premium digest december 2010 development of integrity management strategies for pipelines

7/27/2019 Premium Digest December 2010 Development of Integrity Management Strategies for Pipelines

http://slidepdf.com/reader/full/premium-digest-december-2010-development-of-integrity-management-strategies 1/6

13 PiPelines international digest | deCeMBer 2010

teChniCal

IntroductionThe process o assessing the integrity o pipelines involves

carrying out a detailed study in which the sae remaining lie o

a pipeline is determined and an integrity-management plan is

produced. The ocus o such an assessment is on a metal-loss

inspection o the pipeline using intelligent in-line inspection

(ILI) tools. The inspection fndings are combined with a review o

corrosion management activities in order to diagnose the likely

causes o corrosion, or other damage, and identiy appropriate

preventative measures to minimise urther deterioration o

the pipeline. Based on determined corrosion growth rates,

predictions o uture repairs together with mitigation andre-inspection requirements can be determined. From this,

pipeline operators are able to produce an optimised inspection,

maintenance, and repair strategy in order to ensure the integrity

o their assets and extend their sae working lie while improving

pipeline saety and reliability.

This article demonstrates the important role o in-line

inspection data within the ollowing key stages o an integrity

management strategy:

• Corrosion assessment;

• Corrosion growth assessment; and,

• Remnant-lie assessment and corrosion management

strategy.

The ollowing sections describe the practical application o

each o these activities using the case study o a pipeline section

operated by Jadranski Natovod Plc, Joint Stock Co ( JANAF Plc)

in Croatia.

Background to the case studyJANAF Plc, based in Croatia, manages an international crude oil

transportation system, designed and built in the period rom 1974

to 1979, rom the port and terminal o Omišalj on the island o Krk,

Croatia, supplying both local and oreign refneries in Eastern and

Central Europe (Figure 1).

MACAW Engineering and Rosen have worked with JANAF in

developing its pipeline integrity management strategy or severalpipelines within its system. MACAW has produced integrity

assessments o these pipelines based on in-line inspections

perormed by Rosen, and one such example is presented as a

case study in this article in order to demonstrate the particular

methodologies o pipeline integrity assessment and their

application. This article has been developed in collaboration with

Rosen and JANAF.

The case study or this article concerns the 179 km long, 36 inch

diameter section o the pipeline system rom the Omišalj terminal

to the Sisak terminal. Detailed inormation was provided by

JANAF to enable a thorough assessment to be completed.

Development of integritymanagement strategies for pipelines

With the demand or energy rom an ageing pipeline inrastructure, there is an increasing need to ensure the integrity o assets and extend their sae remaining lie. Gathering sufcient knowledge about the pipeline and knowing how best toanalyse the available inormation has become critical to ensuring the long-term integrity o pipelines. Inspection using in-line intelligent tools provides the clearest picture o the condition o the pipeline. The inormation gathered rom theseinspections can then support numerous integrity related activities, including corrosion assessment, corrosion growthassessment, and remnant-lie assessment and corrosion-management strategy. This article describes the particularmethodologies and demonstrates their application using experience gained with Jadranski Natovod Plc.

By Ivan Cvitanovic, Jadranski Naftovod (JANAF) Plc, Croatia;Uwe Thuenemann, Rosen Europe BV, Netherlands; and,

Cheryl Argent, Chris Lyons, and Andrew Wilde, MACAW Engineering Ltd, UK

Figure 1: Overview of JANAF’s pipeline system. Figure 2: Corrosion concentration at a low point in the elevation prole.




teChniCal

Corrosion assessmentThe aim o the corrosion assessment stage is the diagnosis o

potential corrosion mechanisms. Data relevant to the assessment

are collected, including all available inormation relating to the

basic design, operation, inspection, maintenance, and repair

inormation or the pipeline. These data include:

• Pipeline location and route (pipeline elevation profle,

satellite imagery data, etc.);

• Operational data (MAOP, MOP, product details, chemical

treatments, etc.);

• Inormation on external coating types, specifcations,

application methods and any above-ground coating-survey data;

• Cathodic protection (CP) survey results;

• Previous internal and external inspection data or the pipeline;

• Operational history and details o any repairs;

• Pipeline design and construction details (design criteria,

pipe type/grade, wall thickness, weld data, hydrotest

conditions, etc.); and,

• Current integrity-management plans or the pipeline, orexample, current inspection strategy, routine maintenance

activities, repair criteria, etc.

A review o the ILI fndings is perormed in order to ascertain

the current condition o the pipeline. By considering the shape

(such as pitting, grooving) and location o corrosion within the

line (internal, external, top-o-line, bottom-o-line, girth weld,

etc.), and the collected data on the pipeline, the likely nature

and characteristics o the reported corrosion are diagnosed. In

this way the potential causes o corrosion that are a threat to the

integrity o the pipeline can be identifed.

Case study Rosen Europe conducted a metal-loss internal inspection o

the pipeline segment rom Omišalj to Sisak during May and June

2009, using its Hi-Res MFL and XYZ mapping inspection tool

(CDG) and its Hi-Res axial aw inspection tool (AFD). The pipeline

was inspected in two separate sections: a 74.5 km section between

Omišalj and Dobra, and a 104 km section between Dobra and

Sisak. The pipeline had previously been inspected by Rosen in

2003 using its corrosion detection tool (CDP).

Internal corrosionInternal corrosion in liquid hydrocarbon pipelines usually

ollows a distinctive distribution reecting the origin and location

o ree water within the pipeline. Some o the consistent patterns

o corrosion that occur in liquid pipelines are as ollows:

• Water carry-over: water enters the pipeline as a separate

phase and creates corrosion in the bottom o the line

immediately downstream o the inlet.

• Water separation: water enters the pipeline as a water–in–oil

emulsion which breaks down over time to cause corrosion in

the bottom o the line some distance rom the inlet.

• Water pooling: the corrosion distribution in a pipeline can be

modifed by water pooling at low points whether the origin is

rom carry over or separation.

• Water hold-up: i the elevation profle o a pipeline includes

steep up-slopes in the direction o ow water hold-up can

occur at the oot o these slopes. The corrosion distribution

is similar to that or water pooling but the oil-water interace

tends to be rather more turbulent and corrosion pitting canbe concentrated at this interace.

The pattern o internal corrosion eatures within the pipeline

was consistent with water carry-over, water pooling, and water

hold-up. The corrosion had developed immediately downstream

rom the inlet, diminishing with distance, indicating water

had entered the line in a separate phase, carried-over rom the

Omišalj terminal. Local concentrations o corrosion were present

depending on the topography o the line, with the water pooling

at low points and water hold-up at the base o upslope sections.

One such concentration is shown in Figure 2, where water pooling

has occurred in the bottom o the pipeline at a low point in the

elevation profle.

The product within the pipeline has a very low water content

o 0.1–0.2 per cent volume/volume (v/v). The issues described

above are not normally seen in lines where the water content

is below 0.5 per cent v/v, and it was thereore concluded that

the distribution o internal corrosion eatures in this pipeline

may have been caused as a result o operational issues at some

stage during the lie o the pipeline. Indeed, during the period

o wartime rom 1991–95 the pipeline was let dormant with

the product shut-in. Such conditions would have undoubtedly

avoured corrosion via the pooling o water. Consequently, it was

considered likely that the internal corrosion eatures were no

longer active. This was investigated in detail during a comparison

o the two sets o inspection data and the fndings are discussedlater in this article.

External corrosionThe pipeline is protected against external corrosion by an

external coating and an impressed-current CP system. The

coating on the longitudinally-welded sections o the line is mill-

applied Densolene polyethylene, whereas or the spirally-welded

sections the coating was feld applied. Polyken tape was used

locally where repairs have been made.

The majority o external corrosion was reported in the bottom

hal o the pipeline and located in the spirally-welded sections

where a feld-applied tape coating had been applied. JANAF’s CPmonitoring had shown that adequate protection levels were being

achieved and thereore the most likely cause o the corrosion was

CP shielding which occurs in areas where wrinkling or sagging Figure 4: Example demonstrating search window technique to account

for location reporting accuracy.




teChniCal

o the tape coating allows water and soil ingress beneath the

coating. The CP system can become shielded rom the active

corrosion by the coating.

A concentration o external corrosion was associated with

an above-ground section o line where the pipeline was not

protected by the CP system and the external paint coating was

the only orm o corrosion protection. Consequently, any coating

deects in this area would have been at risk rom atmospheric

external corrosion and it is likely that this was the cause o the

corrosion concentration. The corrosion in this area was identifed

at an early stage and the corrosion had not impacted on the

integrity o the pipeline.

JANAF indicated that the pipeline was subject to stray current

intererence rom numerous DC railway crossings. Stray currents

can represent an external corrosion risk and it is understood

that the pipeline was directly aected up to a distance o

approximately 50 km and indirectly inuenced up to a distance

o approximately 100 km. To proactively counteract these eects

JANAF perormed a modernisation o the pipeline CP systemby installing remote-controlled transormer-rectifer (T/R) units

along the line in order to prevent corrosion in these areas. A key

requirement o the inspection comparison was to review the

eectiveness o these mitigation measures by confrming the

extent o any corrosion activity in these sections o the pipeline.

Corrosion growth assessmentThe corrosion growth assessment (CGA) was conducted in

three stages:

• Feature matching;

• Inspection signal comparison; and,

• Detailed eature resizing.

Each stage o the CGA process is described in detail below.

Feature matchingFeature matching involves the automatic comparison o

two or more sets o inspection data using the eature list

rom each inspection. Beore repeat sets o inspection data

can be compared, they must frst be aligned to a common

‘master’ distance. To ensure sucient accuracy in the distance

correlations, each girth weld should be matched in both o

the inspections. Aligning two sets o data using other pipeline

eatures such as valves, tees, or bends may not provide a

suciently-accurate distance correlation.

Once the inspection data sets are aligned, they can be

compared to identiy eatures that have been reported in both

inspections. For each corrosion eature reported in the recent

inspection, the previous inspection data was reviewed to generate

a list o possible matches. This process takes account o the

errors associated with the axial and circumerential positions by

defning a search window around the eature.

Feature matching provides a quick way o comparing repeat

sets o inspection data and can provide useul inormation that

is applicable to the whole pipeline, not just a sample o the

reported eatures. However, all inspection tools contain a degree

o variation which will aect the accuracy o the inormation

obtained rom a eature matching exercise. Variations include:

• Reporting threshold/detection capability;

• Feature-sizing accuracy; and,

• Axial and circumerential location accuracy.

Unless the above actors are considered within the corrosion-growth comparison, they can result in incorrect matches and

unrealistic corrosion rates. Each actor is discussed below.

The next stages o the corrosion-growth assessment (signal

comparison and detailed resizing) are aimed at removing or

reducing the aect o these variations, and improving the

accuracy o the CGA, and are described in the ollowing.

Reporting threshold/detection capability Inspection results tend to have a reporting threshold (typically

10 per cent wt) imposed upon them due to the large number

o shallow metal-loss indications recorded. In addition, the

sophistication o ILI tools has increased dramatically since their

introduction approximately 40 years ago; consequently, their

ability to detect a given eature has also improved. Both o these

points should be considered when using the results o eature

matching, since eatures that have been reported or the frst

time may have been present but not detected or reported by the

previous inspection; they may not necessarily have grown.

This uncertainty is dealt with by conducting stages 2 and

3 o the CGA process – signal comparison and detailed eature

resizing.

Figure 4: Normalised signal amplitude based on xed reference

points such as girth welds. Figure 5: Signal comparison example.




teChniCal

Feature-sizing accuracy Currently, ILI tools typically report general corrosion with

an accuracy o +/-10 per cent o the wall thickness with a

confdence limit o 80 per cent, i.e. 80 per cent o the time, the

reported depth will be within +/-10 per cent o the actual eature

depth.

The accuracy o a comparison o the depths o a eature

reported in two inspections is dependent on the sizing accuracy

o both tools. An increase (or decrease) in the reported depth

o a eature between inspections can thereore be attributed to

either real corrosion growth or to a variation in depth sizing in

each inspection.

The sizing accuracy o each tool can be used to calculate

the minimum corrosion rate that would be considered to be

statistically signifcant. Below this statistically-signifcant limit,

an increase in reported depths can be attributed to either real

corrosion growth or just variation in the sizing.

For example, consider a general corrosion eature1 located in

10.31-mm wall thickness pipe reported in both 2003 and 2009.Assuming both inspection tools have a depth sizing accuracy o

+/-10 per cent wt at 80 per cent confdence, then the accuracy

o the calculated corrosion growth rate is +/-0.22 mm/a (95 per

cent confdence level). This accuracy can be improved upon

by conducting stage 3 o the CGA process: the detailed eature

resizing.

Axial and circumferential location accuracy The location accuracy is dealt with in two ways in the eature-

matching process: frstly by aligning the eatures in terms o

reported distance, and secondly by using a search window

to match eatures reported by the previous inspection, as

illustrated in Figure 3.

In addition, inspection vendors dier on how they report

eature locations. For example, some vendors use the top let

corner o the eature ‘box’ (as seen in Figure 3), while others

report the centre o the ‘box’. When comparing multiple vendor

inspections the eature data should be adjusted so that all

inspections are reerencing the same location.

Inspection signal comparisonThis stage o the CGA involves the comparison o the raw

inspection signals in order to validate the growth rates

calculated by the eature matching process. Areas selected

or signal comparison include areas o high corrosion growth,apparent ‘new’ growth, and corrosion ‘hotspots’. The signal

comparison eliminates some o the uncertainties related to

the eature-matching process. It is possible to confrm whether

eatures identifed as ‘new’ by the eature matching were in act

present at the time o the previous inspection, although not

reported as they were sized below the reporting threshold. The

signal comparison also validates the matching process.

In order accurately to compare MFL signal data, it is

necessary to normalise the signal amplitude based on a fxed

reerence point (i.e. eatures that will not have changed between

inspections), such as eatures that have been repaired by

composite wrap, eatures that have been recoated immediatelyater the previous inspection, artifcial deects, or girth welds,

(see Figure 4). The signal comparison is not limited to same-

vendor inspections; however, additional data manipulation may

be required such as re-orientation o the signals.

An example signal comparison is shown in Figure 5 that

shows growth o existing and new eatures.

Detailed feature resizingIn this stage, a selected number o eatures identifed in the

earlier stages are subject to detailed analysis and resizing in

order to ensure equivalent interpretation o the inspection

signals in both inspections. This stage makes use o the latest

sizing model and algorithm compensation, and takes into

account the assessment o systematic errors and a calibration

correction. These processes are discussed in urther detail

below.

It should be noted that this process is conducted by the

inspection vendor as access to proprietary sotware and sizing

models is required. This type o analysis is only possible or

repeat inspections by the same vendor.

Interpretation of inspection signalsPart o the data-analysis process or a given eature involves

the interpretation o the shape o the MFL signals. In certain

instances, there may be more than one plausible interpretation:

or example, a certain MFL signal shape could be caused by

either one, wide eature or by two very close pinhole eatures.

A dierence in the interpretation o a given signal can result in a

dierence in reported depths.

Sizing model and algorithm compensation

Sizing models and sizing algorithms are regularly reviewedand improved based upon additional pull-through tests and

in-feld verifcation data. The use o dierent sizing models or

algorithms can result in small sizing dierences. Thereore the

older sets o inspection data are re-analysed using the latest

sizing models and algorithms.

Assessment of systematic errors and calibration correctionIn addition to the small inuence rom varying sizing models,

each ILI is aected not only by scattering o results, but also by

small shits or osets o the calibration, and these are reerred

to as systematic errors. These are corrected using verifcation

data and eatures known to have remained the same between

inspections, such as repaired eatures, artifcial deects andmilling eatures.

The detailed resizing process described above signifcantly

improves the accuracy o the calculated growth rate. This

is demonstrated by the example quoted previously where

a corrosion growth rate calculated by the eature-matching

process would have an accuracy o +/-0.22 mm/a. The corrosion-

growth rate associated with this same eature ollowing detailed

resizing would have an improved accuracy o +/-0.09 mm/a

(at 95 per cent confdence).

Corrosion growth assessment summary

Each stage o the CGA improves upon the accuracy o theprevious one. Feature-matching gives an indication o areas o

1. As dened by Pipeline Operator’s Forum ‘Specications and requirements for intelligent pig inspection of pipelines’, January, 2005.




teChniCal

highest activity, and matched eatures give a good indication

o the actual growth rate. However unmatched eatures can

give an overestimate o corrosion rate as they may not be new

growth but existing eatures that were below the reporting

threshold o the previous inspection.

The signal comparison stage can confrm whether unmatched

eatures are in act evidence o new growth and eliminate any

erroneous high rates. However, the signal comparison is still

dependant on the sizing accuracy o both inspection tools. The

detailed eature resizing ensures equivalent interpretation o

the inspection signals and reduces the associated sizing errors,

thus signifcantly improving the accuracy o the calculated

corrosion rates.

Case study In the CGA conducted on JANAF’s pipeline all eatures were

included in the eature-matching process, the signifcant areas

o corrosion growth identifed by eature matching were then

subject to signal comparison. From the fndings o the signalcomparison 101 eatures were selected or detailed resizing.

The CGA revealed that there was no evidence o internal

corrosion growth throughout the pipeline, thus confrming

the conclusions o the corrosion assessment that the internal

corrosion occurred was historical and most likely occurred

when the pipeline was shut-in with product.

There was no evidence o signifcant corrosion growth in the

areas believed to be aected by stray current. This confrmed

that the control measures implemented by JANAF were

mitigating the problem.

The CGA confrmed evidence o external corrosion growth in

the spiral-welded sections o pipe where a feld-applied tape

coating was used. As described previously, the CP monitoring

data indicated that adequate protection levels had been

achieved, and thereore the most likely cause o this corrosion

was CP shielding. Low-level corrosion growth was identifed

on the above-ground section that was not protected by the CP

system.

Pipeline segmentationLocal variation in ground and coating conditions along the

pipeline route will lead to a variation in the rate o external

corrosion growth. To reect this variation, the pipeline was

segmented or the application o corrosion growth rates in the

uture integrity assessment. The maximum growth rates oreach segment were applied to all reported eatures within that

segment. It is understood that the majority o eatures will grow

at corrosion rates less than the maximum rate. However, the

highest corrosion rates were confrmed during the CGA and due

to the random nature o corrosion growth, it was not possible to

identiy eatures that may grow at these high rates in the uture.

The segmentation was driven by the ollowing contributing

actors:

• Corrosion concentrations (hotspots);

• Corrosion growth rate;

• Changes in consequences area;

• Pipeline topography;• CP data;

• Soil-resistivity data;

• Changes in coating type; and,

• Changes o pipe manuacturer.

Future integrity and corrosionmanagement strategy

Based on the fndings o the corrosion assessment and CGA

described above, an eective uture integrity and corrosion

management strategy can be developed.

The corrosion growth rates are applied to the eatures, which

are then assessed in terms o their impact on the immediate

and uture integrity o the pipeline using relevant code

guidance (or example, Modifed B31G, Detailed RSTRENG2).

Deect-assessment techniques are not described in detail within

this article. A schedule o uture repairs is generated by this

process.

The active corrosion mechanisms identifed by the corrosion

assessment and CGA enable an appropriate mitigation/

rehabilitation strategy to be developed. In the case o JANAF,the active corrosion mechanisms identifed were CP shielding

and atmospheric corrosion on the above-ground section where

the pipeline cannot be protected by the CP system.

The above-ground section was investigated and active

external corrosion was identifed. This was mitigated by sand

blasting and application o an improved paint coating system

in the aected areas.

It was not possible to mitigate against CP shielding thereore

a sectional recoating/replacement schedule was developed

based on calculated time to repair or individual eatures. In

most cases sectional recoating prior to the date by which the

eatures would require repair is a more cost-eective solution

than repair.

ConclusionsThe main conclusions o this article are:

1. The good communication links set up between

operator, inspection vendor and consultant were

key in developing a ocused and eective integrity-

management strategy.

2. The comprehensive historical and current pipeline

data provided by JANAF enabled accurate diagnoses o

corrosion mechanisms.

3. The detailed corrosion-growth assessment clearly

identifes the active and dormant corrosionmechanisms.

4. A very high degree o accuracy or calculated corrosion

growth rates can be achieved by conducting signal

comparison and detailed resizing.

5. The assessment conducted by MACAW demonstrated

that the internal corrosion mitigation and stray current

mitigation implemented by JANAF were eectively

controlling corrosion activity.

6. The detailed corrosion-growth assessment together

with the detailed pipeline data allowed pipeline

segmentation o high risk and high growth areas or

appropriate application o growth rates in the utureintegrity assessment.

2. PRCI Report PR 3-805 ‘A modied criterion for evaluating the remaining strength of corroded pipe’ 22nd December, 1989.




teChniCal

7. The corrosion assessment and corrosion growth

assessment build up a very clear picture o the current

condition o the pipeline and can be used to predict the

uture condition o the pipeline and any repairs and

rehabilitation that may be required.

8. In most cases, it is more cost-eective to prevent or

limit urther growth by implementing corrosion-control

measures than it is to conduct repairs. Mitigation can be

targeted at the active corrosion mechanisms identifed by

the assessment and a re-inspection interval can be set to

confrm the eectiveness o such actions.

This article was presented at the Pipeline Technology

Conerence held in Hannover, Germany, in April 2010, and

organised by EITEP.

premium digest december 2010 development of integrity management strategies for pipelines

Documents