4-1.1 fs sand CPT - 1.3 fs deltaic soils CPT - 0.3-0.4 FR cohesive-behaviour soils CPT - 0.1-0.5 FR intermediate-behaviour soils CPT - 0.1-0.6 FR cohesionless-behaviour soils CPT - 0.2-0.6 Ic cohesive-behaviour soils CPT - 0.2-0.5 Ic intermediate-behaviour soils CPT - 0.6 Ic cohesionless-behaviour soils CPT - 0.3-1.2 N sand SPT - 2.4 w clay, loam lab. testing 170.0 1.6-12.7 wL clay, loam lab. testing - 1.6-8.7 �´ clay lab. testing - 1.6 � clay, loam lab. testing - 2.4-7.9 e organic silty clay lab. testing - 3.0 �´p organic silty clay lab. testing 180.0 0.6 KS dry sand fill PLT 0.3 - ln(DR) sand SPT 67.0 3.7 n sand - 3.3 6.5 * su=undrained shear strength; qc=cone tip resistance; qt=corrected cone tip resistance;

qc1N=dimensionless, stress-normalised cone tip resistance; fs=sleeve friction; FR=stress-normalised friction ratio; Ic=CPT soil behaviour classification index; N=SPT blow count; w=water content; wL=liquid limit; �´=submerged unit weight; �=unit weight; e=void ratio; �´p=preconsolidation pressure; KS=subgrade modulus; DR=relative density; n=porosity

** VST=vane shear testing; CPT=cone penetration testing; CPTU=piezocone testing; SPT=standard penetration testing; PLT=plate load testing

Useful guidelines on typical “total” coefficients of variation of many common design soil strength properties

have been summarized by Phoon and Kulhawy (1999b) and are given in Table 7 for reference. Once again, tabulated values should not be applied uncritically, i.e. if it has not been assessed that the variability related to site conditions, testing methods and transformation models are significantly similar to those used to obtain specific literature values. Table 7. Approximate guidelines for design soil parameter variability (Phoon & Kulhawy (1999b). Design parametera Testb Soil type Point

COV (%) Spatial avg. COVc (%)

Correlation Equationf

su(UC) Direct (lab) Clay 20-55 10-40 - su(UU) Direct (lab) Clay 10-35 7-25 -

17

su(CIUC) Direct (lab) Clay 20-45 10-30 - su(field) VST Clay 15-50 15-50 14 su(UU) qT Clay 30-40e 30-35e 18 su(CIUC) qT Clay 35-50e 35-40e 18 su(UU) N Clay 40-60 40-55 23 su

d KD Clay 30-55 30-55 29 su(field) PI Clay 30-55e - 32 φ Direct (lab) Clay, sand 7-20 6-20 - φ (TC) qT Sand 10-15e 10e 38 φcv PI Clay 15-20e 15-20e 43 Ko Direct (SBPMT) Clay 20-45 15-45 - Ko Direct (SBPMT) Sand 25-55 20-55 - Ko KD Clay 35-50e 35-50e 49 Ko N Clay 40-75e - 54 EPMT Direct (PMT) Sand 20-70 15-70 - ED Direct (DMT) Sand 15-70 10-70 - EPMT N Clay 85-95 85-95 61 ED N Silt 40-60 35-55 64 a - su = undrained shear strength; UU = unconsolidated-undrained triaxial compression test;

UC = unconfined compression test; CIUC = consolidated isotropic undrained triaxial compression test; su(field) = corrected su from vane shear test; φ = effective stress friction angle; TC = triaxial compression; φcv = constant volume φ; Ko = in-situ horizontal stress coefficient; EPMT = pressuremeter modulus; ED = dilatometer modulus

b - VST = vane shear test; qT = corrected cone tip resistance; N = standard penetration test blow count; KD = dilatometer horizontal stress index; PI = plasticity index

c - averaging over 5 meters d - mixture of su from UU, UC, and VST e - COV is a function of the mean; refer to COV equations in Phoon & Kulhawy (1999b) for details f - Quality of correlation affects the COV of design parameters. The correlation models used are

referenced by the equation numbers in Phoon & Kulhawy (1999b) 3.3 Model uncertainties 3.3.1. Ultimate limit state Robust model statistics can only be evaluated using: (1) realistically large scale prototype tests, (2) a sufficiently large and representative database, and (3) reasonably high quality testing where extraneous uncertainties are well controlled. It is common to correct for simplifications in the calculation model using the following multiplicative form (as exemplified by laterally loaded drilled shafts):

Hm = M·Hu (11)

in which Hm = “measured” lateral capacity (more precisely, capacity interpreted from load test), Hu = ultimate lateral capacity computed using limit equilibrium analysis, and M = model factor, typically assumed to be an independent log-normal random variable. It is well known that many different models exist for the computation of Hu.

The distributions of the model factors for Hm determined using 2 different criteria (HL or Hh) and Hu computed from 4 different lateral soil stress models are shown in Fig. 11. Note that M < 1 implies that the calculated capacity is larger than the measured capacity, which is unconservative. If Hm is defined as the hyperbolic capacity (Hh), M < 1 is most likely unsafe as well since there is no reserved capacity beyond Hh and it is mobilized at very large displacements. Capacity model (soil type)

Lateral or moment limit (HL) Hyperbolic capacity (Hh)

18

Reese (1958) (clay)

0.0 0.8 1.6 2.4 3.2HL/Hu(Reese)

0

0.1

0.2

0.3

0.4

Rel

ativ

e Fr

eque

ncy Mean =

S.D. =COV =

n =pAD =

0.920.270.29720.633

0.4 1.2 2.0 2.8 3.6

Hh/Hu(Reese)

0

0.1

0.2

0.3

0.4

Rel

ativ

e Fr

eque

ncy Mean =

S.D. =COV =

n =pAD =

1.420.410.29740.186

Broms (1964a) (clay)

0.0 0.8 1.6 2.4 3.2HL/Hu(Broms)

0

0.1

0.2

0.3

0.4

Rel

ativ

e Fr

eque

ncy Mean =

S.D. =COV =

n =pAD =

1.490.570.38720.122

0.4 1.2 2.0 2.8 3.6

Hh/Hu(Broms)

0

0.1

0.2

0.3

0.4

Rel

ativ

e Fr

eque

ncy Mean =

S.D. =COV =

n =pAD =

2.280.850.37740.149

Randolph & Houlsby (1984) (clay)

0.0 0.8 1.6 2.4 3.2HL/Hu(Randolph & Houlsby)

0

0.1

0.2

0.3

0.4

Rel

ativ

e Fr

eque

ncy Mean =

S.D. =COV =

n =pAD =

0.850.240.28720.555

0.4 1.2 2.0 2.8 3.6Hh/Hu(Randolph & Houlsby)

0

0.1

0.2

0.3

0.4R

elat

ive

Freq

uenc

y Mean =S.D. =COV =

n =pAD =

1.320.380.29740.270

Broms (1964b) (sand)

0.0 0.8 1.6 2.4 3.2HL/Hu(simplified Broms)

0

0.1

0.2

0.3

0.4

Rel

ativ

e Fr

eque

ncy

Mean =S.D. =COV =

n =pAD =

0.880.360.41750.736

0.4 1.2 2.0 2.8 3.6

Hh/Hu(simplified Broms)

0

0.1

0.2

0.3

0.4

Rel

ativ

e Fr

eque

ncy

Mean =S.D. =COV =

n =pAD =

1.300.490.38770.141

Figure 11. Distribution of model factors (Phoon & Kulhawy 2005). 3.3.2. Serviceability limit state Phoon et al. (2006) reported a probabilistic characterization of load-displacement curves using an augered cast-in-place (ACIP) pile load test database. The normalized hyperbolic curve considered in their study is expressed as:

19

byay

QQ

STC += (12)

in which Q = applied load, QSTC = failure load interpreted using the slope tangent method, “a” and “b” = curve-fitting parameters, and y = pile butt displacement. Note that the curve-fitting parameters are physically meaningful – the reciprocals of “a” and “b” are equal to the initial slope and asymptotic value, respectively. Each continuous load-displacement curve can be reduced to two curve-fitting parameters (plotted as a single point in Fig. 12). The scatter in the load-displacement curves is captured by the scatter between “a” and “b”. If the values of “a” are plotted as a histogram, a non-uniform distribution will be obtained (Fig. 13). For example, values close to 5 mm are more frequently encountered. The standard approach is to fit one of the many classical distributions (e.g. lognormal distribution) to the histogram. The crucial point to be emphasized here is that such an approach implicitly assumes that “a” and “b” are statistically independent random variables. However, Fig. 12 clearly shows that the variation of “a” (variation along y-axis) and the variation of “b” (variation along x-axis) are coupled. In other words, it is incorrect to assume that “a” can vary independently of “b”. The correct probabilistic model in this case is not independent random variables but a bivariate random vector. Computational details on the construction of this bivariate random vector are given elsewhere (Phoon and Kulhawy, 2008).

0

4

8

12

16

20

0.2 0.4 0.6 0.8 1.0 1.2b parameter

a pa

ram

eter

(mm

)

Augered cast-in-place pile (compression)

0

4

8

12

16

20

0.2 0.4 0.6 0.8 1.0 1.2b parameter

a pa

ram

eter

(mm

) claysand

Spread foundation (uplift)

a

b

No. of tests = Mean =

Standard deviation = Coefficient of variation =

Mean = Standard deviation =

Coefficient of variation =

Correlation =

40 5.15 mm 3.07 mm 0.60 0.62 0.16 0.26 -0.67

a

b

No. of tests = Mean =

Standard deviation = Coefficient of variation =

Mean = Standard deviation =

Coefficient of variation =

Correlation =

85 7.13 mm 4.66 mm 0.65 0.75 0.14 0.18 -0.24

20

0

1

2

3

4

5

0.4 0.6 0.8 1.0 1.2 1.4b parameter

a pa

ram

eter

(mm

) claysand

Drilled shaft (uplift)

0

1

2

3

4

5

0.4 0.6 0.8 1.0 1.2 1.4b parameter

a pa

ram

eter

(mm

)

Pressure-injected footing (uplift)

a

b

No. of tests = Mean =

Standard deviation = Coefficient of variation =

Mean = Standard deviation =

Coefficient of variation =

Correlation =

48 1.34 mm 0.73 mm 0.54 0.89 0.063 0.07 -0.59

a

b

No. of tests = Mean =

Standard deviation = Coefficient of variation =

Mean = Standard deviation =

Coefficient of variation =

Correlation =

25 1.38 mm 0.95 mm 0.68 0.77 0.21 0.27 -0.73

Figure 12. Correlation between hyperbolic parameters (Phoon et al. 2006; 2007)

0 5 10 15 20a parameter (mm)

0

0.1

0.2

0.3

Rel

ativ

e Fr

eque

ncy Mean =

S.D. =COV =

n =pAD =

5.15 mm3.07 mm0.60400.183

0.0 0.5 1.0 1.5 2.0b parameter

0

0.1

0.2

0.3

Rel

ativ

e Fr

eque

ncy Mean =

S.D. =COV =

n =pAD =

0.620.160.26400.887

Figure 13. Marginal distributions of hyperbolic parameters (Phoon et al. 2006). 4. OTHER POTENTIAL ROLES OF RISK-BASED ANALYSIS One of the important requirements in forensic geotechnical engineering is the identification of failure mechanisms. A possible role for risk-based analysis can be discerned from the simulation studies detailed below. Popescu et. al (1997) studied the effects of spatial variability of soil properties on soil liquefaction for a saturated soil deposit subjected to seismic excitation. They compared standard results derived from deterministic inputs and probabilistic results derived from stochastic finite element analyses. They concluded that both the pattern and the amount of dynamically induced pore water pressure buildup are strongly influenced by the spatial variability of soil parameters. For the same average values of soil parameters, more pore water

21

pressure build up is predicted by the stochastic model than by the deterministic model, which is attributed to a water injection phenomenon triggered by the presence of loose pockets in the spatially variable soil deposit. Griffiths et. al (2002) demonstrated that failure mechanisms of a strip footing are considerably different in the case of uniform soils and spatially variable soils. The spatial correlation structure of the foundation soil in terms of autocorrelation structure has an important effect on the failure mechanism. More extensive studies are presented by Kim (2005). Goldsworthy (2007) presented a fairly comprehensive risk-based study of the effect of site investigations on foundation failures. One conclusion is that an optimum site investigation minimizing the risk of failure can be planned based on the variability and autocorrelation structure of the foundation soil. The above studies and other similar efforts demonstrate that failures can be studied in a more realistic way using spatially variable soils, rather than the traditional uniform or layered soil profiles. Notwithstanding this, a significant number of failures does arise from human failures and a lack of proper quality assurance and control plans. Bea (2006) suggested that the failure development process can be categorized into three phases viz, initiating, contributing and propagating causes and a proper risk management strategy is necessary using observational methods. A risk-based approach to forensic geotechnical engineering may include elements such as: i. Detailed soil investigations in the area in the form of vertical and horizontal soil profiles to formulate

plausible hypotheses concerning prevalent/relevant failure mechanisms. The number and spatial location of boreholes and spacing should be adequate to provide a proper estimate of the mean, variance and autocorrelation properties of in-situ soils.

ii. Analysis of the actual failure mechanism in relation to results from probabilistic studies. iii. Reanalysis of different loading and the associated variability. iv. Re-examination of quality assurance plans which include specifications on quality control of materials and

construction, construction sequences proposed and adopted, observations of performance depending on the nature of project such as deformations, pore pressures etc. which are expected in important projects; and

v. Reliability-based computational back-analysis. 5. CONCLUDING REMARKS This chapter seeks to explore if reliability/risk concepts are potentially useful to forensic geotechnical engineering. This study is very preliminary given the dearth of previous literature. One key aspect that distinguishes geotechnical engineering from structural engineering is the natural variability of geo-materials. Within this context, Leonards (1982) definition of “failure” – “unacceptable difference between expected and observed performance” – cannot be evaluated in a meaningful way using deterministic methods. In broad terms, forensic engineering is related to the investigation of failures with the view of rendering an opinion regarding responsibility. Hence, a statistical measure of “unacceptable difference” (specifically, a difference not explainable by underlying variability) should provide useful additional information in the formulation of such an opinion. A preliminary statistical framework is presented to quantify the difference between expected and observed performance in the presence of unavoidable and potentially significant geotechnical variabilities. It should be emphasized that the quality of the statistical analysis can only be as precise, accurate and meaningful as the engineer’s characterization of uncertainties. Other potentially useful results in the recent reliability and risk literature are highlighted. The intention of this chapter is to stimulate further discussions and research in this important but somewhat overlooked area. REFERENCES Bea, R. (2006). Reliability and human factors in geotechnical engineering. Journal of Geotechnical and

Geoenvironmental Engineering, ASCE, 132(5): 631-643.

22

Brown E. T. (2006). Forensic engineering for underground construction. Proceedings of the ISRM International Symposium 2006 and the 4th Asian Rock Mechanics Symposium, Singapore 8 - 10 November 2006, Chapter 1, 1-16.

Chilès, J-P. and Delfiner, P. (1999). Geostatistics – Modeling Spatial Uncertainty. John Wiley and Sons, New

York. Duncan, J.M. (2000). Factors of safety and reliability in geotechnical engineering, Journal of Geotechnical and

Geoenvironmental Engineering, ASCE, Vol.126, No.4, pp. 307-316. Goldsworthy, J. S. (2007). Quantifying the risk of geotechnical site investigations. PhD Thesis, University of

Adelaide. Griffiths, D V, Fenton, G. A and Manoharan, N. (2002). Bearing capacity of rough rigid strip footing on

cohesive soil: probabilistic study. Journal of Geotechnical and Geoenvironmental Engineering, Vol. 128(9), 73-755.

Jones, A.L., Kramer, S.L. & Arduino, P. (2002). Estimation of uncertainty in geotechnical properties for

performance-based earthquake engineering. PEER Report 2002/16, Pacific Earthquake Engineering Research Center, University of California, Berkeley.

Kim, H. (2005). Spatial variability in soils: stiffness and strength, PhD Thesis, Georgia Institute of Technology,

2005. Kulhawy, F.H. & Trautmann, C.H. (1996). Estimation of in-situ test uncertainty. Uncertainty in the Geologic

Environment - From Theory to Practice (GSP 58), ASCE, New York, 269 – 286. Kulhawy, F.H., Phoon, K.K. & Prakoso, W.A. (2000). Uncertainty in the basic properties of natural

geomaterials. Proc. 1st International Conference on Geotechnical Engineering Education and Training, Sinaia, Romania, 297-302.

Leonards, G. A. (1982). Investigation of failures. Journal of the Geotechnical Engineering Division, ASCE,

108(GT2): 187-246. Paikowsky, S.G. (2002). Load and resistance factor design (LRFD) for deep foundations. Proc. International

Workshop on Foundation Design Codes and Soil Investigation in view of International Harmonization and Performance Based Design, Tokyo, Japan, 59 – 94. Balkema, Netherlands.

Phoon, K.K, Kulhawy, F. H & Grigoriu, M. D. (1995). RBD of Foundations for Transmission Line Structures.

Report TR-105000, Electric Power Research Institute (EPRI), Palo Alto. Phoon, K. K. & Kulhawy, F. H. (1999a). Characterization of geotechnical variability. Canadian Geotechnical

Journal, 36(4):612-624. Phoon, K. K. & Kulhawy, F. H. (1999b). Evaluation of geotechnical property variability. Canadian

Geotechnical Journal, 36(4):625-639. Phoon K.K. and Kulhawy, F.H. (2005). Characterization of model uncertainties for laterally loaded rigid drilled

shafts. Geotechnique, 55(1), 45-54. Phoon, K. K. & Kulhawy, F. H. (2008), Serviceability limit state reliability-based design, Chapter 9, Reliability-

Based Design in Geotechnical Engineering: Computations and Applications, Taylor & Francis, April 2008, 344-383.

Phoon, K.K., Chen, J.-R. & Kulhawy, F.H. (2006). Characterization of model uncertainties for augered cast-in-

place (ACIP) piles under axial compression. Foundation Analysis & Design: Innovative Methods (GSP 153), ASCE, Reston, 82-89

23

Phoon, K. K., Chen, J. R. & Kulhawy, F. H. (2007), Probabilistic hyperbolic models for foundation uplift movement, Probabilistic Applications in Geotechnical Engineering (GSP 170), ASCE, Reston, CDROM.

Popescu, R. Prevost, J. H and Deodatis, G. (1997). Effects of spatial variability on soil liquefaction: some design

recommendations. Geotechnique, 47(5):1019-1036. Rétháti, L. (1988). Probabilistic solutions in geotechnics. Elsevier, New York. Silva, F., Lambe, T. W. and Marr, W. A. (2008), Probability and risk of slope failure. Journal of Geotechnical

and Geoenvironmental Engineering, ASCE, 134(12): 1691-1699. Sowers, G. F. (1993). Human factors in civil and geotechnical engineering failures. Journal of Geotechnical

Engineering, ASCE, 119(2): 238-256. Uzielli, M. (2004). Variability of stress-normalized CPT measurements and application to seismic liquefaction

initiation assessment, University of Florence (Italy). Thesis downloadable at: http://www.georisk.eu. Uzielli, M., Lacasse, S., Nadim, F. & Phoon, K.K. (2006). Soil variability analysis for geotechnical practice.

Proc., Second International Workshop on Characterisation and Engineering Properties of Natural Soils, Singapore. Balkema, Netherlands, 1653 – 1752.

Vanmarcke, E.H. (1983). Random Field: Analysis and Synthesis, MIT Press, Cambridge, Massachusetts.

- 1 -

OKOKOKOKOKOKOKOKOK TTTTTTTTTEEEEEEEEEDDDDDDDDDIIIIIIIII MIMIMIMIMIMIMIMIMINNNNNNNNNEEEEEEEEESSSSSSSSSIIIIIIIIITTTTTTTTTEEEEEEEEE

TaTaTaTaTaTaTaTaTabbbbbbbbbuuuuuuuuubbbbbbbb

biiiiiiiiilllllllll MMMMMMMMMooooooooounununununununununttttttttt HHHHaHaHaHHHaaaaaagegegegegegegegegennnnnnnnn

TTTTTTTTTAAAAAAAAABUBUBUBUBUBUBUBUBUBIBIBIBIBIBIBIBIBILLLLLLLLL

WWWWeWeWeWWWestestestestestestsssttteeeeeeeeerrrrrrrrrnnnnnnnnn

LaLaLaLaLaLaLaLaLaeeeeeeeee PrPrPrPrPrPrPrPrProoooooooovvvvvvvvviiiiiiiiinnnnnnnnnccccccccceeeeeeee

e dididdiddididid NiNiNiNiNiNiNiNiNinnnnnnnnngggggggggeeeeeeeeerrrrrrrrruuuuuuuuummmmmmmmm DaDaDaDaDaDaDaDaDarrrrrrrrruuuuuuuuu

iii eeeeee PPPPoPoPoPPPoooooorrrrtrtrtrrrttttttMMMMMMMMMooooooooorrrrrresbesbesbeeeesbesbesbsssbbbyyyyyyyyy TTTeTTeTTTTe OkOkOkOkOkOkOkOkOk

BIBIBIBIBIBIBIBIBIGGGGGGGGGEEEEEEEEE

PAPAPAPAPAPAPAPAPAPUPUPUPUPUPUP

PUPUAAAAAAAAA NNNNNNNNNEEEEEEEEEWWWWWWWWW GGGGGGGGGUUUUUUUUUIIIIIIIIINNNNNNNNNEEEEEEEEEAAAAAAAAA

KIKIKIKIKIKIKIKIKIUNUNUNUNUNUNUNUNUNGGGGGGGGGAAAAAAAAA rrrrrrrrr eeeeeeeee vvvvvvvvv iiiiiiiii 0 50 100 RRRRRRRRR

LLLLLLLLLakeakeakeakeakeakeakeakeake MMMMMMMMMuuuuuuuuurrrrrrrrrrrrrrrrrraaaayayayaaayyyyyy ddddddddd

kilometres nnnnnnnnn aaaaaaaaa klll kkklkkll MaMaMaMaMaMannnnnnddddddaaaaaa kklkll MaMaMannndddaaa ccccccccc rriii rrirrii rririi

SSSStttttt

puapuapuapuapuapua AIAIAIAIAIAIAAAIIIAMAMAMAMAMAMAAAMMMBBBBBBBAKBAKBAKAAAAAAKKKKKK

SSSSSttt

PPapuaPPapuaPPapua WWWWWeWeeeestestststsstterereeeerrrrnnnnnn PrPrPrPrPPrroooooovvvvvviiiiiinnnnnncccccceeeeee aaaaaa

PPP WWWeeestststereerrnnn PrPrProoovvviiinnnccceee WaW Wa W WWasssaaapppeeeaaa

OKOKOKOKOKOKOKOKOK TTTTTTTTTEEEEEEEEEDDDDDDDDDIIIIIIIII MIMIMIMIMIMIMIMIMINNNNNNNNNEEEEEEEEESSSSSSSSSIIIIIIIIITTTTTTTTTEEEEEEEEE

TTTTTTTTTAAAAAAAAABUBUBUBUBUBUBUBUBUBIBIBIBIBIBIBIBIBILLLLLLLLL

dididdiddididid NiNiNiNiNiNiNiNiNinnnnnnnnngggggggggeeeeeeeeerrrrrrrrruuuuuuuuummmmmmmmm

iii eeeeee TTTeTTeTTTTe OkOkOkOkOkOkOkOkOk

BIBIBIBIBIBIBIBIBIGGGGGGGGGEEEEEEEEE

KIKIKIKIKIKIKIKIKIUNUNUNUNUNUNUNUNUNGGGGGGGGGAAAAAAAAA rrrrrrrrr eeeeeeeee vvvvvvvvv iiiiiiiii 0 50 100 RRRRRRRRR

LLLLLLLLLakeakeakeakeakeakeakeakeake MMMMMMMMMuuuuuuuuurrrrrrrrrrrrrrrrrraaaayayayaaayyyyyy ddddddddd

kilometres nnnnnnnnn aaaaaaaaa klll kkklkkll MaMaMaMaMaMannnnnnddddddaaaaaa kklkll MaMaMannndddaaa ccccccccc rriii rrirrii rririi

SSSStttttt

puapuapuapuapuapua AIAIAIAIAIAIAAAIIIAMAMAMAMAMAMAAAMMMBBBBBBBAKBAKBAKAAAAAAKKKKKK

SSSSSttt

PPapuaPPapuaPPapua WWWWWeWeeeestestststsstterereeeerrrrnnnnnn PrPrPrPrPPrroooooovvvvvviiiiiinnnnnncccccceeeeee aaaaaa

PPP WWWeeestststereerrnnn PrPrProoovvviiinnnccceee WaWaWaWaWaWaWaWaWasssssssssaaaaaaaaapppppppppeeeeeeeeeaaaaaaaaa

StStStStStStStStStuuuuuuuuurrrrrrrrrttttttttt IIIIIIIIIssssssssslllllllllaaaaaaaaannnnnnnnnddddddddd FFFFFFFFFlllllllllyyyyyyyyyRRRRRRRRRiiiiiiiiivvvvvvvvveee aaaaaaaaa eeeeeerrrrrrrrr WaWaWaWWWWWWasuasuasuasuasuasusssuuuaaaaaaaaaeeeeeeeee nnnnnn aaaaaaaaa iininiiin iiiiii uuiuiuuui LeLeLewwwaaadddaaa KiKiKiKiKiKiKiwwwwwwwaaaaaaaiiiiiii IIIIIIIssssssslllllllaaaaaaannnnnnnddddddd

DDD'D'D'D'''AAAAAAllllllbbbbbbeeeeeerrrrrrttttttisisiiisisss IsIsIIIIssssllllllaaaaaannnnnndddddd KiKiwwaaii IIssllaanndd

iii sssssssss GuGuGu DD'D''AAAlllbbbeeerrrtttisiiss IsIIsslllaaannnddd LeLeLeLeLeLewwwwwwaaaaaaddddddaaaaaa GGGG G G nenenenenenenenene wwwwwwwww TATATATATAPPPPPPIIIIIILALLLALALAAA TA TATATAPPPIIILALALA

dododododododododo eeeeeeeee nnnnnnnnn NNNNNNNNN SewSewSewSeSeSeSeSeSewwwwwweeeeeeeeerrrrrrrrriiiiiiiiimmmmmmmmmaaaaaaaaabbbbbbbbbuuuuuuuuu IIIIII III aaaaaa SS S SSS iiiiii

TaTaTaTaTaTaTaTaTabbbbbbbbbuuuuuuuuubbbbbbbbbiiiiiiiiilllllllll MMMMMMMMMooooooooounununununununununttttttttt HHHHaHaHaHHHaaaaaagegegegegegegegegennnnnnnnn WWWWeWeWeWWWestestestestestestsssttteeeeeeeeerrrrrrrrrnnnnnnnnn LaLaLaLaLaLaLaLaLaeeeeeeeee PrPrPrPrPrPrPrPrProoooooooovvvvvvvvviiiiiiiiinnnnnnnnnccccccccceeeeeeeee

DaDaDaDaDaDaDaDaDarrrrrrrrruuuuuuuuu PPPPoPoPoPPPoooooorrrrtrtrtrrrttttttMMMMMMMMMooooooooorrrrrrrrresbesbesbeeeesbesbesbsssbbbyyyyyyyyy

PP Paa a p

p p uu u aa a

III I nn n dd d oo o nn n ee e ss s ii i iaa a

PP P aa a pp p uu u aa a NN N e

e e ww w GG G

uu u ii i ini nn ee

aa

OO Okk k

TT T ee e dd d iii ii iiiii ii i

Case Studies 1 TURNING HINDSIGHT INTO FORESIGHT

(LEADERSHIP LESSONS FROM A LANDSLIDE)

Mike Marley BE, MEngSC, FIEAust, RPEQ, ACIS Principal – Golder Associates Pty Ltd 611 Coronation Drive, Toowong, Qld

mmarley@golder.com.au INTRODUCTION

Tom Peters (one of the great business thinkers of our time) proclaims in his “Leadership Essentials” that “great leaders are great story tellers”. He states” “We need stories, Riveting Tales that fire the imagination of ---- as-yet-reluctant heroes-in- waiting”.

This is the story of a forensic engineering study to determine the cause of a major landslide in P.N.G., of the combination of good luck and good management that led to a successful outcome, and of the lessons the author took from the study which have informed his views on project management during the years since the study.

Drawing on Peters’ “Leadership Essentials” (the ability to inspire, liberate, achieve) some lessons in leadership are highlighted, and the importance of populating the leadership literature with similar stories of real projects, as a means of recruiting and nurturing the next generation of inspirational leaders, is emphasised.

THE STORY

Background

The Ok Tedi gold and copper mine was developed in the remote Star Mountains in the Western Province of P.N.G., close to the Irian Jaya border (Figure 1). A tailings storage dam was included in the mine development plan, and construction of this dam was on the critical path for the mine construction programme to ensure that key milestones could be met for the project’s financial model.

OOKK TTT EEDDIIII MMMIIIINNN EEE SSS IIII TTT EEE

TTT AAA BBB UUU BBBIIIILL

TTaabbuubbiiiiillll

WWW eee ssstttteee rrr nnn PPP rrrrooo vvviiiiinnn ccceee

MMM ooo uuunnn tttt HHHaaa gggeee nnn

LLaaee

NNNiiiiiiiiiinnggeerrrr uuu mmm

BBIIIIIGGEE

DDaarr uuu PPoorrrrtttt MMM ooorrrr eee sss bbb yyy

KKIIIIUUNNGGAA

PPAAPPUUAA NNN EEEWWW GGG UUUIIII NNN EEEAAA

LLLaaakkkeee MMuurrrraayy

0 50 100

killometres

MMM aaannn ddd aaa

AAIIIIAAMMBBAAKK

WWeesstteerrnn PPPrrr ooo vvv iiinnn ccc eee

WWaass aaappp eee aaa

SSStttt uuurrrr tttt IIIIIsss lllllaaannn ddd

DD'''''''AAAllllbbb eee rrr ttt iiiiisiss IIIIsss lllllaaannn ddd

WWW aaa sss uuu aaa

LLeeww aaaddd aaa TTAAPPIIII LLL AAA

SSS eee www eeerrrr iiiiimmm aaa bbb uuu SSuuiiiii

DDD AAA RRR UUU

KKiiiiiwww aaaiiiii IIIIIsss lllllaaannn ddd

PPaarr aaammm aaa IIIII sss lllll aaannn ddd

- 2 - DDaarrrr uuu IIIIIsss lllllaaa nnn ddd

Figure 1 A fast track approach was adopted, requiring investigation, and design to proceed concurrently with construction. This approach increases the overall risk associated with the work, and requires the design programme to be sufficiently flexible to accommodate design changes where investigations identify conditions significantly different from assumptions adopted in the preliminary design.

The site is located in one of the most remote and challenging environments in the world: the terrain is rugged; the rainfall extreme (up to 11m per annum); and the jungle is almost impenetrable.

Lesson 1: Understand and respect the environment in which you are operating

History

Preparation for construction of the tailings dam commenced in May 1983.

Excavation for the dam footprint commenced in early November 1983.

In mid December 1983, (when approximately 250,000m³ of excavation had been completed) a landslide occurred in the eastern side of the valley. Approximately 3.4 million m³ was involved in the slide mass. This was followed in early January 1984 by a much larger slide (approximately 35 million m³) covering an area approximately 800m wide and 1.2km long) (Figure 2).

Extent of December 1983 Sl

Extent of January 1984 Slid

Figure 2

An investigation of the slides was initiated (specifically to asses whether the footprint excavation was the primary cause, and whether there were grounds for an insurance claim under the contractors insurance).

Initial Assessment

Initial assessment revealed:

– significant variability in the magnitude of translational movement within the slide mass:

– ratio of excavation volume to slide mass volume for first slide of approximately 1:15;

– ratio of excavation volume to slide mass volume for second slide of approximately 1:154;

- 3 - – the disposition of the slide mass with respect to excavation was particularly asymmetrical;

– slide had apparently occurred on a very low angle failure plane (Figure 3).

Figure 3 There was considerable doubt as to whether the excavation was the sole (or even a significant) contributor to the slide. Other potential causes canvassed at the time include earthquake and/or an intense rainfall event. Clearly assessment of the cause would not be straight forward.

Assembly of a team with a wide range of geo-scientific expertise was the essential first step.

Lesson 2: Complex tasks require high performance teams

The Investigation

Geologically, the Star Mountains are part of the deformation zone associated with the active boundary between the Pacific an Indo-Australian tectonic plates (Figure 4). At this boundary the sub-oceanic crust of the Pacific Plate plunges beneath the more stable continental crust of the Indo-Australian plate causing uplift (Figure 5). There is considerable seismic and volcanic activity associated with this plate sub-duction and the associated rapid uplift has resulted in thick layers of limestone and mudstone, formed from marine deposits, being exposed in the rugged heights of the Star Mountains. A simplified cross-section through the PNG shows the impact of this tectonic activity (Figure 5), including the high central mountain spine and zones of low amplitude folding to the north and south of the spine. This reminder of the mechanics of continental creation proved to be the first vital clue in understanding the cause of the slide.

Figure 4

- 4 -

Figure 5 Initial helicopter reconnaissance surveys to assess the magnitude and extent of the damage caused by the slide also highlighted an unusual feature of the Ok Ma valley (viz. that the valley was broadly leaf-shaped with a narrow, meandering entrance gorge upstream of the dam site, and another similar exit gorge downstream of the site). This feature was repeated regularly throughout the surrounding area, and this repetition of topographic form proved to be the second important clue in understanding the slide.

Detailed study of the aerial photography of the site and its surrounds revealed a number of regional characteristics, in particular several sets of linear features crossing the folded topography surrounding the site (Figure 6). These lineaments (identified under the stereoscope as distinct partings in the tree cover) are indicators of geological structures such as faults, fractures and shears in the rocks underlying the area.

Figure 6 The orientation of the predominant sets of defects was noted.

Lesson 3: Take time to look at the big picture – get above the trees so you can see the form of the forest.

Review of the considerable volume of existing investigation data and records showed:

- 5 - – no significant earthquake or rainfall events that were likely to have initiated the slides,

– the stratigraphic sequence at the site (Figure 7) with the upper levels of the sequence removed by erosion exposing the

Warre limestone as a relatively thin cap overlying approximately 1000m of dark fine grained mudstone (Pnyang formation) in which the slide occurred.

– failure could not be readily explained based on measured strength in mudstone.

Figure 7 Field investigations concentrated on detailed geological mapping of the site, and a drilling and instrumentation installation programme. A number of inclinometers and piezometers were installed throughout the slide mass to record any on-going movement of the slide mass, as well as details of fluctuations in groundwater levels which have a major influence on slope stability.

Airphoto interpretation and field mapping yielded valuable understanding of the regional geomorphology. The leaf-shaped valleys were found to be the results of transverse erosion through the anticlinal folds formed by the compression and uplift caused by regional tectonic forces.

A mechanism was deduced for the valley formation processes as shown in Figure 8:

– Streams originating in overlying formations (since removed)erode down through the upper moderately resistant

limestone cap.

– As uplift and erosion continue, the streams enter the underlying less resistant mudstone.

– Progressive erosion of the mudstone and removal of the toe of the valley slope leads to undercutting and breaking up of the limestone cap.

– Evolution of the slope by repetitive slides in the mudstone creates weak zones approximately parallel to the bedding

with very much lower strength than the surrounding materials.

- 6 -

Figure 8

The establishment of the mechanism of formation of the characteristic leaf shaped valley (Figure 9) was the next critical factor in understanding the cause of the slide.

- 7 -

Figure 9 With a widely scattered team, daily debriefings were held to enable the whole team to stay in touch with progress of the various facets of investigation work, each proceeding at its own characteristic pace.

Lesson 4: Good communication is critical to the effective functioning of high performance teams.

Geological mapping of the immediate area surrounding the slide revealed a closely spaced grid of near vertical joints in the mudstone (Figure 10) and the orientations of these defects were found to align closely with the large-scale lineations observed on the aerial photography.

Figure 10 The orientations of these lineations at both macro and micro scale aligned closely with the theoretically calculated directions of failure planes which would be induced in the mudstone by the tectonic compression forces causing the uplift and folding at the site. Laboratory scale triaxial compression tests on mudstone specimens confirmed this.

All of the major defect sets identified at the site were now able to be explained. Figure 11 shows the defects:

– bedding planes originally horizontal, but now distorted by folding,

– axial defects parallel to the fold axis (major tensile zone),

– conjugate joint sets induced by tectonic compression.

- 8 -

Figure 11 Figure 12 shows a diagrammetric representation of a typical wedge-shaped block of mudstone which would be produced as a result of the intersection of the various defects sets at the site.

Figure 12 Instrumentation records showed that movement was continuing in the slide mass and that the movement appeared to be occurring on multiple planes stepping up in elevation progressively in the downstream direction.

Measured values of shear strength on samples recovered from the vicinity of the shear planes identified in the inclinometers confirmed that actual strengths on these planes (formed initially in the valley forming process) was about one third of the average strength of intact mudstone. These measurements confirmed values of shear strength estimated from back analysis of the stability of the slope to be necessary for slides to be mobilised on planes at the locations indicated by inclinometers. They also confirmed that use of the average values of mudstone strength (as had apparently been done by the original designers) would seriously under-estimate the likelihood of instability.

Lesson 6: Averages are statistical constructs. Understand the numbers that went into their calculation.

Piecing all of the data together allowed a model for the slide mechanism to be developed involving uniform planar movement of a relatively few large wedge-shaped blocks of mudstone. Figure 13 shows a schematic representation of the site and the three dimensional spatial relationships between the slide blocks and the terrain surface. Initial failure of a relatively small block at the upstream end following removal of support at the toe of the slope by the footprint excavation in turn removed support from a succession of further blocks (with different failure plane elevations) resulting in progressive enlargement of the failure downstream.

- 9 -

Figure 13 This model actually recreated the form of the landslide remarkably closely.

Lesson 7: Sketch what you think is happening. It allows a better understanding of complex relationship.

LESSONS

The lessons highlighted during this investigation while significant in their own way, and useful in informing the author’s approach to project management since, have greater real value than simply providing guidance for the future solving of similar technical problems.

While technical papers are an essential tool in the advancement of the science that underpins our professional practice, stories of the execution of major iconic projects are considered to be as important (or more important) in the advancement of the art of management and leadership in our profession.

Stories can identify heroes who have executed iconic projects, point to these pioneers, and provide a real-life point of reference for the next generation of aspiring leaders as a guide to how it is done.

According to Peters, the Essentials of Leadership are to :

• INSPIRE (enthusiastic followers);

•

LIBERATE

(team members’ imagination and passion);

•

ACHIEVE

(outstanding results through others).

In these terms the more profound learnings from the study can be set out as follows: • INSPIRE

− Leadership is not a matter of doing excellence------ it is a matter of inspiring excellence in others.

− Recognise your dependence on the effective functioning of a high performance team.

− Honour those team members whose perspective may appear unconventional or off-beat.

(In this study, listening to an initially improbable notion that plate tectonics was a root cause of the slide helped inspire a cascade of lateral thinking breakthroughs).

• LIBERATE

− Create opportunities: encourage the team to apply their talents and grasp these opportunities;

− “Forget command and control; Forget knowing one’s place; Forget hierarchy”

- 10 -

(In this study, liberating the team from the constraints of adhering solely to conventional quantitative engineering analysis was a key to the successful modelling of the slide mechanism).

• ACHIEVE

− Honour those who try to achieve something rather than those who try to stop something wrong being done;

− Achieving outstanding results requires passion, persistence and imagination

(In this study persistence in pursuing the enigma of failure on planes apparently too shallow to allow movement assuming average material strengths, allowed the achievement of a quantitatively supportable slide mechanism).

THE IMPORTANCE OF STORYTELLING

As Peters attests: “A scintillating story makes an abstract strategy real.”

On completion of a WOW project (according to Peters; one that matters; one that makes a difference; one that transforms the organisation; one that it is worth bragging about) write up the story for inclusion in the library of leadership literature.

And then, recount the story (often) to help inspire and liberate aspiring leaders to achieve their potential.

REFERENCES

1) Hollingswoth D’Appolonia, July 1984 “Ok Tedi Project, Ok Ma Tailings Dam Site, Evaluation of Landslide”. Report to Ok Tedi Mining Ltd.

2) Griffiths, J.N. Hutchinson, D. Brunsden, D.J. Petley and P.G. Fookes; “The Reactivation of a Landslide During the OK Ma Tailing Dam, Papua New Guinea”, J.S. Quaterly Journal of Engineering Geology and Hydrogeology Vol. 37, 2004 pp 173-186.

3) Tom Peters “Tom Peters Essentials, Leadership”, DK Publishing, 2005.

U:\MIKE MARLEY\TURNING HINDSIGHT INTO FORESHIGHT-REV0.DOC

1

Case studies 2

Forensic Studies for Failure in Construction of An Underground Station of the Kaohsiung MRT System

Richard N. Hwang 11F, 3, Dunhwa S. Road, Sec. 1, Taipei, Taiwan 10557

wspc@wspc.com

Kenji Ishihara Chou University, Japan

other@tp.ox.uk

Wei F. Lee Taiwan Construction Research Institute, Taipei, Taiwan

other@tp.ox.uk

Abstract: Several buildings collapsed as a result of leakage on diaphragm wall during the excavation for constructing O1 Station of the Kaohsiung MRT System. Resistivity image profiling was carried out to check the quality of diaphragm walls and the effectiveness of ground treatment using CCP. In addition, undisturbed soil samples were taken by using GP-75S sampler and laboratory tests were conducted for determining the characteristics of soils. This paper presents the findings of these studies. 1 THE PROJECT

The construction for Kaohsiung MRT System (KMRTS) was commenced in October 2001. The system comprises 2 lines, i.e., the Red Line in the NS direction and the Orange Line in the EW direction with a total of 37 stations and 3 depots and is expected to be open for revenue services at the end of year 2007.

The station of interest, i.e., O1 Station (Sizihwan Station), is a terminal station at the western end of the Orange Line. It is located on the north bank of the Kaohsiung Harbour as depicted in Fig. 1. Fig. 2 shows the layout of O1 Station which is 287m in length and 16m in width and is connected to a section of tun-nel at its eastern end.

2 THE EXCAVATION

This 2-level station with side platforms is buried at a depth of 5.12 below ground surface. Excavation for the station was car-ried out to a depth of 20m by using the cut-and-cover method of construction. Diaphragm walls of 800mm in thickness were in-stalled to a depth of 39m in Zones A and B and walls of 1000mm in thickness were installed to a depth of 37.5m in Zones C and D. The pit was propped by six levels of struts as depicted in Fig. 3.

On 9 August 2004, a sinkhole was formed behind Panel S60M on the south side when excavation already reached a depth of 15m and 4 buildings (3-story) collapsed due to leakage of the diaphragm wall. Excavation was halted and measures were taken to improve the watertightness of diaphragm walls. Excavation in Zone B was resumed in July 2005 but a sinkhole was formed shortly at the same location for the second time. Additional ground improvement work was carried out and exca-vation resumed at the end of November 2005. The excavation

was successfully completed and sub-structure erected in June 2006.

3 GROUND CONDITIONS

Fig. 4 shows the soil profile obtained at Borehole WB-11. Al-though the thicknesses of sublayers vary from place to place, the sequence shown in the figure is quite typical. Soil properties adopted in designs are listed in Table 1.

Kaohsiung Harbour

Ai River

Kaohsiung Harbour

Ai River

Fig. 1 Location of O1 Station, Kaohsiung MRT System

2

Table 1 Basic soil properties adopted in design of diaphragm walls

Unit weight

KN/m3 C’

kPa Φ ’

degrees Su kPa

SF 20.1 0 28 CL 18.9 0 27 18 SM 20.2 0 32 CL 19.5 0 29 85 CL 19.5 0 28 113

Table 2 Results of chemical tests on soils

Hole Depth, m PH Organic Cl-

(%) SO4

-2

(mg/kg)SO3

-2

(mg/kg)OA-1 20 7.7 7.38 0.053 2176 5.1 OA-3 13 9.5 3.96 0.038 99.6 11.9 OA-4 15 8.3 5.29 0.11 324.7 7.5 OS-1 16 8.3 1.8 0.049 341.4 159 OS-2 16 8.3 1.8 0.049 323.3 2.6

Table 3 Results of chemical tests on groundwater

Hole Depth

m PH Cl-

(mg/l) SO4

-2

(mg/l) SO3

-2

(mg/l) Ca+2

(mg/l) Ca+2

(mg/l)OA-1 15 7.9 180 112 <2.5 59.7 80.3 OA-3 15 7.2 5030 680 <2.5 113 2480 OS-1 17 6.8 15800 2450 <2.5 984 5530 OS-2 15 7.7 17300 2450 <2.5 1030 4780

Table 4 Classification of aggressiveness of groundwater (ref?)

Cl- (mg/l) SO4-2 /SO3

-2 (mg/l) mild 0 ~ 1,000 0 ~ 150 moderate 1,000 ~ 5,000 150 ~ 1,500 severe > 5,000 1,500 ~ 10,000 very severe > 10,000

The subsoils on west coast of Kaohsiung City are young sedi-ments with high water contents and low plasticity. Such soils can easily be softened once disturbed, or liquefied when sub-jected to steep hydraulic gradients and, therefore, failures of trenches were quite common during installation of diaphragm walls. In many cases, mini-piles and/or micro-piles were in-stalled to prevent trench from collapsing. Even so, necking still frequently occurred and reduced the sectional areas of dia-phragm walls. One row of CCP piles was indeed used along the perimeter of the area to be excavated for maintaining the stabil-ity of trenches prior to the installation of the diaphragm walls.

In some cases, the soil pockets were continuous from back to back of the walls and became water paths. Ground treatment had to be applied behind defective diaphragm walls to stop the in-gress of groundwater into pits.

The site was a salt pan a century ago. It was later used for fish farming for decades. Groundwater table was much lowered as fish farmers relied groundwater as supply because surface water was contaminated. Ingression of sea water increased the high chloride content in groundwater. It has been suspected that the quality of diaphragm walls may deteriorate as a result of chloride attack and, therefore, chemical contents are routinely

determined in major projects to check whether groundwater is aggressive. Table 2 shows the results of chemical tests on soils and Table 3 shows the results of tests on groundwater in bore-holes in the section of route between O1 and O2 Stations. The aggressiveness of chemical attack is classified as “severe” in ac-cordance with Table 4 which ….(ref. )

EntranceA

EntranceB

Zone A Zone B Zone C Zone D

O1 StationCut-and-cover

Tunnel

N

4 buildings (3-story)collapsed

Leakage

EntranceA

EntranceB

Zone A Zone B Zone C Zone D

O1 StationCut-and-cover

Tunnel

NN

4 buildings (3-story)collapsed

Leakage

Fig. 2 Layout of O1 Station, Kaohsiung MRT System

4 THE EVENT OF 9 AUGUST 2004

Water spurted from the bottom of the excavation in front of Panel S60M, refer to Fig. 2 for location, on the southern wall at 13:20 on 9 August 2004 as excavation reached a depth of 15m. A sinkhole of 3m maximum depth was formed behind the dia-phragm wall and the area affected was around 500m2. Four 4-story buildings collapsed in less than an hour and were demol-ished overnight. Several low-rise shops were severely damaged and were demolished sometime later.

01

2

3

4

5

6

15.78m

Ground Level

GL-19.85m

Diaphragm Wall

Zones A & Bt = 800mmL = 39m

Zones C & Dt = 1000mmL = 37.5m

GL-5.12m

01

2

3

4

5

6

15.78m

Ground Level

GL-19.85m

Diaphragm Wall

Zones A & Bt = 800mmL = 39m

Zones C & Dt = 1000mmL = 37.5m

GL-5.12m

Fig. 3 Excavation scheme and retaining system

3

JSG Piles

(#1, 3 and 5 were in place #4 was being installed)

1 2 3 4 5

S58M S60M

Bored piles 800mm diameter at 1m spacing

GL-1m to GL-30m

JSG piles 1200mm diameter at 1m spacing

GL-12m to GL-29m

JSG piles 1200mm diameter at 1m spacing

GL-12m to GL-29m

JSG Piles

(#1, 3 and 5 were in place #4 was being installed)

1 2 3 4 5

S58M S60M

Bored piles 800mm diameter at 1m spacing

GL-1m to GL-30m

JSG piles 1200mm diameter at 1m spacing

GL-12m to GL-29m

JSG piles 1200mm diameter at 1m spacing

GL-12m to GL-29m

Plan

SM

10<N<28

CL/ML20<N<27

CL

2<N<3

GL –15m

Diaphragm Wallt = 800mm to GL-39m

JSG Piles

GL –12m

GL –29m

Defective betweenGL-15.95m to GL-16.55

GL –15m

JSG PilesCL

SM

10<N<28

CL/ML20<N<27

CL

2<N<3

GL –15m

Diaphragm Wallt = 800mm to GL-39m

JSG Piles

GL –12m

GL –29m

Defective betweenGL-15.95m to GL-16.55

GL –15m

JSG PilesCL

Section Fig. 4 The event of 7 July 2005

One row of 11 bored piles was installed behind Panels S58M

to S60M. Pumping tests were performed in November 2004 to see if there were other defective panels. A total of 3,285 cubic meters of water was drawn from 6 wells and water levels inside the excavation closely monitored at 60 wells. The groundwater table inside the excavation dropped; on an average of 2.2m as a result of pumping. The recovery of water levels was monitored for more than 10 days, however, the desired purpose was not achieved as the locations of leakages could not be identified (Ho, et al., 2007).

5 THE EVENT OF 7 JULY 2005

To be on the safe side, one row of JSG piles was added along the entire perimeter of the station. Furthermore, the joints be-tween JSG piles were treated by using CCP piles. Pumping tests were again performed subsequently to confirm the effectiveness of these measures. The results were not satisfactory as the rate of recovery of groundwater inside the excavation was only slightly smaller than that obtained previously (Ho, et al., 2007). It was decided to add more JSG piles at the back of Panel S58M. Three new piles were installed without problem. As No. 4 pile, refer to Fig. 9, was installed on 7 July 2005, groundwater brought a large quantity of soil into the pit. A nearby hospital was endangered and the patients in the hospital were urgently evacuated as a precautionary measure. It how-ever survived with only minor damage. The sinkhole was about 1m in depth and settlement spread over an area of about

1,000m2.Subsoils at this site, being closer to the sea, are even worse than those at O2 Station so the incident was not a sur-prise.

6 FORENSIC STUDIES

Subsequent to the event of 9 August 2004, the contractor en-gaged Kaohsiung Professional Civil Engineers Association, Kaohsiung Professional Geotechnical Engineers Association and Kaohsiung Architects Association on 16 August to form a joint committee to investigate the causes of failures. After the event of 7 July 2005, Taiwan Construction Research Institute (TCRI) was engaged by the contractor in December 2005 to further in-vestigate the causes of failures.

6.1 Investigation by the Professional Associations

The committee studied the field records of diaphragm wall installation and checked the structural design of diaphragm walls. Upon the recommendation of the committee, additional soil exploration was carried out to determine ground conditions and tests were carried out to determine the quality of diaphragm walls. The design was found to be adequate and the failure was attributed to defects in wall panels.

Coring was carried out on Panels S59F, S60M and S61F in September 2005 to check the integrity of the walls. Panel S60M appeared to be in good quality within the depth of coring of 16m. Segregation of concrete was detected in S59F at depths of 15.95m to 16.55m. Coring was abandoned at Panel S61M as rebars were encountered at a depth of 6m. Laboratory tests were carried out on the cores obtained from Panels S59F and S60M and an average unconfined compressive strength of 58.7 MPa (with a standard deviation of 83.8 MPa) was obtained on the 11 specimens from Panel S59F and an average strength of 713.1 MPa was obtained on the 9 specimens from Panel S60M, while the design strength of concrete is ? MPa.

Chemical tests on the cores indicate chloride contents vary-ing from 0.016 to 0.022 kg/m3, less than the tolerance of 0.30 specified in CNS 3090 for ready-mixed concrete.

The thicknesses of Panels S60M and S61F and the locations of rebars were confirmed to be adequate by using radar. How-ever, Panel S59F was found to be defective.

6.2 Investigation by Taiwan Construction Research Institute

In addition to routine tests, such as CPT and SPT, TCRI carried out resistivity image profiling (RIP) of diaphragm walls and laboratory tests on soil samples taken by using GP-75S sampler which is a tool specifically developed for undisturbed sampling of difficult soils.

Cross-hole resistivity image profiling was carried out at Panel S72MF. Basically, electrical current is applied to the ground through two current electrodes and a electrical field is generated. The resulting voltage difference between two poten-tial electrodes is measured to determine the resistance of the ground. By moving the current electrodes and the potential electrodes, ground resistances can be measured for different situations. With the modern computer technology, the data ob-tained can be analyzed to obtain the resistivities of underground materials at various locations and the results can be can be pre-sented in a graphic form.

Table 4 shows the typical values of resistivities of different types of soils and rocks. As can be noted, the resistivities for concrete vary from 10,000 ~ 40,000 ohm-m, while those for

4

saturated sands, silts and clay vary from 15 to 1,000 ohm-m. The resistivity of sea water is as low as 1 to 5 ohm-m. It is therefore possible to identify defects, if any, in diaphragm walls accordingly.

Pole-pole array was adopted and 3 holes were sunk on each side of the panel as shown in Fig. 5.Sensors were installed at 1m intervals to a depth of 40m below ground surface. (more de-scription on operation)

Table 4 Resistivities of different material

Material Resistance, ρ ( Ω -m) Fresh water 10,000 Sea water 1 ~ 5 Coarse sand, gravel (dry) 20,000 ~ 80,000 Coarse sand, gravel (saturated) 1,000 ~ 5,000 Sand (dry) 5,000 ~ 20,000 Sand (saturated) 200 ~ 1,000 Silt (dry) 400 ~ 2,000 Silt (saturated) 30 ~ 200 Clay (saturated) 15 ~ 30 Sandstone 100 ~ 8,000 Granite, gneiss 7,000 ~ 15,000 Concrete, gabbro 10,000 ~ 40,000 Quartzite 5,000 ~ 10,000

One row of JSG piles were previously installed at the back of

the panel as shown in Figs. 6 and 7. To further improve the wa-tertightness of the wall system, CCP piles (not shown) were in-stalled at the joints to fill up the gaps, if any, between JSG piles. RIF was carried out between depth of 11m and 40m before the installation of CCP and again after the installation of CCP to see the differences made.

Fig. 8 shows the results obtained for a longitudinal section (XZ plan, at Y = ) of the wall panel and soil pockets are identi-fiable.

B1

1.8m

1.57m 1.8m

1.35m 1.2m 1.4m

1.7m 1.9m 1.95m

BH-1 BH-2 BH-3

BH-4 BH-5 BH-6

X

Y

S72MF S73M

2.2m

B1

1.8m

1.57m 1.8m

1.35m 1.2m 1.4m

1.7m 1.9m 1.95m

BH-1 BH-2 BH-3

BH-4 BH-5 BH-6

X

Y

S72MF S73M

2.2m

Fig. 5 Resistivity image profiling at Panel 72

S72

JSG

開挖區

Y

X

0

X=1.0

X=3.0

X=1.5X=2.0

X=2.5

X=3.5

Z

Y=1.0Y=1.5

Y=2.5Y=3.0 Y=2.0Y=3.5

Crosshole RIPCrosshole RIP

StationExcavation

S72

JSG

開挖區

Y

X

0

X=1.0

X=3.0

X=1.5X=2.0

X=2.5

X=3.5

Z

Y=1.0Y=1.5

Y=2.5Y=3.0 Y=2.0Y=3.5

Crosshole RIPCrosshole RIP

StationExcavation

Fig. 6 3D presentation of Resistivity Image Profileing

Crosshole RIP

JSG 12m

40m17m

Crosshole RIP

DiaphragmWall

Crosshole RIP

JSG 12m

40m17m

Crosshole RIP

DiaphragmWall

Fig. 7 Side view of Resistivity Image Profiling

JSG Gr out

GL

GL

GL

GL

Low resistivityindicating soil pocket with sea water

Low resistivityindicating soil pocket

Resistivity (ohm-m)10,00018030 0

JSG Gr out

GL

GL

GL

GL

Low resistivityindicating soil pocket with sea water

Low resistivityindicating soil pocket

JSG Gr out

GL

GL

GL

GL

Low resistivityindicating soil pocket with sea water

Low resistivityindicating soil pocket

Resistivity (ohm-m)10,00018030 0

Resistivity (ohm-m)10,00018030 0

Fig. 8 Interpretation of Results of Resistivity Image Profiling

5

X=1.0 X=1.5 X=2.0 X=2.5 X=3.0 X=3.5

X=1.0 X=1.5 X=2.0 X=2.5 X=3.0 X=3.5

X=1.0 X=1.5 X=2.0 X=2.5 X=3.0 X=3.5

X=1.0 X=1.5 X=2.0 X=2.5 X=3.0 X=3.5

Fig. 10 Results of Resistivity Image Profiling in YZ sections

Y=1.0 Y=1.5 Y=2.0 Y=2.5 Y=3.0 Y=3.5

Y=1.0 Y=1.5 Y=2.0 Y=2.5 Y=3.0 Y=3.5

Y=1.0 Y=1.5 Y=2.0 Y=2.5 Y=3.0 Y=3.5

Y=1.0 Y=1.5 Y=2.0 Y=2.5 Y=3.0 Y=3.5

Fig. 11 Results of Resistivity Image Profiling in XZ sections

After CCP Treatment

Pr evi ous JGP Tr eat ment

New CCP Tr eat ment

Resistivity (ohm-m)10,00018030 0

Before CCP Treatment After CCP Treatment

Pr evi ous JGP Tr eat ment

New CCP Tr eat ment

Resistivity (ohm-m)10,00018030 0

Resistivity (ohm-m)10,00018030 0

Before CCP Treatment

Fig. 12 3D presentation of results of Resistivity Image Profil-ing

Fig. 13 compares the results of RIP before and after CCP

treatment.

ACKNOWLEDGMENTS

The authors are grateful to Da-Cin/Shimizu Joint Venture for the authorization to publish the data presented herein.

REFERENCES

Ho, S. K., Chou, C. C., Chen, D., Chung, L-J, Liao, Z-B, Chen,

Y-Y (2007), The pumping tests at KMRT O1 Station, Proc., 2007 Cross-Strait Symposium on Geotechnical Engineering, April 15~17 (in Chinese)

CASE HISTORY 4

IMPORTANCE OF UNDERSTANDING LANDFORMS WHEN

DEVELOPING ON LANDSLIPPED TERRAIN R.P. Thimpson

Coffey Geotechnics Ltd INTRODUCTION The landscape in northern England, U.K, includes many areas where there are relict landslides which are a reflection of former glacial or periglacial conditions. Valley landslides were typically oversteepened by meltwaters and the rapid dropdown of glacial lakes following the breaching of ice or moraine dams. As the climate became warmer, stability of the hill slopes returned but was often only marginal. Increased demand for building land has led to areas of former landslipping being considered for development. However, the first stage is to recognise and understand that the landforms are actually relict landslides and that the margins of safety against renewed movement are modest. Before even embarking on an intrusive ground investigation, it is important to review both the geology and the landforms so that the possible, if not probable factors dictating the stability can be established. These include the depth to the plane of sliding, the shear strength operating on this plane, and the groundwater conditions. The sensitivity of the slope to renewed changes needs to be understood so that the appropriate range of hypotheses that could again trigger failure is identified. Two cases are described below in which development of adjacent areas of hillside was planned. Each had somewhat different landforms and therefore required separate failure hypotheses to ensure properly targeted ground investigations were conducted and suitable geotechnical works designed and constructed. Site A The first site was on east-facing hillside and was planned for multi-storey industrial development, which was to form a further part of an already developed location (ref1). No alternative sites were available. Published information (ref2) on the geology of the area indicated that it comprised inter-layered shaley mudstones and sandstones belonging to the Carboniferous period. A particularly distinctive marine fossil band, Gastrioceras cumbriense, was noted as being present in the relevant stratiagraphy. The site itself lay about 8km south of the ice sheet at its maximum glaciation. It is thought that a glacial lake had been formed which subsequently drained rapidly and caused landslipping of the valley side. The movement would have been on the shaley mudstone layers. As the water levels fell the reprofiled valley side adopted a stable condition again, albeit marginal. Concerning the topography of the area, the lowest part of the valley side had slopes of about 10 degrees which gradually rose to about 15 degrees. However the site for the development itself was backed by a mound with slopes of about 40 degrees which in turn was followed by a stretch of level ground before the main hillside rose at an angle of about 60 degrees. The stretch of level ground and the mound were tentatively classified as comparable to graben and horst structures. The graben was backed by faults with the main slope and with the back face of the horst while the

ground downslope of the front face if the horst was also landslipped terrain, see fig. 1. The development was to be sited in the last area, downslope of the horst. The proposed construction included a semi-basement which entailed excavation works that had a serious potentonal to create renewed landslip movement. It was therefore essential to gain the fullest understanding of the ground conditions and test the hypothesis that the location was a series of relict landslides. The initial step was to propose a ground model in which the depth to the plane of former sliding was comparable to the width of the graben i.e. the level stretch of ground (ref 3). A second step was to then propose that the mound or horst had undergone a planar slide on gently dipping plane consistent with the general dip of the strata. In addition it was recognised that it would be benefit if the marine fossil band could be detected. Once the hypothetical ground model for the area upslope of the site to be developed was established, the next stage would be to focus on the site itself so that it could be safely excavated and that suitable foundations could be designed. Boreholes were drilled firstly in the area of the supposed graben. The core revealed steeply inclined mudstone layers intermixed with more broken rock, consistent with slope movement. At a depth of nearly 20m, similar to the width of the graben, intact shaley mudstone rock was encountered which thereby demonstrated that the base of the graben had been reached and allowed location of the plane of former sliding to the established. Sandstone was encountered after a further 20m of drilling. The second part to be investigated was the horst block. This was found to comprise essentially intact layers of gently dipping shaley mudstone. Marginally above the plane of sliding, as determined from the investigation in the area of the graben block, a marine fossil band was recovered which was identified as being the Gastruiceras cynvruese band. Having found the fossil band, its plan position and elevation was measured at this location. By relying on the published value for the geological dip of the strata a search was made for possible evidence of the fossil band in the main intact hillside. Fortunately there was a deeply incised narrow gorge with fresh exposures of rock and the fossil band was able to be located. Its exact plan position and elevation in this second location was measured so that an accurate appreciation of the dip of the undisturbed strata could be determined. The third part that was targeted for the investigation was the site selected for the development and the area immediately downslope. With the knowledge that the fossil band was likely to be present and that it lay marginally above the plane of sliding , the drilling was conducted with great care. On the site itself, landslide debris was initially encountered and close to the base of this fossil band was again met and was directly underlain by mudstone with one thin layer, exhibiting a polished surfacing consistent with material that had been subject to former sliding. This was carefully sampled for subsequent laboratory testing. It was determined that the residual shear was between 11 degrees and 13 degrees, the slight variation being a reflection of testing under a range of effective stress conditions (ref 4). Immediately downslope the interface between the landslipped debris and the intact mudstone was at a slightly steeper grade. This raised particular cause for concern regarding its stability, noting that the toe of this slope had been cut away with only a low height retaining wall being in place at this location. By this stage the surface topography was known and the original hypothetical ground model was proven with the location and strength of material relating to the plane of

sliding established. The final main factor was information the ground water conditions. There were two possible sources. Firstly the underlying sandstone rock, which would be relatively permeable, could collect rainfall and run-off from the main hillside and feed this from below into the overlying mudstone, in effect giving rise to an unward flow and critical porewater pressures at the plane of sliding. Alternatively, the ground surface of the graben block was seen to become flooded after heavy rain. The water could then, possibly , drain down through the broken rock fragments forming the block until the water met the relatively low permeability intact mudstone at a depth of the plane of sliding. The water would then flow along this plane in preference to downwards through the intact mudstone. Targeted instrumentation and careful monitoring established that this later hypothesis was most likely to be correct. Furthermore it revealed that the downslope of the mound, on the site for the development itself, the phreatic surface could lie very close to the current ground surface. Stability analyses of the existing slope gave varying factors of safety with the area downslope of the proposed development site having only marginal stability, of the order of 1.1. The next stage was to consider the implications of excavating the site for the development. Again such works were found to give rise to unacceptably low margins of safety in respect of the stability of the mound which could potentially move on its former plane of sliding. Finally the weight of the proposed development would render unstable the downslope section of the slope. The design to address the various geotechnical problems was divided into three parts. On the uphill side of the proposed excavation it was decided to install a contigious bored pile wall with temporary ground anchors. For the permanent condition the piles would be propped by the base slab of the development. This part was designed to ensure that movement of the uphill side did not occur under any of the foreseeable temporary or permanent conditions. The next part concerned the foundations of the building itself. It was important that not only was adequate bearing capacity achieved ,but that the building weight did not impose load on the downslope area. Accordingly bored pile foundations were selected with the piles being founded below the plane of sliding. To increase stability of the downslope area itself, trench drains were designed to lower the ground water levels. By this means the factor of safety against instability was raised to at least 1.3 even under the most severe conditions. Conclusions for Site A i) The initial ground model and associated failure hypothesis were rigorously

researched and were based on a comparatively complex series of landslipped blocks.

ii) The location of the plane of sliding was provisionally estimated together with the groundwater conditions and formed the basis of a tentative model.

iii) Ground investigations and detailed forensic geotechnical examination of the data obtained confirmed the validity of the original model and failure hypothesis.

iv) With this information to hand, the appropriate geotechnical design involving an anchored bored pile wall, bored pile foundations and drainage trenches could be determined and implemented.

v) The case history A has shown that a comparatively complex landslip area can be effectively and safety developed provided the mechanism of the landslip is fully understood.

Site B The second site included a section of the access road to site A and was also on a hillside that included landforms consistent with relics landsliding. Similar geological conditions to site A existed with the solid strata comprising mudstone ad sandstone. The general topography of the area included a gentle convex toe section of the slope which then rose at a gradient of about 10 degrees, increasing to about 15 degrees before reaching a level section or plateau area at about 100m distant from the tone, see fig. 2. Beyond this the ground rose relatively steeply. Some excavation works in the toe area of the slope had been carried out already to allow constructon of the access road to be started. The contractor seemed to be peridiocally finding that after excavating a section on one day to a specified profile he was returning the next day only to find that he had to repeat the works to achieve the required cut profile. The problem was noted as being more marked in wet weather. Examination of the overall slope area revealed that fissures in the ground were appearing about 100m upslope of the toe at the junction between the top of the 15 degree slope and the plateau area. There was little doubt that the excavation works for the access road had triggered movement of a relict landslide. In addition it was the lower stage being disturbed by the construction works. However the detailed mechanism controlling the behaviour of the slope was not known. Two possible hypotheses were proposed. The first model involved a relatively shallow failure surface at about 3m depth with surface run-off ponding on the plateau area and then infiltrating the ground to give rise to periodic increases in the pore pressure on the failure surface. The second model included a failure surface at 6m to 7m depth with the pore pressures being controlled by variations in groundwater flow rates and therefore rainfall. Boreholes were drilled on the landslipped area in order to determine the actual ground conditions with depth and piezometers were installed so that the groundwater conditions could be recorded. The piezometers were also designed to act as basis slip indicators. In addition movement surface monitoring pins were placed spanning the zone at the edge of the landslipped surface so that magnitude and direction of movement between the landslip and adjacent stable ground could be established. Remedial solutions to stablise the movement were planned to be centred around the installation of drains to reduce the porewater pressures on the sliding plane. However until the ciritical plane and associated water pressures were known, it was not possible to proceed with the design of a solution let alone its implementation. One aspect that became apparent from the borehole investigation was that a geological fault diagonally traversed the centre of the landslip area so that mudstone rock was displaced downwards over the lower section of the landslip. At the fault interface, therefore, more permeable sandstone rock was juxtaposed against less permeable mudstone rock. The boreholes also revealed that non-solid deposits, being drift or landslipped material, extended to about 6m to 7m over most of the landslipped area increasing to about 10m depth near the head of the slip. However there was no horizon that could be conclusively determined as being the slip plane, other than possibly the interface between non-solid and the solid deposits.

A rockfill berm was placed at the toe of the slope to provide a modest improvement in the short-term stability of the slope. It was not until about nine months afterwards that there was a period of particularly heavy rainfall. The piezometers revealed that there was a marked increase in pore pressure at the interface of the non-solid and solid deposits. Futhermore the slip indicators in the piezometer tubes confirmed that movement was taking place on this horizon. The original ground model and its associated failure hypothesis were duly modified to embrace the information that had been established regarding the location of the movement plan, the presence of the faulting and that this fault dictated the critical porewater pressures. From this information, a 7m deep drainage trench was designed and constructed running diagonally across the slope and positioned immediately downslope of the line of the fault. This ensured that under very wet conditions the groundwater was drained away from the fault zone before it had the opportunity of reaching and disturbing the stability of the toe of the slope. With the toe stablised, the upper pat of the slope was effectively buttressed so that it likewise was rendered stable. Conclusions for Site B

i) The intial ground model and associated failure hypotheses were reasonably based on the site being a straightforward two-stage relict landslip.

ii) The details relating the plane of sliding and critical porewater pressures were not known but two tentative models were proposed.

iii) Ground investigations revealed that the original ground model required modification in order to properly understand the behaviour of the landslip.

iv) With this information to hand, appropriate drainage works to address the critical porewater pressures could be determined.

v) The tentative drainage solutions based on either of the original modesls would have been seriously flawed and would not have provided stability to the landslip.

vi) The case history B has shown that an apparently relatively straightforward landslip may, in fact, be complex and that rigorous forensic geotechnical engineering is essential to gain a correct understanding.

Overall Conclusions

1. The examples for the two sites, which concerned relict landslides, demonstrate that it is essential from the outset to develop well thought out and researched failure hypotheses together with ground models and that this should be done prior to intrusive ground investigations being carried out.

2. On the basis of the hypotheses, the ground investigations can be target to give the optimum results. It is essential to review and modify, if necessary, the initially chosen hypothesis in the light of the findings from the investigations. Provided this is achieved, the geotechnical aspects of a development can be designed and implemented with known levels of safety.

3. One of the sites, Site A, was complex but the rigorous forensic geotechnical examination of the findings from the borings and field exposures allowed the original ground model and associated failure hypothesis to be confirmed.

4. The other site, site B, appeared to be straightforward two-stage landslide. However the two originally proposed possible failure hypotheses were found not to match the actual ground conditions as revealed by intrusive investigations. It demonstrated the requirement to carefully examine the information obtained and then to modify the original hypothesis accordingly.

5. From the examples it is evident that if proper failure hypotheses and targeted ground investigations are not conducted, the subsequent development could lead to renewed slope movement. This is exemplified, in particular, by the initial works that were carried out at site B.

References

1. Leach B.A. and Thomson R.P. (1987), The influence of underdrainage of the stability of relic landslides. 9th ECSMFE Dublin Vol.1pp 447-450, Balkema

2. The Geology of the Country around Huddersfield and Halifax (1930), Explanation of Eological Sheet 77, Memoirs of the Geologifcal Survey of England & Wales. DSIR His Majesty’s Stationary Ofrice London.

3. Holmes A. (1965), Principles of Physical Geology. Second Edition pp 1046-1051, Nelson.

4. Thompson R.P (1991) Stabilisation of a landslide on Etruia Marl.International Conference on Slope Stability, Isle of Wight UK. Pp 403-408. The Institution of Civil Engineers. Thomas Telford.

1/7

Case Studies 5

Monitoring in Forensic Geotechnical Engineering - Dam Construction –

- Yoshi Iwasaki, Ph.D., Dr of Eng, P.E.

Managing Director Geo-Research Institute, Osaka

1 Observational Procedure The text book of “Soil Mechanics in Engineering Practice, Second Edition” by K. Terzaghi and R.Peck(1967) added Chapter 12 Performance Observations to the first edition. It was the first to point the enrolment of the monitoring soil and structural performance during construction work as to provide evidence in lawsuits. It states that a fair decision is to be expected only if the causes and the nature of the mishap are known. If the contractor or owner can prove that he has anticipated the undesirable condition, has observed its progress during construction, and has done everything possible to avoid it, he is in a much more favorable position than if the condition has taken him by surprise. The element of surprise not only injures his professional reputation, but it may also injure his financial standing. When the third edition by the authors of K.Terzaghi, R.Peck, and G.Mesri was published, the Chapter 12 in the second edition was deleted and refers observational procedure only in a single paragraph of “10.5 Observations during Construction,” where no legal aspects was mentioned for usage of monitoring.

2 Enrolment of observation in forensic geotechnical In the third edition of “Soil Mechanics in Engineering Practice”, it is stated that “in earth-dam and foundation engineering the permanent structures are designed before the construction operations start, and the consequences of unanticipated sources of trouble do not appear until the structure is in an advanced state of construction or is in service.” This statement sounds observational method is not useful for earth dam and foundation engineering. The author will review several case histories that are constructed with or without monitoring including forensic problems and try to consider any meaning of monitoring in dam construction. In very early stage of dam construction, St.Francis Dam, Los Angeles County, California, U.S.A. was constructed as a concrete gravity dam in March 1926. During the

2/7

initial filling, cracks in the dam began to appear and was tried to seal the cracks. During the final stage of the second filling in March 1928, new leakage was found and the dam was failed. The St.Francis Dam is considered the greatest American civil engineering failure of the twentieth century. Approximately 450 mostly sleeping people died by flood caused by the failure of the dam. When the dam was constructed, dam engineering was just to start to develop based upon these failures and lessons learnt. Another example is Vajont dam, Italy that was a concrete arch dam constructed in 1962 and failed in 1963.

2.1 Vajont Dam Vajont Dam is a dam completed in 1961 under Monte Toc, 100km north of Venice, Italy. It was one of the highest concrete dams in the world measuring 262m in height with 27m in base thickness and 3.4m in top thickness. Shifting of rock was noticed during the first filling in February 1960 before final completion of the dam of September 1960. By March 1960, the level of the reservoir reached 130 above the level of the river, when the first small slide was observed.

Table-1 The Sequence of Filling/Draw Down leading to Failure

date water level

movement rate

mass movement/land slide

1960/2 1960/3 130m1960/10 170m 3.5cm/day potential huge slide area recognized

Filling 1960/11/4 180m 0.7millon slid 185m 8cm/day

The First

Stage Draw Down

135m 0.1cm/day

1961/10 135m 1962/2 185m

Filling

1962/11 235m 1.2cm/day1962/11 1cm/day

The Second Stage Draw Down

1663/2 185m 0cm/day

1663/3 185m 1993/5 231m 0.3cm/day1993/6 237m 0.4cm/day1993/7 240m 0.5cm/day1993/8 240m 0.8cm/day

Filling

1993/9 245m 3.5cm/day1993/10 245m

1993/10/9 235m 20cm/day

The Third

Stage Draw Down

1993/10/9 22:38 heavy rain 260million m3 slid

3/7

In October 1960, the level reached 170m, a rapid increase in the rate of displacement to approximately 3.5cm/day was observed. At this time, a huge joint of 2km in length was found to open up in an area of slope 1700m long and 1000m wide, suggesting very large slide was mobilized. On November 4, 1960, there took place a land slide of 700,000 cubic meters at the level of 180m. The water behind the dam was lowered to 135m and the movement was reduced to 1mm/day. The safety of the slope of the dam was discussed and calculations ordered by management showed that catastrophic failure was unlikely and the reservoir allowed to refill water under controlled monitoring. Thus the reservoir was filled and slowly emptied three times. On October 9, 1963 at 10:38 pm, the combination of third drawing down of the reservoir and heavy rains triggered a big landslide of about 260million cubic meters of forest, earth, and rock, which fell into the reservoir. The resulting splashed water caused some 50million cube meters of water to push up the opposite bank and destroyed the Village Casso, 260m above the lake level before overtopping the concrete dam of 250m in height. The water, estimated about 30million m3 then fell more than 500m onto the downstream villages of Longarone, Pirago, Villanova, Rivalta, and Fae destroying them and killed some 2500 persons. The court condemned eight persons in guilty not to take any action of evacuation of the inhabitants. In the case of Vajon dam, monitoring was being carried out, however, it was very difficult to predict the failure of the slope.

2.2 Teton Dam Teton dam constructed and failed in1976, Idaho, U.S.A. is another case that has no monitoring except measurement of deformation.

Teton dam, Idaho, the major feature of the multipurpose Teton Basin Project of the United States Bureau of Reclamation, failed on June 5, 1976 while the reservoir level was 1m below the spillway sill. The dam was a center core zoned earth fill structure, with a height of 93m above the river bed and 123.4m above the lowest foundation. The instrumentation of Teton dam was not prepared because performance records were considered to be available for dams “constructed of similar material and on similar foundations” and only settlement monuments were provided.

2.2 Takase Dam

4/7

Takase dam is a central core earth-rock dam, 176m in height, 362m in crest length built in 1978 by Tokyo Electric Power Company Inc. as a part of Takase River Hydraulic Power Development Project developed in Takase Basin, central Japan.

Fig.1 Pore pressure measurements at Takase dam during initial filling (1)

5/7

Fig.2 Pore pressure measurements at Takase dam during initial filling (2)

Fig.3 Change of equi-pore water pressure line during the first filling (Dec.,’78, Feb.,79, Apr.,79, and July, 79)

Instrumentation of Takase dam was arranged to provide pore pressure, earth pressure, displacement in dam section, and surface targets for measurement of surface movements. In placing the pore pressure gauges for control of the dam construction work, it was decided to concentrate mainly on the core and transition zones, where comparatively

6/7

high pore pressure was expected to develop, and also to a central extent on the inner shell zone, which has a relatively high fine grain content. Pore pressure gauges for dam security control during and filling of the reservoir were located on the core contact area and within the foundation bedrock, and also in positions that enable tracking of the downstream phreatic line. Fig.1 and 2 show changes of pore water pressure during the initial filling water in the reservoir. The filling started in December, 1978 and reached nearly completion in July, 1979. Each figure shows the measured pressure at three different heights within the embankment of EL=+1107m, EL=+1165m, and EL=1135m. Fig.3 shows change of equi-contour line during the filling (Dec,’78(left top)-Feb., 79(left bottom, Apr.,79(right top), and(July,79(right bottom )). It is seen that the distribution of the pore pressure in the field is sound condition from which no leakage is expected.

2.3 Dams that were saved from failure based upon monitoring ASCE Task committee on instrumentation and monitoring dam performance summarized various case histories that showed saved dams from catastrophic failures based upon instrumentation and monitoring. The following three examples are from the case studies shown in the guidelines for Instrumentation and Measurements for Monitoring Dam Performance. Findlay(1988) reports a dam for a hydroelectric project in 1986 included the installation of instrumentation and monitoring system. Pneumatic piezometers were installed to monitor hydrostatic pressure during and following the initial filling. It was found the reading of two sensors increased equivalent with rising water elevation. Piping might have occurred if the head condition were continued. The reservoir was immediately dewatered. The pore water gages responded with the same amount of the dewatered level. The two piezpmeters were installed near a contraction joint. Modification was made for inadequate sealing and a safe condition was achieved. Only after one year of a dam of USBR(Fontenelle Dam) in 1965, piping of the embankment and surface erosion caused a partial failure. Damage to the 36.6m high, 1626m long rolled-earth embankment was repaired and the dam put back in service. In 1982, instrumentation and monitoring indicated the presence of new seepage. A cut-off wall of about 60cm thick and 48m deep along the whole embankment of about 1.6km was installed to stop the leakage. (ENR,Nov.26, 1987) Navajo Dam of U.S. Bureau Reclamation was detected seepage problem during the initial filling in 1963. Seepage reading of the dam had been increased until the installation of a 100cm thick, 120m deep, 135m long cutoff wall.(Civil Engineering,

7/7

March 1987) It is clear that the observational procedure has been applied successfully in dam constructions based upon monitoring of the dam behavior during the first filing as the final stage of construction or even later phase under service stage.

3 Conclusions If you search for “Teton Dam,” through internet Home Page, you will find hot discussions still going on the true reason that made the dam failed. If the performance was monitored, the discussion might be more fruitful and practical based upon real data. Failure of Teton Dam clearly shows the measurement of power pressure within dam section is necessary and if it had been measured, the failure of the dam itself had been prevented. Vajon dam shows inadequate theory and practice of the slope failure in geotechnical engineering. The statement in the textbook of “Soil Mechanics in Engineering Practice” the consequences of unanticipated sources of trouble do not appear until the structure is in an advanced state of construction or is in service” should be replaced by “the construction of a dam continues until the initial filling water and should be confirmed as assumption by design or ready to be modified based upon the monitoring results during the filling water in the dam.

References: W.L.Chadwick(1977), Case Study of Teton Dam and Its Failure, Proceedings of the Ninth International Conference on Soil Mechanics and Foundation Engineering, Case Study Volume, Tokyo, Japanese Society of Soil Mechanics and Foundation Engineering, pp431-455 R.Takai, T.Iwakata, and Y.Miyata (1977), Results of Soil Tests and Measurements during and after Construction of the Takase Dam, Proceedings of the Ninth International Conference on Soil Mechanics and Foundation Engineering, Case Study Volume, Tokyo, Japanese Society of Soil Mechanics and Foundation Engineering, pp497-554 ASCE Task Committee(2000)”Guidelines for Instrumentation and Measurements for Monitoring Dam Performance” Findlay, R.Craig (1998), “Hydrostatic Pressure at a Soil-Structure Interface,” Proceedings of the Second International Conference on Case Histories in Geotchnical Engineering, University of Missouri-Rolla, St.Louis

Case Study 7

FORENSIC ANALYSIS ON RE WALL DISTRESS: A CASE STUDY

Sivakumar Babu, G. L.1, Murthy, B. R. S.2 and Singh, V. P.3

1. Associate Professor, Dept. of Civil Engg., IISc, Bangalore, India, 560 012, gls@civil.iisc.ernet.in

2. Retd. Professor, IISc, Bangalore, India, 560 012, murthybrs@gmail.com 3. Research Associate, Dept. of Civil Engg., IISc, Bangalore, India, 560 012, vikasps@civil.iisc.ernet.in

ABSTRACT

The article presents a forensic analysis of the observed distress of an RE wall. The investigation involved field observations, extensive laboratory testing of backfill soil properties and reinforcement material (i.e. geogrids), analysis of failure cause and justification for the soil nailing technique as a rehabilitation measure. The article also highlights soil nailing technique as a viable alternative for rehabilitation of failed reinforced earth (RE) structures.

1. INTRODUCTION

Forensic geotechnical engineering (FGE) may be defined as the discipline that prepares geotechnical engineers to investigate failures of structures such as embankments, pavements, foundations and earth retaining systems, in purview of professional ethics and legal bindings. FGE being a relatively new terminology in geotechnical engineering profession, it is necessary to create general awareness and understanding about the subject among geotechnical engineers, clients, and contractors.

In this article, a case study of RE walls constructed for a flyover is presented wherein a

detailed investigation was sought by the client with regard to the examination of the distress noted during the construction. The methodology adopted to carry out detailed investigation involved following sequence of steps.

1. Survey and characterization of the distress i.e. site visit and observations. 2. Diagnostic tests i.e. collection of material (backfill and geogrid) samples for laboratory testing to

evaluate the behavior of soil and the strength characteristic of geogrids used. 3. Development of most probable failure hypothesis and cross check with original design. 4. Back analysis of RE walls stability using conventional methods, and rigorous numerical analysis

using computational code. 5. Suggestion and justification for the rehabilitation measure in the form of driven soil nails.

2. SURVEY AND CHARACTERIZATION OF DISTRESS

Site visits were made to (a) identify and study locations of distress, (b) collect the samples of backfill and geogrids for laboratory testing, and (c) procure a preliminary report on the possible cause of distress from the client/contractors. For the purpose of identification, the two ends of

flyover were named as the Siliguri end and the Dalkola end. The construction work of the RE wall started in June 2006 and the Siliguri end of the flyover was completed in April 2007 and the Dalkola end by August 2007. During June –October 2007, the area experienced severe rains which were of unprecedented magnitude causing distress in the RE wall that was being constructed. On the Siliguri end, distresses in RE wall were noted in the following forms: i) damage to reinforced earthwork ii) damage near abutment leading to misalignment of panels, and iii) sand flow was noticed and one of the panels got dislocated in the region near the design height of 8 to 8.8 m. On the Dalkola end, the client informed that the sand removal led to overstressing of the connection, resulting in the breakage of grid and panels. A few photographs of the distress locations in the RE walls taken during the site visits are shown in Figures 1(a)-(c).

Figure 1(a): Damage near flyover abutment

Figure 1(b): Bulging of facing panel

Figure 1(c): Damage near top facing panel and the approach slab

3. DIAGNOSTIC TESTS ON MATERIAL SAMPLES

3.1 Test Results of Backfill Soil Samples

Backfill soil samples collected during site visits were tested in the laboratory and used to obtain shear strength parameters in both the undisturbed and saturated conditions. Table 1 shows the summary of the shear strength parameters obtained from laboratory tests on backfill soil samples.

Table 1: Shear strength parameters of backfill

Soil state Bulk density (kN/m3)

Peak friction angle (degrees)

Residual friction angle (degrees)

Cohesion (kPa)

In situ condition

17.50

40.0

35.1

0

In situ condition

16.08

39.1

33.2

0

Saturated condition

17.50

30.54

29.8

0

It is noted that the shear strength parameters for the in situ condition are higher than the design value of 32o. Tests on saturated sand indicate that there is a marginal decrease in the friction angle. The permeability coefficient of the backfill corresponding to the bulk density of 17.50 kN/m3

is in the range of 0.089 mm/sec. It is likely that the rate of infiltration during the intense rainfall was higher than this value.

3.2 Strength and Permittivity Characteristics of Geogrids Samples

The permittivity of the geogrid (geotextile) used is 1.85 sec-1 and the thickness of the geogrid is 1.68 mm which corresponds to permeability coefficient of 3.1 mm/sec. For proper filter behavior, as per NCMA (1996) guidelines, the geogrid permeability coefficient needs to be not less than 10 times higher than the permeability coefficient of the backfill. In this case, the geogrid is highly permeable and the factor of safety with respect to filter criteria is in the range of 3.48 (3.1/10 x 0.089) and hence, the geogrid used was satisfactory and a separate drainage medium behind the facing panels was not essential. Further, the design strength and other properties of geogrid samples were assessed by another agency and the result supplied by them showed that the used geogrids satisfied their technical specifications.

4. DEVELOPMENT OF MOST PROBABLE FAILURE HYPOTHESIS

Rainfall leads to saturation of the backfill and/or pore water pressure development which needs to be addressed properly. The effect of rainfall on reinforced wall can be understood from design charts for reinforced soil slopes for dry and saturated conditions developed by Jewell (1990). Figures 2(a) and (b) show the minimum required force coefficient (kreq) for reinforced soil structures with slope angles varying from 0 to 90 degrees. The state of pore water pressure is expressed in terms of pore pressure coefficient ru = u/γz, where u is the pore pressure, γ is backfill bulk density and z is the depth at the point of considerations of pore water pressure (u) and vertical stress (γz). The pore pressure coefficient varies from 0 to 0.5 and dry/fully drained backfill has ru = 0, whereas presence of saturated water in the backfill is indicated by ru = 0.5.

It is expected that minimum required force expressed in terms of (kreq) in the reinforcement

for equilibrium is lesser for fully drained backfill. From Figures 2(a) and (b), it can be noted that for fully drained backfill, for the slope angle 900 representing wall, for design friction angle of 320, the kreq value is 0.31. This value increases to values in the range of 0.60 when pore pressure in the backfill exists. The pore pressure generated in the backfill in a momentary condition as a result of heavy downpour and is not considered in design. Thus, when rain of unpredicted magnitude occurs and if the infiltration rate is higher than rate of discharge, seepage pressures build up and momentarily induce additional pressure on the retaining system. The local bulging observed in a few locations in the RE walls of the flyover is attributed to the above phenomenon.

Further, for a completely covered surface in the plan area of the wall, percolation of water is

not allowed. For this condition, using design friction angle of 32o, the minimum required force coefficient (kreq) for reinforced soil structure of 90o (vertical wall) is 0.31 and the corresponding force required to be supplied by the reinforcement at a depth of 10.40m is (kreq)γH is 0.31x17.5x10.4 = 56.42 kN/m. It is noted that the geogrids provided at this depth have higher strength than required value of 56.42 kN/m. However, under the influence of pore pressures, the

minimum required force coefficient (kreq) is in the range of 0.64 and hence, the force momentarily can increase to a maximum value of 116 kN/m, applying considerable force on the reinforcement

and panels which could have contributed to the distress noted.

It is also noted that for the fourth facing panel from the top, two types of geogrids (SS120 and SS150) which have different stress-deformation characteristics were used. The tensile force

required for stability is marginally in the same range whereas elongations required to obtain the same force in each geogrid layer are different. This could have contributed to marginal drift in the facing panels under the influence of earth pressures.

Figure 2: Jewell (1990) design charts for reinforced slopes (a) ru = 0; (b) ru = 0.5.

5. BACK ANALYSES OF RE WALLS STABILITY

The back analyses to examine the stability of the geogrid reinforced retaining walls was carried out using conventional method (FHWA 2001) and by performing rigorous finite element analysis using a computational code (Plaxis 2006). Various sections of the RE walls with different design height of wall (e.g., 10.40 m, 9.60 m and 8.80 m) were analysed separately. In the present paper, results of the analyses corresponding to the maximum design height (i.e. 10.40 m) are presented and discussed. Table 2 shows the material properties and other parameters adopted for the study.

Table 2: Material properties considered for analysis

Parameter Value Design height of the wall, H (m) 10.40 Unit weight of wallfill, γ (kN/m3) 18 Angle of internal friction of wallfill, φ (degree) 32 Surcharge load, ws (kPa) 22 Length of geogrids member, L (m) 7.80 (7.20) Vertical spacing of geogrids members, SV (m) 0.80 (0.365) Geogrid-soil interaction coefficient, μ 0.85 Design strength of geogrid members As per the data supplied by the client Note: Figures in bracket are for the bottom most layer of geogrid.

It is to be noted that the properties of foundation and backfill are assumed to be same as that of wallfill. Tables 4 and 5 shows the typical results. Figures 1 and 2 shows the finite element models of the RE wall of 10.40 m design height with and without use of soil nailing as remedial measure respectively.

5.1 Results of Conventional Analysis

RE wall is found safe against overturning failure with minimum eccentricity of 0.85 mm

which less than the permissible vale of 1.30 mm (i.e. L/6). Also, wall is safe against sliding failure having a minimum factor of safety = 2.50 which is more than the minimum recommended value (i.e. 1.50). Further, no evidence of bearing capacity failure were reported or found during site visit. Thus, it can be concluded that the RE wall is externally stable. Table 3 shows the result of the internal stability analyses. According to FHWA (2001), minimum recommended values for factor of safety against pullout FSP and tensile failure FST of the reinforcement are 1.50 and 1.30, respectively. It can be seen (indicated in bold) from Table 3 that the FSP values for the top four geogrid layers and FST values for the bottom four geogrid layers are less than the minimum recommended values.

Table 3: Results of conventional analyses for section with design height 10.40 m

Geogrid layer number ( from

top )

Depth z (m)

Design tensile strength of geogrid, Pd (kN/m)

Factor safety against

pullout, FSP

Factor of safety against tensile failure,

FST

1 0.44 29.70 0.81 4.36 2 1.24 29.70 0.99 2.85 3 2.04 39.60 1.19 2.79 4 2.84 39.60 1.41 2.17 5 3.64 59.50 1.64 2.62 6 4.44 59.50 1.91 2.17 7 5.24 59.50 2.19 1.82 8 6.04 74.40 2.51 1.93 9 6.84 74.40 2.85 1.66

10 7.64 74.40 3.24 1.43 11 8.44 74.40 3.68 1.24 12 9.24 74.40 4.16 1.09 13 10.04 74.40 4.35 0.90

5.2 Results of Finite Element Analysis

As mentioned earlier, the section of RE wall with design 10.40 m was simulated using finite element computational code Plaxis (2006). Mohr-Coulomb model was used to represent soil behaviour and geogrid layers were simulated using ‘Geogrid’ structural elements (yellow color) available in the code. Details on simulation procedure can be found in the user manual of the computational code. Figure 3 shows a typical example of numerically simulated RE wall.

Figure 3: Typical example of numerically simulated RE wall.

The results of the finite element analysis are shown in Table 4. From Table 4, it can be seen that the bottom geogrid layers are not safe against tensile failure. Further, the maximum displacement of the wall is observed as 26.53 mm.

Table 4: Results of finite element analysis for section with design height 10.40 m

Geogrid layer (from top)

Depth z

(m)

Design strength of geogrid, PD

(kN/m)

Axial force developed in each member

(kN/m)

Horizontal displacement

(mm)

FST

1 0.44 29.70 16.34 26.53 1.82 2 1.24 29.70 16.56 26.44 1.79 3 2.04 39.60 12.21 26.42 3.24 4 2.84 39.60 16.52 26.39 2.40 5 3.64 59.50 19.20 26.36 3.10 6 4.44 59.50 18.99 26.32 3.13 7 5.24 59.50 22.36 26.27 2.66 8 6.04 74.40 32.05 26.22 2.32 9 6.84 74.40 37.06 26.16 2.01

10 7.64 74.40 46.30 26.10 1.61 11 8.44 74.40 60.73 26.04 1.23 12 9.24 74.40 78.99 25.98 0.94 13 10.04 74.40 133.59 25.94 0.56

5.3 General Observation from Back Analyses

From Tables 3 and 4, it is evident that in the bottom layers of geogrids, the axial force developed is more than design tensile strength of geogrids; at depths beyond 6m as the factor of safety is close to 2 or less. The height of the wall above the ground level is in the range of 8m. Hence, in this range it is preferable to increase the factor of safety with respect to tension, and, it is recommended that soil nailing could be employed.

6. SOIL NAILING AS REHABILATION MEASURE

FHWA (2003) may be referred for detailed information on the analysis, design and construction of soil nail walls. Figure 4 shows the numerically simulated model of RE wall with soil nails. Table 5 shows the results of finite element analyses of the RE wall, additionally reinforced with 25 mm diameter horizontally driven nails of length 7.80 m spaced at 0.80 m vertically between geogrid layers and 1.00 m horizontally along the length of the wall. From Table 5, it is evident that improvement in the values of factor of safety against tensile failure FST is observed. For the bottom most geogrid layer FST value has increased from 0.56 (for geogrids alone) to about 1.11 (for geogrids along with soil nails) which almost two times improvement. It is also observed that maximum axial force developed in the bottom most soil nail is about 58.08 kN which yields a factor of safety value against nail tensile failure FST equal to 3.53. Hence, use of soil nailing significantly improves the stability of RE wall and is therefore, suggested as the possible remedial measure for short term and long term stability of RE wall. Soil nails may be provided where the factor of safety with respect to tensile forces is less than 2 and it can be noted from Table 4 that soil nailing is required at the top as well as bottom. It may also be noted that the wall height above the ground level is 8m. Hence, the results presented in Table 5 needs to be considered for 8 m depth only.

Figure 4: Numerically simulated RE wall with soil nails (black elements indicate soil nails).

Table 5: Results of finite element analysis of RE Walls with soil nails*

Geogrid layer (from top)

Depth z

(m)

Design strength of geogrid, PD

(kN/m)

Axial force developed in each member

(kN/m)

Horizontal displacement

(mm)

FST

1 0.44 29.70 8.57 13.07 3.46 2 1.24 29.70 9.12 12.86 3.25 3 2.04 39.60 6.38 12.75 6.20 4 2.84 39.60 7.07 12.69 5.60 5 3.64 59.50 7.50 12.65 7.93 6 4.44 59.50 6.00 12.63 9.91 7 5.24 59.50 6.72 12.61 8.85 8 6.04 74.40 10.42 12.58 7.14 9 6.84 74.40 13.87 12.55 5.36

10 7.64 74.40 18.96 12.53 3.92 11 8.44 74.40 25.15 12.50 2.95 12 9.24 74.40 35.04 12.48 2.12 13 10.04 74.40 67.12 12.47 1.11

* - 25 mm diameter driven nails of length 7.80 m spaced at 0.80 m vertically between geogrid layers and 1.00 m horizontally.

6.1 Advantages of Using Soil Nails as Rehabilitation Measure

As indicated from the results of the numerical analyses of RE wall with soil nails (see Table 5), the contribution of soil nails in improving the stability of the RE is found to significant. Introduction of soil nails helped in the aspects:

1. Redistribution of axial forces among soil nails and geogrid layers and thus improving the tensile

capacity of the original reinforcement (i.e. geogrids). 2. Reduction in net force at facing connection thereby improving connection stability. 3. Reduction in the lateral displacement of the RE wall to almost 50%. 4. Improvement in the overall stability of the RE wall against any unforeseen event such as

unprecedented heavy rainfall.

7. FINAL RECOMMENDATIONS AND PRECAUTIONARY MEASURES

The analysis shows that the unprecedented rains during construction (temporary phenomenon) resulted in very high lateral pressures contributing to minor movement of panels in the area. The geotextile provided for filtration was adequate. The backfill properties measured from undisturbed samples were satisfactory. However, it is essential that proper drainage measures should be implemented so that the backfill is not saturated during the remaining part of construction.

Soil nailing to improve the stability of RE wall is recommended as the rehabilitation measure. Driven soil nails of 20 mm diameter and length 5.6 m are suggested near the Siliguri end abutment. In general, nail lengths of 0.7H may be provided at locations of distress. The height shall be reckoned as the height above the ground level. Two nails per facing panel may be provided when the height exceeds 6 m. It is to be noted that the above recommendation about the provision of soil nails is based on the similar analysis performed for other sections of RE wall with smaller heights, keeping in view the cost of rehabilitation process, and ease of application.

7.1 Suggested Precautionary Measures

1. It is essential that the nailing process should not leave the facing ugly. 2. As noted during the field observations, gap was not maintained between the top of the facing

panel and the approach slab, creating additional loads on the facing panels. This contributed to movement of the facing panel. It is suggested that gap may be created such that the stress transfer is avoided.

3. It is reiterated that normal conditions considered in design shall be ensured during execution. 4. Unexpected loads in the form of seepage/pore pressures, loading on the facing from approach

slab shall be avoided. 5. It is advisable that during impending monsoon, the embankment area be covered properly and

handle local problems if any from outside only.

8. CONCLUDING REMARKS

The article presents a case study of failed RE wall, wherein, a through investigation about the cause of distress observed during the construction of RE walls was sought by the client. Although, this study do not addresses the legal issues related to the investigation, it attempts to describe the general procedure that may be followed in the forensic geotechnical engineering.

9. REFERECNES

FHWA. (2001). Mechanically Stabilized Earth Walls and Reinforced Soil Slopes: Design and Construction Guidelines. Report FHWA-NH1-00-043, U. S. Department of Transportation, Federal Highway Administration, Washington D. C.

FHWA. (2003). Geotechnical Engineering Circular No. 7 - Soil Nail Walls. Report FHWA0-IF-03- 017, U. S. Department of Transportation, Federal Highway Administration, Washington D. C.

Jewell, R. A. (1990). Revised design charts for steep reinforced slopes. Proceedings of the Symposium on Reinforced Embankments: Theory and Practice, Thomas Telford, 1-30.

NCMA. (1996). Design Manual for Segmental Retaining Walls. National Concrete Masonry Association, 2nd edn, NCMA, Herndon, VA.

Plaxis (2006). Reference Manual. Delft University of Technology & PLAXIS B.V., The Netherlands.

Case Study 8 DISTRESS TO REINFORCED EARTH WALL EMBANKMENT: A CASE STUDY

Prof. T.G. SITHARAM Professor and Chairman, PMG,

Prof. K.S. SUBBA RAO

Professor (retd) Department of Civil Engineering Indian Institute ofScience

Bangalore – 560012

SUMMARY

In this case study, the distress to the Road over bridge (ROB) with a 6 lane highway

(about 30m wide) constructed using reinforced earth wall technique with a maximum height

of about 11m is discussed. A view of constructed RE wall is shown in Fig1. The longitudinal

cracks as seen on the ROB-II are shown in Figure 2. The typical cross-sectional details of

ROB are given in Fig 3. After the completion of the ROB-II, it was opened for traffic on

15.07.2007. During inspection on 20th September 2007, some longitudinal cracks were noted

on the pavement parallel to the reinforced earth wall face (As in Fig 2) on both Right

Carriage Way (RCW) and Left Carriage way (LCW) of the road between chainages 195.0m

and 390.0m (see Fig 4). On both sides of the road cracks have appeared just at the end of the

GI mesh reinforcements of the RE wall which is at a distance of 5.0 to 8.5 m from the

Reinforced earth wall panels. The maximum width of crack was mostly 1 to 3mm and at one

place it was 9mm. Immediately, the longitudinal cracks were filled up and rolled. At that

time, some cavities have also been observed as shown in fig 4. The asphalt filling up which

had been done on the observed cracks showed partial success in that the crack on the left

carriage way from chainage 195 to 295 did not reappear. But the one on the right carriage

way from 295 to 390 did reappear. In addition, on the 19th of October, 2007 new hair line

cracks were observed on the left carriage way and the right carriage way as indicated in the

Figure 2. These hair line cracks were seen to be confined to the top bituminous concrete

layer, however extending upto Ch 405 on LCW and Ch +205 on RCW.

The observation on the verticality of the end panels were also recorded from

26.09.2007 to 15.10.2007. The recorded movements during the above mentioned period

remained fairly stable and showed no evidence of perceptible distress propagation. The

movements mostly are in the range of 5 to 30 mm which are generally expected in RE walls

for mobilization of shear forces on the soil-reinforcement interfaces.

In order to get some additional information on the compacted state of the backfill,

geotechnical investigations at three more locations in the ROB section of the carriage ways

were suggested and carried out.

Fig 1. View of RE wall which is constructed for ROB

Fig 2. Longitudinal cracks on the carriage way

Fig 3. Typical cross section of the ROB Fig 4. Cavity observed on RCW

Fig 5. Longitudinal cracks as seen on both RCW and LCW of ROB –II

Most of the boreholes in the reinforced earth wall embankment section show a good

penetration resistance with depth but only in one o the borehole (which is close to the

longitudinal crack on the right carriage way) shows some evidence of loose pockets and with

very low penetration resistance (at a depth of 4.5m from the road surface) and the presence of

boulders (between depths of 5 to 8m from the road surface). The results of ground penetrating

radar survey indicate the presence of loose pockets and cavities in the top layers of

compacted fill.

Design of RE wall was correct and the Factor of Safety (FS) for the global stability is

high, which rules out the global stability concern (as expressed by the designer and the

owner) for the ROB-II section with distress. GPR survey has also not indicated any formation

of slip circle type of failure passing through the embankment section. The design calculations

for the retained earth wall consisted of internal stability checks (reinforcement rupture and

reinforcement pullout) external stability checks (for overturning and sliding) and they are

found to be in order. It is unlikely that the longitudinal cracks that appeared might have been

caused due to these internal and external stability issues. Some settlement might have been

caused due to poor foundation (as observed in BH-10 with values of N=10 corresponding to a

depth of 8.5m from the road level). However, this may not be very significant as we did not

observe any level difference across the crack.

Selection of borrow materials and Compaction control was found to be the cause of

distress and cracks. Most of the cracks are confined up to the bottom of the granular sub base

layer. Granular sub base layer seems to be most affected layer. From the detailed

investigations (forensic), the cause of longitudinal cracks are attributed to:

a) From the median area or the voids that might have been present in the top layers of

pavement (BC, DBM and BM), rain water got seeped into the granular sub base layer and

further onto subgrade which might have caused differential settlement of soil layers at the

junction of reinforcement section and un-reinforced section of the embankment and resulted

in the form of longitudinal cracks on the pavement. This is evident from the differential

settlement in various soil layers as observed in GPR survey.

b) These longitudinal cracks are mainly confined in the top 1.5m from the road

surface. However, there are indications that these cracks might have extended up to 4 to 5m

into the compacted soil (as indicated in one of the boreholes) during the heavy rainy days and

subsequent seepage at some locations, where it had small boulders causing a high

permeability zone. At some locations, densities of the backfill material obtained in the

embankment section did not meet the desired values.

c) Selection of granular sub base material and improper drainage through the granular

sub base layer.

Cement grouting is recommended for the compacted backfill and has been done all

along the longitudinal cracked zones at regular intervals of 3m c/c. c. The median portion was

recommended to be made impermeable by providing LDPE layer along with PCC of about 50

mm thick all along the length of the median to prevent water entering the pavement.

Premixed seal coat comprising of a fine aggregate premixed with bituminous binder was also

recommended for sealing the voids in a bituminous surface laid to the specified levels, grade

and camber for the full width of the carriage way.

Case Study 9

EXPERIMENTAL AND NUMERICAL TECHNIQUES FOR ANALYSIS OF SOIL NAILED STRUCTURES

K. Rajagopal1 and T. Rambabu2

1Professor, Dept. of Civil Engineering, IIT Madras, Chennai 600036, e-mail: gopalkr@iitm.ac.in 2 T. Rambabu, M/s Utraco-Ryobi Ground Engineering Pvt. Ltd., Chennai-87, e-mail:

utracoryobi@yahoo.com

ABSTRACT

This paper discusses the comparison in the response of soil nailed retaining walls and slopes obtained by different methods i.e. experimental, limit-equilibrium and finite element analysis. The paper describes some laboratory tests in which 1.75 m high soil nailed retaining walls were subjected to failure by applying uniform surcharge pressure. Their response was studied by three different methods and comparisons were made. Later, the paper describes the stabilization of deep excavations at a construction site in Chennai using pre-stressed nailing technique. The construction site is surrounded by fully-developed residential areas. Any slope failure at the site will result in loss of foundation support to the adjacent structures which may lead to many legal problems. After carefully considering various methods for stabilizing the excavated faces, it was decided to use pre- stressed grouted nails for stabilizing the slope faces. This paper discusses the results from different parametric studies and the results obtained. The implications of different possible alternatives are discussed in the paper.

1. INTRODUCTION

The use of soil nailing is a new and innovative technique that can be used even in very

difficult site conditions. It involves in driving in steel rods into the soil beyond the rupture plane to bind the active soil mass to the interior mass. This technique falls under the category of earth reinforcement and consists of passive bars, named nails, which are generally used to retain excavations, and to stabilize unstable slopes. This technique has been widely used in different types of projects, e.g. soil slopes (Nicholson and Boley 1985, Bruce and Jewell 1987), retaining structures (Stocker et al. 1979, Shen et al. 1981, Guillox and Schlosser 1982), tunneling and other civil and industrial projects. Babu and Singh (2009) have described some latest techniques for the design of nailed soil walls and their verification by finite element based numerical methods. Gosavi et al. (2009) have described some analysis method for nailed soil cuts and their verification by laboratory model tests. The design and construction aspects of soil nailing for stabilizing of slopes and excavations was presented by Elias and Juran (1991).

In this paper, the response of the nailed soil walls is studied through different methods, viz. limit equilibrium method based on bi-linear wedge analysis, laboratory model tests and finite element analyses. The results are analysed for comparing the accuracy of different methods and the implications for the field performance. Later, one case study of a real field case is described in the paper.

3

44 FGE 2009, Bangalore

2. BI-LINEAR WEDGE ANALYSIS METHOD

This limit equilibrium analysis method assumes the failure surface to consist of two surfaces

intersecting within the soil mass. The reinforcement elements are assumed to develop only tensile forces and the effect of the bending stiffness is ignored in the analysis. The length of the reinforcement element beyond the rupture wedge is defined as the embedment length and the pullout resistance is calculated based on this length of reinforcement only. These tensile forces are divided into the components parallel and perpendicular to the assumed rupture plane. The normal force and tangential component in each reinforcement element crossing the potential surface are added to the resisting forces mobilised in the soil when determining the factor of safety of the entire mass. The factor of safety of the soil mass is computed as the ratio between the available and the required forces.

The above calculations are performed repeatedly for different failure surfaces using a computer programme developed in C++ language. The computer program is interactive and reads the input data through the console and the results are saved in a file. The output from the program consists of the factors of safety at each and every rupture plane considered in the analysis. The program also gives the minimum factor of safety for the given configuration.

3. EXPERIMENTAL PROGRAMME

All the retaining wall tests were performed in a specially constructed retaining wall test facility of 2000 mm long, 740 mm wide and 2000 mm height. This facility consists of a mass concrete base slab of 100 mm thickness and two longitudinal side walls that are anchored into the base slab. The side walls of the facility were constructed with hollow concrete blocks of size 400 mm × 200 mm × 200 mm. These blocks were anchored into the foundation soil through tie rods of 12.5 mm extending through the concrete floor. During the tests, steel channels (ISMC 200) were fixed across the two side walls to create a self-reacting rigid frame consisting of the base slab, side walls and the steel channels as illustrated in Figure 1.

The test set up is constructed 1.0 m below ground level and 1.0 m above the ground level for ease of handling large quantities of soil during the tests and also from the stability consideration of the two side walls. A free space of 500 mm was left in front of these walls for construction of the test wall and for access to provide necessary instrumentation.

The backfill of the model retaining walls was subjected to uniform surcharge pressure by inflating an air bag placed between the backfill surface and the cross-channels connected between the two side walls. The top of the air bag was covered with a non-woven geotextile and a 25 mm thick plywood plank was placed between the geotextile and the steel channels to distribute the load uniformly on the soil. This air bag was specially made of double faced neoprene coated nylon reinforced fabric.

An air compressor capable of supplying air at a maximum pressure of 1000 kPa was run continuously during the test to supply pressurised air. The air is routed through a non-return valve. The regulator was adjusted to control the pressure of the air in the air bag to apply the pressure in small increments.

45 FGE 2009, Bangalore

The lateral displacements of the wall facing were measured through LVDTs that were connected to display unit to directly read the displacements. The vertical and lateral soil pressures were measured using strain gauge type pressure cells of 300 kPa capacity that have an accuracy of 0.1 kPa. The loads developed in the soil nails at the wall facing were measured through pre-calibrated load rings fabricated within the geotechnical engineering laboratory of IIT Madras. These load rings were made of 45 mm wide steel pipes (85 mm internal diameter and 2.5 mm wall thickness) fixed with resistance type strain gauges on diametrically opposite sides. Strain gauges were also fixed at four locations on the soil nails to measure the strains along the length of the nails. These strain gauges were protected from damage during the installation by placing them in side a thin plastic membrane wound around the soil nail.

3.1 Construction of the Model Retaining Walls The soil used in the present investigation was a dry river sand with coefficient of uniformity (Cu) of 1.96, coefficient of curvature (Cc) of 1.05 and effective size of particle (D10) of 0.25 mm. The soil can be classified as poorly graded sand with the letter symbol SP according to the Indian soil classification system. The minimum and maximum dry unit weights of the soil are 14.97 and 18.5 kN/m3. The direct shear tests have given the peak friction angle of the soil as 41°.

0.2 m

inflatable air pressure bag

reaction frame wooden plank and nonwoven geotextiles

facing panel

LVDT

2.0 m

sand load ring G.L

pressure cells

strain gauge

2.0 m

Figure 1. Set up for model retaining wall tests

The model walls were then built in the retaining wall test facility with the same configuration to evaluate the accuracy of the limit equilibrium method of analysis of this type of earth structures. All

46 FGE 2009, Bangalore

the tests were performed with a model wall height of 1800 mm. This has left a clear gap of 200 mm to accommodate the air bag between the surface of the backfill and the cross channels to apply the uniform surcharge pressure.

The front facing was made of individual facing blocks (8 rows and 3 columns). The soil nails used in these test walls consisted of 12.5 mm diameter plain mild steel rods. The length of the soil nails was varied from 600 mm to 1600 mm. The number of soil nails also was varied from 9 to 24.

All the tests were conducted at a relative density of the backfill soil of 70%, which corresponds to a peak friction angle of 41°. The backfill sand was placed in the tank in layers of 225 mm thick. For, each lift, the amount of soil needed to produce the desired dry density (at relative density of 70%) was weighed out and placed loosely in the tank. Each layer was then compacted manually using a drop weight of 65 N falling on a wooden plank of plan size 250 mm × 600 mm from a height of 350 mm. This procedure of soil compaction was arrived at by trials to achieve a uniform soil density throughout the test tank. The compacted density of the sand was monitored by collecting samples in steel cups of 90 cm3 volume placed at different depths in the sand bed. The difference in the density of the soil at different locations was found to be less than 1%.

The front facing units were temporarily supported laterally using cement blocks. After the backfill soil was placed to the full height of the model wall, the bricks in each layer were successively removed after the nails in that layer were installed. The nails were installed by driving. A template was used to guide the inclined nails at 10° below the horizontal. After the nails were driven in, the facing panels were connected to the nails through pre-calibrated load rings as shown in Figure 1. This method of construction of the test walls corresponds to the top-down construction technique adopted for slope stabilisation works in the field.

3.2 Test Procedure

The constructed walls were subjected to failure by applying uniform surcharge pressure on the backfill. The surcharge pressure was applied in small increments until the failure of the wall by controlling the flow of air through the non-return valve. The applied surcharge pressure was monitored through the vertical pressures read from the pressure cell fixed at the back end of the wall. Each increment of surcharge was kept constant until the deformations under that particular load increment have reached a steady state.

As the surcharge pressure was increased, the lateral deformations and the load in the load rings were observed to increase. As the failure pressure was approached, the load in the load ring connected to the upper row of the soil nails became constant and had suddenly decreased at some surcharge indicating the pullout failure of the nails at that elevation. The lateral earth pressures measured behind the facing panels also had shown a similar trend. At that stage, the walls were deemed to have failed due to the pullout of the soil nails. The top row of nails became so loose that they could be pulled out by hand. The test was stopped at that stage and the entire test set up was dismantled. Some of the initial tests were repeated again with the same nail configuration to ascertain the repeatability of the test results. Very small differences were obtained in the results obtained from such tests. The tests were repeated with different nail configurations to examine the influence of the length, inclination and spacing of nails.

47 FGE 2009, Bangalore

4. FINITE ELEMENT ANALYSIS OF TEST WALLS

All the finite element analyses in this research work were carried out using the general purpose finite element program GEOFEM developed at IIT Madras, Rajagopal (1995). The behaviour of retaining walls supported by soil nails was analysed by two dimensional finite elemental analysis using plane strain idealisation by smearing the effect of soil nails over 1 m width perpendicular to the plane of the analysis. The influence of some of the important parameters which influence the behaviour of these walls was studied through parametric studies. In order to compare the finite element results with those from experimental work, the exact dimensions used in the laboratory experiments were employed in these analyses.

The finite element mesh used for the analysis of nailed walls on sand consisted of eight-node quadrilateral elements to represent the soil and the front facing units, 3-node bar elements to represent the soil nails and six-node joint elements of zero thickness to represent the interfaces between the front facing and soil and between the nails and soil. The elastic-plastic behaviour and the failure of the soil were simulated by using Mohr-Coulomb yield criterion.

The data from the pullout tests was used to estimate the interface friction angle between the soil nails and the soil. The interface elements consist of two stiffness values, one in the shear direction and the other in the normal direction. The shear stiffness was given a large value to maintain the compatibility between the nail and soil before the failure and was set to a small after the failure to allow for relative deformation. The stiffness in the normal direction was set to a large value to maintain the contact in the normal direction at all stages of the analysis. The interface element is not allowed to develop tensile stresses in the normal direction.

The construction of the wall was simulated by placing the soil in the mesh and applying the self weight as gravity forces with a Ko of 0.50. During this stage, the facing panels were laterally supported by external props. The resulting deformations and strains were set to zero at the end of this stage. During the next stage of analysis, the front props were removed and uniform pressure was applied on the backfill surface in small increments to simulate the pressure loading in the laboratory tests. The analysis was iterated to reduce the out-of-balance forces at each load step to less than 0.5%. The failure of the wall was defined at the stage at which the out-of-balance force increases to a high value and divergence takes place in the solution. The finite element analyses were performed with nail lengths of 0.6, 0.8 and 1.2 m, and with nail inclinations of 0°, 10° and 20°.

5. COMPARISON OF DIFFERENT RESULTS

Typical lateral deformations along the height of the wall at different surcharge pressures are shown in Figure 2 for 1.2 m long horizontal nail. The forces measured in the load rings were found to increase with increase in surcharge pressures, Figure 3. At the ultimate surcharge pressures the top row of facing panels deformed excessively accompanied by the sudden drop of force in the load ring. At that stage, the top row of nails became so loose that they could be pulled out by hand easily. This behaviour shows that the top row of soil nails had failed in pullout. The lateral pressures measured behind the facing units also showed a linear increase with surcharge pressure until failure and showed a sudden decrease at failure. In general, the walls with inclined nails failed at higher surcharge pressures because of the higher pullout capacity of these nails. In all the walls, the lateral wall displacements have decreased towards the

48 FGE 2009, Bangalore

elev

atio

n (m

m)

forc

e in

load

ring

s (N

)

toe of the walls. This is because the length of embedded portion (where the nail resistance is developed) beyond the rupture plane increases towards the toe of the wall. As the deformations are lower towards the toe of the wall, the relative deformations between the soil and nail are also lower and hence the mobilised force in the nail is lower in the bottom row of nails as shown by the results in Figure 3.

2000

1500 applied pressure (kPa)

1000

500

0

0

10

20

30

40

50

60 5 10 15 20

lateral displacement (mm)

Figure 2 Lateral displacements at different surcharge pressures (1.2 m long horizontal nail)

2000

elevation (mm)

1500

1000

1687.5 1237.5 1012.5 562.5 112.5

500

0

0 20 40 60 80 surcharge pressure (kPa)

Figure 3. Load ring forces in soil nailed wall (1.2 m horizontal nail)

The measured lateral earth pressures behind the facings panels are shown in Figure 4. It may be observed that the lateral earth pressures have also increased with the applied surcharge pressure until

49 FGE 2009, Bangalore

points of maximum strain locations over the length of the nail (L=1.2 m)

elev

atio

n (m

)

the collapse state of the model walls. When the failure pressure was approached, the facing panels had excessive lateral outward deformations leading to a decrease in the lateral earth pressures in the top most pressure cell. The reduction of earth pressures was taken as one indication of the failure of the walls.

The comparison of the internal rupture plane measured in the laboratory tests and those predicted by the finite element analysis and the bilinear wedge mechanism is shown in Figure 4. The points with the peak strain in the nails were joined together to obtain the possible rupture plane in both FEA and experiments while the surface with the least factor of safety is taken as the rupture plane in the case of limit equilibrium analysis. It could be seen that the rupture plane predicted by the bilinear wedge method is widely different from the measured surface. This may be another reason for the low ultimate pressures estimated by the limit equilibrium method.

The responses predicted by the finite element analyses with 0.6 m long nails placed at three different inclinations are shown in Figure 5. These responses closely matched those measured in the laboratory tests. It could be observed that the deformations decrease (at the same surcharge pressure) with increase in the nail inclination as also noted from the experimental results. This result clearly shows that the finite element model is able to replicate the improvement in the performance of nailed soil retaining walls with the inclination of the nails.

The comparison between the measured ultimate surcharge pressures on the soil nailed model walls and the corresponding pressures predicted by the limit equilibrium and the finite element methods is presented in Table 1. It can be observed that the pressures estimated from the limit equilibrium method are lower than the measured values in all the cases while the values predicted by the finite element analyses are reasonably close to the measured values. One of the reasons for the discrepancy between the predicted and the estimated results could be the effect of sidewall friction.

2.0

1.5

1.0

0.5

0.0

0.0

0.50

fem exp design method 1.0

Figure 4. Internal rupture surfaces from different methods

50 FGE 2009, Bangalore

pres

sure

(kPa

)

Table 1. Comparison of ultimate surcharge pressures on retaining wall

Configuration of soil nails

Limit Equilibrium analysis (kPa)

Laboratory model tests (kPa)

FE analysis (kPa)

1.6 m (horizontal) 100 130 125

1.2 m (horizontal)

1.2 m (10° inclination)

50

65

60

80

55

85

0.8 m (horizontal)

0.8 m (10° inclination)

35

45

45

60

40

55

0.6 m (horizontal)

0.6 m (10° inclination)

25

35 35

45 30

45

80

60

40

20

0

0 10

20o inclination

10o inclination

horizontal

20 30

40 displacement (mm)

Figure 5. FE predicted retaining wall behaviour with different nail inclinations

6. CASE STUDY OF A PRE-STRESSED NAILED EXCAVATION

At one construction site in North Chennai, it is required to make deep excavations for construction of two basement floors of a shopping complex. The proposed construction site is surrounded by one to two storey residential and commercial buildings. Due to heavy rain fall in the area, the surface rain water collects into the construction area. Hence, dewatering system had to be installed to keep the construction area dry.

51 FGE 2009, Bangalore

In general, the soil profile at the site consists of four distinct layers. The top one meter or so from the ground level is loose silty reddish brown fine to medium sand. This upper soil layer is followed by a three meter thick medium dense to dense brown silty sand of fine to medium grade with average SPT values of 20. This second layer is underlain by a firm to medium dense brown silty sand of 3 m thickness with average SPT values of around 10. The soil layer below consists of silty sand of approximate thickness 6 m thick extending up to 13 m. This layer is a medium dense to dense and is brownish in colour with an average SPT of 30. This layer is followed by dense to very dense, hard silty sand with average SPT values in excess of 50.

The ground water table at the site is around 3 to 4 m from the ground level as per the local information. However due to effective dewatering at the site, the water table is maintained as a depth lower than 8 to 9m below the ground level i.e. at one to two meters below the excavated ground surface. Assumed profiles and proposed slope is indicated in Figure 6.

Figure 6. Schematic sketch of the proposed excavation and ground support

The factor of safety of the slope without any stabilization comes to 0.63 which is unacceptable. The factor of safety of the slope was increased to 1.20 by providing pre-stressed soil nails as illustrated above. This factor of safety was adequate as the excavated slopes are only of temporary nature. The details of the soil nails provided at the site are shown in Table 2. The factors of safety obtained from finite element analyses are slightly larger than those obtained from these analyses.

Table 2. Details of the pre-stressed soil nails at different depths

Row name Location of soil nails

Force in kN/m

Free length(m)

Required fixed length(m)

Total required length (m)

Provided length (m)

A B C D E

-1.0m -2.5m -4.0m -5.5m -7.0m

10 20 25 30 40

3.0 2.0 1.5 1.0 0.0

1.5 3.0 3.75 4.5 4.2

4.5 5.0 5.25 5.5 4.2

6.0 6.0 6.0 6.0 6.0

52 FGE 2009, Bangalore

The slope stability analysis was performed using a limit equilibrium based slope stability software “DC slope” from DC Foundations, Germany. A uniform surcharge of 20 kPa was assumed at the ground level to take account of the loads from moving machinery etc. The results from these analyses were independently verified from finite element analysis using the program PLAXIS-2d. Typical result from one analysis of limit equilibrium analysis is shown in Figure 7.

Figure 7. Typical result from program DC slope

The surface was treated by spraying 50 mm thick concrete to prevent surface erosion. Some typical pictures from the site are shown below to illustrate the nail pre-tensioning and the beam arrangement.

Figure 8. General view of the construction site

53 FGE 2009, Bangalore

Figure 9. Pre-tensioning of the soil nail after the grouting 7. CONCLUSIONS

Any design of soil structure is complicated due to several unknown parameters. Hence empirical methods are employed to come out with safe designs. Invariably, these are based on earlier experience or on some limit equilibrium analysis. The current study has shown that the stability of the nailed wall or excavation is very much influenced by the length and inclination of the soil nails and pre-tension effects.. Some of the major conclusions that can be drawn from this study are as follows:

• The design based on the bilinear wedge failure mechanism is always on the conservative side. • The length and inclination of the soil nail plays significant role in the overall behaviour of nailed

soil wall. • The nail should project sufficiently beyond the rupture plane to develop significant pullout

capacity and also to intercept the potential rupture planes. • The internal rupture plane in the soil nailed walls can be idealised as a bilinear surface and can

be obtained by drawing a vertical line at 0.3H behind the facing and an inclined line through the toe at (45+φ/2)°.

• The influence of the nail inclination is very well represented by the finite element models. ACKNOWLEDGEMENTS

The financial support for this work from the Ministry of Human Resources and Development, Government of India, New Delhi is gratefully acknowledged. Some part of the work reported in this paper was performed by Mr. G.V. Ramesh as part of his thesis work for Master of Science degree by research at IIT Madras.

RFERENCES

Babu, G.L.S. and Singh, Vikas Pratap (2009) Appraisal of Soil Nailing Design, Indian Geotechnical Journal, Vol. 39, No. 1, 81-95.

Bruce, D.A. and R.A. Jewell (1987) Soil nailing: application and practice – part 2. Ground Engineering, The journal of British Geotechnical society, Vol. 20, No.1, 21- 28.

54 FGE 2009, Bangalore

Elias, V. and Juran, I. (1991) Soil Nailing for Stabilisation of Highway Slopes and Excavations,

FHWA-RD-89-198, Federal Highway Administration, McLean, Virginia, USA. Gosavi, M., Swami Saran and Mittal, S. (2009) Model Tests for validation of analysis of nailed cuts,

Indian Geotechnical Journal, Vol. 39, No. 1, 96-115. Guilloox, A. and G.Notte (1982) Experiences on a Retaining structures by Nailing. Proceedings of

8th European Conference on soil Mechanics and Foundation Engineering, Helsinki. Nicholson, P.J. and D.L. Boley (1985) Soil nailing supports excavation. Journal of Geotechnincal

Engineering, ASCE, 111, No. 4. 45-47. Rajagopal, K. (1998) Users Manual for the Finite Element Program GEOFEM, Geotechnical

Engineering Division, Department of Civil Engineering, IIT Madras. Shen, C.K., L.R. Herrman, K.M. Romstad, S.Bang, Y.S. Kim and J.S. De Natale (1981) An In Situ

Earth reinforcement Lateral support system. Department of Civil Engineering, Univ. of California, Davis, Report No. 81-03, U.S. National Science Foundation Report No. NSF/CEE- 81059, March, 130 pp.

Stocker, M.F. and G.Riedinger (1990) The Bearing Behaviour of Nailed Retaining Structures, Design and performance of Earth retaining Structures. ASCE Geotechnical Special publications, No.25, New York, 612-628.

Case Study 10 THE FAILURE ANALYSIS AND REHABILITATION OF

DAMAGED RE WALL

Murthy, B. R. S. Retd. Professor, IISc, Bangalore, India, 560 012, murthybrs@gmail.com

1. BACK GROUND

A Geo-grid Reinforced earth retaining wall for a fly-over along the National High

way No 4 was constructed a few years back. The wall height is varying from 2 to 13 meters. At a location of about 7 meter height of the wall there was snapping of connection between facing element and the geo-grid reinforcement. The authorities requested the author to visit the site and suggest a suitable permanent solution for rehabilitation of the failed zone. Also it was requested to find out the reason for failure. The site was inspected along with all concerned. There was not much information available with the project authorities regarding the quality control or the designs not even a sketch of the actual implementation.

2. SITE CONDITION

The failure had taken place over a length of about 10 meters. The failure was

essentially in the form of collapse of the panels due to snapping of the connecting plastic cleats/clamps, perhaps initially of the upper layers. The number of geo-grid reinforcement layers provided was found to be reasonable. There was variation in the grade of geo-grid increasing with depth. The soil appeared to be in good condition as required from an RE wall point of view. In fact the failure plane had not gone beyond a meter from the wall face which was the boundary of the drainage zone of granular material. The soil behind the failed zone was standing near vertical indicating the sufficiency of the reinforcement layers and good quality of compaction. It was reported that the PHI value for the soil was found to be more than 34 degrees.

In such a situation it was difficult to evaluate the reasons for failure. A few

theories were proposed by engineers and local people for the cause of failure. One particular theory was that a heavily loaded truck rammed against the crash barrier. Really no such sign was visible. All the theories were discounted and it was indicative that most of the reasons were only to divert the attention and hence search started for a realistic one.

3. INVESTIGATION

A closer examination of the detailing and the site condition of failure indicated

that the failure has initiated from snapping of upper layers of geo-grid/ connections. This was indicative from large elongations of the upper layers. Further the crash barrier and

11

12 FGE 2009, Bangalore

the friction slab are rigid and continuous where as the pavement beyond the friction slab is flexible. The integrity between them cannot be established by a thin bitumen pavement on the top. The wheel loads will transmit the load and vibrations differently in the two systems. In rigid pavement portion the vibrations are transmitted down into the granular medium whereas the same is damped/ dissipated in the flexible pavement. The transmitted vibration will tend to compact the granular soil resulting in the loss of contact of the friction slab and soil. This will in turn makes the friction slab to transmit the reaction due to the wheel load to the facing element. As the facing element assembly is not designed to transmit the axial loads the reaction will tend the assembly of facing element to BUCKLE causing the snapping of the cleats/ joints between the facing and the geo-grid. This is the primary reason for the failure. Figures A and B indicate the schematic representation of the wall and details of the analysis of failure.

4. REHABILITATION SCHEME

To reconstruct the wall, the following scheme was proposed and implemented and the

same has been presented in the drawings. The essential steps were as follows.

1. Support the failed surface with sand bags and provide support to the exposed crash barrier using wall form jacks.

2. Providing a suitable staging, remove the sand bags to a depth of 1.65 meters. Trim the failure plane to remove loose materials.

3. Drive 6 meter long 20 TOR bars at a spacing of 550 mm c/c (three rows) in both directions.

4. Fasten 50x50x2.8 mm MS grid reinforcement vertically along a defined line which can facilitate fixing of the permanent RE wall panels and horizontally along the lowest layer of nails.

5. Fill the void space beyond the MS grid with 50 mm aggregates as densely as possible.

6. Provide 50mm thick shot-crete on the weld mesh as per relevant IS code. 7. Repeat the steps 2, 3, 4, 5 and 6 to reach the full depth. 8. Install the RCC pre-cast facing panel for one row and fasten the same suitably

with the projected nails. 9. Back fill the gap between the shot-creted surface and the Pre-cast panel with self

compacting coarse sand. 10. Repeat the steps 8 and 9 to reach the top level.

13 FGE 2009, Bangalore

CRASH BARRIER WITH FLEXIBLE PAVEMENT

FRICTION SLAB (RIGID) BITUMIN WEARING SURFACE

RCC FACING GEOGRID LAYERS

GRANNULAR DRAINAGE MEDIUM about 1 meter thick

FIGURE A

14 FGE 2009, Bangalore

CRASH BARRIER WITH FRICTION SLAB (RIGID)

FLEXIBLE PAVEMENT

BITUMIN WEARING SURFACE

REACTIONS

RCC FACING GEOGRID LAYERS

GRANNULAR DRAINAGE MEDIUM about 1 meter thick

FIGURE B

15 FGE 2009, Bangalore

ROAD SURFACE

SAND BAGS TO A SLOPE OF 1:1

FAILED SOIL SURFACE WALL FORM JACKS TO SUPPORT THE CRASH BARRIER

Figure1. The failed reach is protected using sand bags to a slope of 1:1 and supporting the crash barrier with wall form pipe jacks

16 FGE 2009, Bangalore

20 TOR BARS OF LENGTH 0.7H DRIVEN AT SPACING OF 550 C/C IN BOTH THE DIRECTIONS

Figure 2. The sand bags are removed to a depth of 1.65 meters and three rows of nails are driven in to the back fill material.

17 FGE 2009, Bangalore

50x50x2.8 mm MS WELD MESH FASTENED TO THE NAILS AND VOID SPACE BEHIND FILLED CAREFULLY WITH AGGREGATES

50 mm THICK SHOTCRETE AS PER IS 9812

Figure 3. MS weld mesh of size 50x50x2.8 mm is fastened to the nails and at the top of the sand bags. The gap between the soil and the weld mesh is carefully packed with aggregates of suitable size. 50mm thick shot-crete is provided on the weld mesh. Repeat the procedure of removal of sand bags, driving of nails, fastening of weld mesh, backfilling with aggregate and shot-creting to reach the full depth of the wall as indicated in Figures 4 and 5.

18 FGE 2009, Bangalore

Figure 4

19 FGE 2009, Bangalore

Figure 5

20 FGE 2009, Bangalore

PRECAST RCC FACING PANELS ASSEMBLED FROM BOTTOM

THE FACING PANEL IS CONNECTED TO THE NAIL SYSTEM AND THE GAP IS BACK FILLED WITH SAND

Figure 6 The pre-cast RCC facing panels as per the original designs shall be assembled along the original alignment starting from the ground level. The panels are connected to the nails using suitable MS cleats and fastening system. The gap between the shot-crete surface and the RCC pre-cast panel shall be backfilled with good self compacting material like sand. The above sequence shall be followed till the full height is covered.

INTERACTION BETWEEN APPROACH EMBANKMENT ON SOFT GROUND AND ROAD OVER BRIDGE FOUNDATIONS – A CASE STUDY

M.R.Madhav1, B.Vidyaranya2 & G.B. Rajendra Prasad3

1Professor Emeritus, 2Research Scholar & 3Asst. Executive Engineer

1JNTU College of Engineering, 2Osmania University, Hyderabad & 3R&B Amalapuram Division madhavmr@gmail.com & vidyaranya_b@yahoo.co.in

ABSTRACT

The paper presents an interesting instance of approach embankment construction affecting the already constructed Road Over Bridge (ROB) and its foundations. The ground consists of very thick layer of very soft to soft soil extending to a depth of nearly 15 to 16 m underlain by dense sands. As such, the bridge was constructed on well foundations sunk into strong and stiff underlying strata. In spite of this fact, the bridge suffered extensive damage due to the construction of nearly 4.5 m high embankment. The lateral squeezing of the soil is believed to be the cause of the failure/damage to bridge foundations. The case history highlights the need for an integrated approach for the design and construction of a bridge and its approaches.

1. INTRODUCTION

The design of structures on or over soft soils needs special care and concern. The interaction between structures adjacent to each other is an important aspect that should not be ignored. The paper presents a case study of a Road over Bridge which was constructed prior to the construction of the approach embankments. The embankment was constructed subsequently. However, the effect of possible squeezing of soft soil in the longitudinal direction was not considered leading to a major geotechnical problem. The paper presents the details of the bridge, damage caused to it and the results of soil investigations at the site.

2. DESCRIPTION OF THE SITE

The site under investigation is on a bypass road constructed to facilitate free movement of traffic over a railway line and avoid the use of a level crossing. The proposed road connecting two towns passes over an irrigation canal and a railway track approximately 400 meters apart. The irrigation canal runs parallel to the railway track. An embankment of height 6.5 m was to be constructed to connect the ROB and the bridge crossing the canal. The site is surrounded by paddy fields all-round. The ground water table was at a depth of about 1.5 m below the existing ground level at the time of investigation but is known to rise close to GL following monsoon rains.

A road runs parallel to the canal on the side away from the ROB. The bridge structure over the canal was founded on a raft 15 m wide. The bridge settled by about 0.5 to 1.0 m and the piers tilted leading to severe cracking of the deck, piers and the abutments, the bottom of the canal heaved due to the construction of the approach embankment to the bridge on the canal. The sides of canal supporting the road were

lined with concrete panels to support the surrounding soil. Even the panels moved in to the canal by about half a meter. The plan of ROB is shown in Fig. 1. The ROB consists of a single span of 16.0 m long RCC ‘T’ beam slab. The abutments were founded on well foundations two under each. The wells, 4.50 m in diameter, were plugged at depth between 18.30 m to 18.60 m below the ground level. Wing and return walls (Fig. 2) were made in Mass CC M-20 grade up to ground level and Mass CC M-30 grade above ground level with 40 mm grade hard granite stone aggregate.

Figure 1 Plan of the Road over Bridge

Fig 2 Section of Return Wall

6

7 FGE 2009, Bangalore

Foundation provided up to 4.0m below ground level consisted of 2.0 m sand filling well compacted in layers of 300 mm and RCC raft foundation of 0.50 m followed by 1.50 m sub-structure with 1:1 slope on one side and straight face on other side. Figs 1 through 4 show the plan, sectional details of the abutments and the return walls in detail.

Figure 3 Section of Abutment

Figure 4 Section of Wing Wall

The embankment was constructed using locally available soil. A layer of geogrid was laid at the bottom of the fill. The soil used for the embankment appeared to be heavy clay, possibly with slightly higher percentage of fine sand and silt. Sandy soils were used for the central core while clay was used for the remaining portion of the backfill.

The wing walls got settled and heavy cracks were generated. The construction of the bridge was completed in Jan’ 01 but the construction of the embankment was delayed till year 2005. The work was recommenced with work on Abutment No.1 side to a height 6 m from the top of the well cap and to 1.2 m from the bottom of the bed block. A crack was observed at a level of 1.20 m below the top of bed

block. The crack had widened to a width of about 25 mm since then. The top of the abutment moved by about 15 to 20 mm towards the backfill. Earthwork on the rear side of the abutment No. 2 was carried out to a height of 4.4 m from the well cap level. Horizontal cracks 25 mm wide were noticed at 3.60 m 4.80 m and 6.0 m from the top of the slab on wing walls No.4. Figs 5 through 10 shows the tilt generated in the wing walls due to movements caused by embankment construction and Table 1 summarize the movements reported.

Figure 5 Sectional Elevation of Abutment with wing walls

Figure 6 Sectional Elevation of Wing Wall 1

Figure 7 Sectional Elevation of Wing Wall 2

8 FGE 2009, Bangalore

Figure 8 Sectional elevation of Abutment with Wing Walls

Figure 9 Sectional Elevation of Wing Wall 3

Figure 10 Sectional Elevation of Wing Wall 4

Table 1 Details of Movements of Abutments

Horizontal Movements

Loca tion

Settlement at top

Lean Toward Fill Well

Cap Bed

Block W1 270 340 120 20

W2

400

290

100

Butting against

abutment W3 Nil 75 75 40 W4 No

formation 120 20 Nil

3. SITE INVESTIGATIONS

The site was investigated by several agencies prior to the construction of the bridge. The soft clay layer was reported to extend up to a depth of 18.5 m. The water table was close to GL. The site conditions needed fresh investigations to evaluate the soil properties and to examine the reasons for the movements. Fresh boreholes were advanced to a depth a 25- 30 m at three locations close to the ROB structure.

The soil profile indicated black soft clay up to 15 to 18 m depth. Medium to fine sand underlies the soft clay. SPT ‘N’ values were negligibly low and some times equal to zero. The variations of the SPT values with depth are shown in Fig. 12. SPT N values were slightly high (4 to 5) near the top (2-3 m) possibly due to desiccation but reduce to 2 to 3 until a depth of 15 in BH.1, 16 m in BH.2 and 18.0 m in BH.3. N values increased to about 20 to 27 at 18 to 20 m depth range. N values in the lower sand stratum were greater than 40.

The liquid limit values ranged from 72 to 85 while the plasticity index values were of order 45 to 55. The natural moisture content increases with depth over the top 4.0 to 5.0 m and lie in the range of 75 to 90. The liquidity index values are greater than 1.0 at few locations. The shrinkage limit values measured over selected depths are 8.0 for depths less than 11.0 m and about 10 at a depth of 16.0 m. Low shrinkage limit and high plasticity index attest to the fact that the soil is of expansive type and the predominant clay mineral could be montmorillonite. The soils at the site are within the flood plains of the river Godavari. The river over a major part of its length traverses through the Deccan Plateau wherein basaltic formations exist. These rocks normally weather in to expansive clay, locally called Black Cotton clay.

4. IN SITU TESTING

Standard Penetration Tests were conducted in the three boreholes. Results from other agencies indicate N values of 2 to 3 up to a depth of 12.5 m and increase to about 10 to 12 at 17 to 18 m. The N values

9 FGE 2009, Bangalore

Dep

th, z

D ep

th, Z

Dep

th

Dep

th, Z

in the sand stratum were in excess of 30. Another investigation conducted in 2001 also indicated N values of 45 to 60 in the depth range of 19 to 22.

The investigations carried out by JNTU corroborate the above studies. SPT N values were slightly high (4 to 5) near the top (2-3 m) possibly due to desiccation

but reduce to 2 to 3 until a depth of 15 in BH.1, 16 m in BH.2 and 18.0 m in BH.3. N values increase to about 20 to 27 at 18 to 20 m depth range. N values in the lower sand stratum were greater than 40. Fig. 12 shows the variation of the SPT values conducted by various consulting agencies.

JNTU Professional Agency 1 Professional Agency 2

N

0 10 20 30 40 50 60 0

5

10

15

20

N

0 10 20 30 40 50 60 0

5

10

15

20

25

0 20

0

5

10

15

20

25

N

40 60

25

30

30 BH - 1 BH - 2 BH - 3

30

Series1 Series2 .

Fig. 12 Comparison of SPT N Values from Different Agencies

5. RESULTS AND DISCUSSION

The laboratory tests were conducted on the undisturbed soil samples to determine the Atterberg limits, the natural moisture content and the undrained strength of the soil at various depths. Fig 13 shows the variation of the moisture content with depth for the three boreholes investigated. The water content is close to 60% in the top 3 to 4 m depth and increase to 80% in the soft clay layer. The values reduce to about 20 to 25 at depth 20 m within the sandy soils underlying the soft clay. Fig. 14 shows the variation of the Liquidity index with depth for the three boreholes investigated.

Water Content 0 20 40 60 80 100

0

5 10 15 20 25 30 35 BH-1 BH-2 BH-3

Figure 13 Variation of Natural Moisture Content with Depth.

10 FGE 2009, Bangalore

z

-0.5 0.5 LI

0

10

20

1.5 2.5

30 BH-1 BH-2 BH-3

Figure 14 Variation of Liquidity Index with Depth

The displaced soil induced lateral pressures on the surrounding structure as a result of which the wing walls settled and cracked. The damage to the ROB structure is severe due to large lateral pressures generated by the embankment fill.

6. ANALYSIS

The undrained strength of the in situ soft soil is around 10 to 15 kPa. The ultimate bearing capacity of the embankment would be of the order of 55 to 65 kPa. An embankment of 4.5 m height generated a vertical stress of 83 kPa (=4.5x18.5 kN/m3). The construction of the embankment induced plastic flow and failure of the soft soil. Since the soil is highly plastic clay and has low bearing capacity, the soil was squeezed laterally due to embankment loading. The lateral movements were less across the embankment possibly because of the berms and basal reinforcement provided. However, there was no such restraint in the longitudinal direction towards the wing walls and the abutments. Hence the substructure moved significantly.

7. CONCLUSIONS

The case study illustrates an interesting interaction between an already constructed bridge abutments and embankment constructed subsequently by another agency with no communication between the two. The stability of the embankment was ensured in the transverse direction by the provision of berms and basal geogrid reinforcement. The interaction between the embankment and the abutment in the longitudinal direction was not considered resulting in large unacceptable movements of the bridge structure rendering it dysfunctional.

Case Study 12

AN INVESTIGATION INTO THE COLLAPSE OF A PORTION OF SOIL NAILED WALL

Prof. T.G. Sitharam,

Professor of Geotechnical Engineering, and

Chairman, Centre for infrastructure, Sustainable Transportation and Urban Planning

Indian Institute of Science, Bangalore – 560012.

Email: sitharam@civil.iisc.ernet.in

SUMMARY

The excavation for a site, which has three basements, the soil nailing has been

adopted for retaining the earth so as to have a vertical cut temporarily until a permanent

concrete retaining wall is constructed. On March 12th, west side of the soil nailed wall was

collapsed (see Fig 2) for a length of 20m and road on the adjacent property was also

damaged. A temporary restoration scheme (See Fig 4) consisting of earthen embankment fill

and sand bags fill was already adopted as an immediate measure. Geotechnical report clearly

indicates that the top soil (is about 0.5m) and lateritic red soil (up to a depth of 2.75m to

3.5m) from the ground level. Very close to the failure area, the lateritic red soil exists up to

3m depth below which disintegrated rock is present. Soil gradation also clearly indicates

presence of uniform sand (less than 2mm particle size), silt and clay sizes non presence of

any gravels or boulders in the area. N values are invariably greater than 15 in insitu condition.

Driven soil nailed walls have been constructed as a temporary measure all along the

boundary to retain the earth for a depth of about 10m. 20mm dia TOR steel nails of length 5m

long at spacing of 450 mm c/c were driven in to the ground as shown in cross section. 50mm

thick shotcrete of 1:2:2 grout with 50x50x2.6mm wire mesh was adopted (see Fig 2 for the

final finished wall surface). Nailing was done in stages of 1.5m deep at every stage. This was

reported to be repeated up to a depth of hard rock. Typical cross section of the soil nailed

wall is shown below in fig 1.

The soil nailing has been reported to be completed on 14th October 2008. On March

12th, west side of the soil nailed wall was collapsed for a length of 20m and road on the

adjacent property was also damaged. The typical diagram clearly indicates that the failure

(yellow colored area) is only in the top 3m of the soil mass which corroborates well will the

geotechnical borehole data. The fig 2 clearly indicates there is no failure below 3m from the

ground level (Ground RL is at 99.85 m). The failure surface does not pass through the

disintegrated rock mass. Figure 3 shows the collapsed area in plan.

39

Fig 1. Cross Section of the soil nailed wall and Plan view

4 

   

Fig 2. Typical Cross-section of the failed slope

Fig 3. View of the soil nailed wall after collapse (left fig) and restored slope with sand bags (right photo)

40 FGE 2009, Bangalore

Fig 4. Temporary scheme adopted

Considering the adopted temporary restoration scheme with embankment (see

photograph in Figure 3 and Fig 4), the non-availability of working space on the west side

road, and noise/vibration due to piling, it was decided to work within the property of the

client from inside their land initially to restore the slope.

The following scheme (as shown in Figure 5) is suggested as a remedial measure for

permanent restoration of the failed section of the wall (about 20m long). The scheme consists

of construction of a retaining wall at the toe of the temporary embankment, grouting of the

pervious embankment to take the applied pressures from the bottom approach ramp and

grouting of the sand bags/filled up soil in the failed slope section. As a precautionary

measure, even the temporary embankment and the adjacent road re-constructed using sand

bags has been grouted by drilling holes using clean cement grout, so that it can take the load

of the approach ramp uniformly.

41 FGE 2009, Bangalore

The whole area was reported with oozing water on the face of the wall (See photo in

Fig 2). This might be due to the puncturing of the live manhole at several points due to soil

nailing. Further, the failure is attributed to not providing the deisgned length of the nails in the

execution. The actual length of nails provided were shown through analysis shorter than what

has been proposed to have a factor of safety less than 1. The present paper highlight the scheme

of restoration suggested by the consultant rather than the cause of failure. Consultant was

involved only after the failure in the month of April 2009. This presentation describes the scheme

of restoration. The scheme involves a retainng wall away from the boundary (inside the clients

area) along with a backfill and a retaing wall at the toe of temporary embankment coupled with

constructing approach ramp in stages. As this scheme did not cause any disturbance to

the neighbouring areas, the scheme was accepted and executed. The

disadvantage of the scheme is that the client lost a small area of about 6 to 8 car parks.

Fig 4. The final view of the restored retaining wall with the approach slab

42 FGE 2009, Bangalore

CASE STUDY 13

THE CASE OF A TILTED CHIMNEY – A FORENSIC ANALYSIS

RAO.V.V.S. Nagadi Consultants Pvt. Ltd., New Delhi.

THE PROBLEM: A R.C. Chimney of 45m.height having a diameter of 5m. at GL was constructed in a sugar factory in the year 2001. By about June 2002 the chimney was observed to have tilted by 450mm. at the top. To correct the tilt, sand bags were piled up on the diametrically opposite side of the direction of tilt. The chimney regained the original plumb, but it developed wide vertical cracks at three locations on the chimney throughout its height. To develop an appropriate repair solution, detailed investigation was conducted to determine the cause of tilt. DATA COLLECTION: From the design as well as “as constructed” drawings, it was observed that the chimney had been constructed as per designs. Detailed borehole investigations had been conducted and proper design parameters had been used. However, there was no borehole data at the location of the chimney. Hence bore hole investigations were performed around the chimney. The sub-soil profiles were uniform. The profile is shown in fig. 1.

Fig.1.SUB - SOIL PROFILE

-0.0

-1.0

-2.0

-3.0

-4.0

-5.0

-6.0

-7.0

-8.0

-9.0

-10.0

-11.0

-12.0

-13.0

-14.0

-15.0

DEP

TH IN

MET

RES

BH 3

-5

-4

-10

-16

-22

-20

-26

-28

-37

………………………………………………………………………………………………………………………………………………………………………………………………………………………

FILL

E.G.L

E.G.L.: Existing Ground Level

LEGEND : 5 Observed 'N' Values Water Table

GREY SAND(90-99%) WITH TRACES OF SILT (1-10%)

GREY CLAYEY (4-12%) SANDY (9-22%) SILT (65-82%) WITH GRAVEL (1-10%)

FILLED UP STRATA (GREY SANDY SILT/SAND WITH BRICK BATS AND DEBRIS)

SOIL PROFILE

The foundation raft was founded at a depth of -3m. The soil below consists of medium dense sand.

During the site visit, it was found that a single storey shed had been built very near to the chimney on the tilted side. Upon enquiry, it was told that the shed was constructed after the factory went into operation, without consulting the design or geotechnical consultants. The crack was observed a few weeks later. The cross-section of the construction was as shown in fig.2.

In plan, the section at GL was as shown in fig.3.

Fig.2.

CONCLUSIONS: The additional loads acting on diametrically opposite ends of the foundation raft resulted in inducing splitting action of the chimney resulting in cracks.

VIEW OF A CRACK VIEW OF CHIMNEY AND SHED

CASE STUDY 14

INFLUENCE OF VIBRATIONS ON INSTALLATION OF

BORED PILES

Dr. N. Santosh Rao Nagadi Consultants Pvt. Ltd., New Delhi

1. INTRODUCTION

This is a case study of a forensic geotechnical investigation to assess the possible cause/s which prevented the satisfactory construction of large diameter bored piles using the bailer technique. The investigation involved first identifying the possible causes, carrying out field and laboratory tests where necessary to assess the same, analysis of the test results and finally arriving at the conclusion. 2. PROJECT DETAILS

The project in this case study is an elevated corridor through the Kishanganj town, Bihar, India, located parallel to the existing road as also the railway line abutting the road. The elevated corridor had been proposed to be constructed in reinforced cement concrete with the deck consisting of box girders supported by two column piers resting on 1m diameter 16m long bored piles. 3. PROBLEM DEFINED

The piling contractor had been unable to ensure satisfactory construction of the large diameter bored piles using the bailer technique, primarily because of frequent collapse of the sides of the bores made for the purpose of construction of the bored piles. As per the conditions of the contract, bentonite slurry was not to be used in the pile bores and use of liner/casing had also not been envisaged on account of the nature of the subsoil as indicated by the report of the geotechnical investigations carried out earlier for this project. 4. POSSIBLE CAUSES

An evaluation of the above problem indicated that one or both of the following could be the possible causes :

1. Subsoil being unsuitable for construction of the bored piles under the constraints put up by the contract.

2. Vibrations transmitted from the trains moving on the nearby railway tracks and vehicle movement on the adjacent road.

5. INVESTIGATION OF POSSIBLE CAUSES a. Subsoil Conditions

To check whether the subsoil conditions at the site of the proposed elevated corridor are similar to those considered in the design and included in the tender documents, a few boreholes had been carried out at select locations along the length of the elevated corridor. A comparison of the typical subsoil profile indicated by these boreholes with the typical subsoil profile indicated by the geotechnical investigation report included in the tender documents is given below:

Geotechnical Investigation Report New Boreholes Depth (m) Soil

Description Range of N-

values Soil

Description Range of N-

values

0 – 5 10 – 15 Fine Sand with Silt 10 – 14

5 – 10 15 – 20 Medium to Coarse Sand 12 – 20

10 – 20

Silty Fine Sand / Sandy Silt

20 – 25 Coarse Sand 25 – 35

20 – 25 - - Coarse Sand with Gravel > 50

The above comparison clearly shows that the actual subsoil conditions differ

significantly from that indicated by the geotechnical investigation report. While an unsupported pile bore may remain intact in a silty sand / sandy silt soil the same will invariably tend to collapse in a medium to coarse sandy soil. Therefore, while based on the results of the geotechnical investigation report, an unsupported pile bore can be considered feasible, the same will be considered difficult to infeasible when the actual subsoil conditions as indicated by the new boreholes are taken into account. b. Vibrations from Rail and Road Traffic

To assess the level of vibrations caused by the rail and road traffic in close proximity to the site of the proposed elevated corridor, vibration measurements had been carried out using geophones (i.e. velocity pick-ups with a sensitivity of 28.3 mv/mm/s) connected either to a storage oscilloscope or to a computer based data acquisition system. The locations of the vibration measurement points have been indicated in the schematic layout plan of the area given in fig.1.

The recorded vibrations from different types of trains as also from trucks plying on the road have been presented in figs. 2 to 6. The absorption coefficient for the subsoil at the site has been determined from the vibrations recorded during the passage of a Rajdhani express train by locating the geophones in a line perpendicular to the railway tracks on the opposite side of the road under the proposed elevated corridor.

From the recorded vibrations and the absorption coefficient determined, the values of the acceleration ratio at various distances from the tracks will be :

Distance (m) : 6 10 14 18 22 Acceleration Ratio : 0.03 0.013 0.009 0.005 0.002

As per Tscheboterioff (1965), dynamic forces caused by pile driving operations having high accelerations in the range of 0.01g to 0.001g have been shown to have caused significant soil settlements to cause distress to the nearby structures and substructures. As the levels of vibrations caused by the rail traffic are in the above range, the vibrations caused by the rail traffic can be considered be a major contributory factor in the collapse of the pile bores. 6. CONCLUSIONS

The investigation of the possible cause/s indicates that both the considered possible causes have contributed to the frequent collapse of the pile bores in that:

1. The subsoil profile given in the tender documents as indicated by the geotechnical investigation report is wrong.

2. The actual subsoil conditions indicate that the subsoil at the site is unsuitable for use of the bailer technique.

3. The agency that carried out the initial geotechnical investigation is the main culprit in having provided wrong data about the subsoil conditions leading to the inappropriate choice of method of construction of the bored piles.

4. The engineers of the highways authority should have properly verified the geotechnical investigation report before going in for tendering of the work and insisting on bored piles with restrictions on method adopted for the same.

5. Vibrations from the rail and road traffic are significant enough to have contributed to the frequent collapse of the pile bores.

6. Use of full length liner would be the only method by which bored piles could be constructed in such subsoil conditions where significant vibrations are generated by rail and road traffic.

Fig.1 : Schematic Layout Plan

Fig.2 : Express Train on Track 3 in E-W direction

Fig.3 : Rajdhani Express Train on Track 3

Fig.4 : Truck Plying on the Road

Fig.5 : Passenger Train on Track 1

Fig.6 : Goods Train on Track 3

ANNEXURE

COMPOSITION OF TC 40: FORENSIC GEOTECHNICAL ENGINEERING

Chairman:Dr.V.V.S.Rao, India dr_rao@nagadi.co.in Secretary: Prof. G.L. Sivakumar Babu, India, gls@civil.iisc.ernet.in Core Members: 1. Prof. Yoshi Iwasaki, Japan, yoshi-iw@geor.or.jp 2. Dr. Richard Hwang, Moh & Associates, Taiwan, richard.hwang@maa.com.tw 3. Mr. David Starr, Golder Associates, Australia, david_starr@bigpond.com 4. Prof. K.K. Phoon, Singapore, cvepkk@nus.edu.sg 5. Mr. P.W. Day, South Africa, day@jaws.co.za 6. Prof. J. Mesci, Hungary, mecsi@pmmk.pte.hu 7. Dr. Jan Hellings, U.K., jan@janhellings.com 8. Mr. D.S. Saxena, USA. dsaxena@ascworld.net 9. Prof. Mihail e Popescu mihail.e.popescu@gmail.com General Members: 1. Prof. L.G.De Mello, Brazil lgdemello@usp.br 2. Dr. Dianging Le, China dianging@whu.edu.cn 3.Prof. Zeliko Arbanas, Croatia zeliko.arbanas@gh.hr 4.Mr. Jack Pappin, Hong Kong jackpappin@arup.com 5. Prof. Lingwei Kong, China lwkong@dell.whrsm.ac.cn 6. Mr. W.K.Pun, Hongkong wkpun@ceddigov.hk 7. Dr. R.Szepeshazi, Hungary szepesr@sze.hu 8. Prof. S. Hamsin, Kazakstan 9. Dr.W.Cichy, Poland wcic@pg.gda.pl 10. Prof. R. Gilbert, USA bob.gilbert@mail.utexas.edu 11. Dr. Chan S. F. Malaysia sfchan@pc.jaring.mv Corresponding Members: 1. Mr. Marco Uzielli Italy, marco.uzielly@ngi.no 2. Mr. Gil Yoon, South Korea glyoon@kordi.re.kr 3. Mr. Masahiro Shirato, Japan, shirato@pwri.go.jp 4. Mr. Makoto Suzuki, Japan, makoto.suzuki@shimz.co.jp 5. Mr. Limin Zhang, Hong Kong, cezhangi@ust.hk 6. Mr. Chris Basile P.E. (USA) christopherbasile@yahoo.com 7. Mr. Grey Stephan P.E. (USA) gstephan@ascworld.net 8. Dr.Phili Pells pells@psmtoo.com.au 9. Mr. Michael Marley, mmarley@golder.com.au 10. Strath Clarke 11. Prof. Mihail Popescu , mepopescu@usa.com THE SUBCOMMITTEES SC 1: Characterization of distress. Convener: Mr. P.W. Day SC2: Diagnostic Tests. Convener: Mr. David Starr SC 3: Back Analysis. Convener: Dr. Hwang SC 4: Instrumentation. Convener: Prof. Yoshi Iwasaki SC 5: Development of failure hypothesis: Convener: Dr. J.Hellings SC 6: Reliability Checks. Convener: Prof. K.K.Phoon SC 7: Legal Issues: Convener: Mr. D.S.Saxena SC 1: CHARACTERIZATION OF DISTRESS. Convener: Mr. Peter W. Day day@jaws.co.za Members: 1. Prof. Luiz de Mello lgdmello@usp.br 2. Jack Pappin jackpappin@arup.com 3. Prof. Bob Gilbert bob_gilbert@mail.utexas.edu 4. Dr. Jan Hellings jan@janhellings.com 5. Nick Shirlaw nshirlaw@golder.com.sg

SC 2: DIAGNOSTIC TESTS. Convener: Mr.David Starr david_starr@bigpond.com Members: 1. Dr Philip Pells, Pells Sullivan & Meynink (PSM) pells@psmtoo.com.au 2. Mr Michael Marley, Golder Associates mmarley@golder.com.au 3. Strath Clarke (he works for PSM ) SC-3: BACK ANALYSIS. Convener: Dr. Mihail Popescu mihail.e.popescu@gmail.com Co-Convener:Dr.RichardHwang. richard.hwang@maa.com.tw Members: 1.Dr.StephenButtling stephen.buttling@hyderconsulting.com 2.Mr.NiclauseShirlaw nshirlaw@golder.com.sg 3.Dr.TingWenHui tingwh@pop.jaring.my 4.Prof.FusaoOka foka@mbox.kudpc.kyoto-u.ac.jp 5.Dr. Chung Tien Chin ct.chin@maaconsultants.com SC 4: INSTRUMENTATION Convener: Dr. Yoshinara Iwasaki yoshi-iw@geor.or.jp Members: 1.Mr. Craig A. Davis (USA) craig.dvis@ladwp.com 2. A. WADA (Singapore) : wada@agacorp.com.sg SC 5: DEVELOPMENT OF FAILURE HYPOTHESIS. Convener: Dr. Jan Hellings. jan@janhellings.com Members: 1. Bill Grose. Ove Arup & Partners (UK). bill.grose@arup.com 2. Rab Fernie. Cementation Foundations Skanska Ltd rab.fernie@skanska.co.uk 3. Kjell Karlsrud. NGI (Norway). kjell.karlsrud@ngi.no 4. Dirk Luger. GeoDelft (The Netherlands). dirk.luger@geodelft.nl 5. Mahmoud Mahmoud. GES Geotech (Canada). mahmoud@gesgeotech.com 6. Darren Page. OTB (UK). geoscience@hotmail.com 7. Roger Thomson. EDGE Consultants UK Ltd. rpt@edgeconsultants.co.uk SC6: RELIABILITY CHECKS. Convener: Prof. Kok-Kwang Phoon (Singapore,) – cvepkk@nus.edu.sg Members: 1. Prof. Sivakumar Babu (India) - gls@civil.iisc.ernet.in 2. Mr. Robert Gilbert (USA) - bob_gilbert@mail.utexas.edu 3. Mr. Dian-qing Li (China) - dianqing@whu.edu.cn 4. Mr. Marco Uzielli (Italy) - Marco.Uzielli@ngi.no 5. Mr. Gil Yoon (South Korea) - glyoon@kordi.re.kr 6. Mr. Masahiro Shirato (Japan) - shirato@pwri.go.jp 7. Mr. Makoto Suzuki (Japan) - makoto.suzuki@shimz.co.jp 8. Mr. Jack Pappin (Hong Kong) 9. Mr. Limin Zhang (Hong Kong) - cezhangl@ust.hk SC 7- LEGAL ISSUES INVOLVING JURIPRUDENCE SYSTEM Convener: Mr. Dhirendra S Saxena P.E. (USA) dsaxena@ascworld.net Members: 1. Mr. Christopher C Basile P.E. (USA) christopherbasile@yahoo.com 2. Mr. Greg A Stephan P.E. (USA) gstephan@ascworld.net