E L S E V I E R Nuclear Engineering and Design 168 (1997) 23-34

Nuclear E needng

Design

Dynamic behaviour of anchors in cracked and uncracked concrete: a progress report 1

M i l t o n R o d r i g u e z a, Y o n g - g a n g Z h a n g a, D i e t e r L o t e z a, H e r m a n L. G r a v e s I I I b,

R i c h a r d E. K l i n g n e r a

a Department of Civil Engineering, The University of Texas, Austin, TX 78712, USA b US Nuclear Regulatory Commission, Washington, DC 20555, USA

Received 16 November 1995

Abstract

In early 1993, the US Nuclear Regulatory Commission began a research program at The University of Texas at Austin, dealing with the dynamic behavior of anchors in cracked and uncracked concrete. In this paper, the progress of that research program is reviewed. The test program is summarized, and work performed to date is reviewed, with emphasis on the dynamic and static behavior of single tensile anchors in uncracked concrete. General conclusions from that work are discussed, and future plans are presented. © 1997 Elsevier Science S.A.

1. Objectives and scope of this paper

The objectives of this paper are to describe the overall organization of this research program, to summarize the test program and to review work performed to date, with emphasis on the dynamic and static behavior of single tensile anchors in uncracked concrete. General conclusions from that work are discussed, and future plans are presented.

2. Background

As stated in the original Request for Proposal of the US Nuclear Regulatory Commission (US-

22 WRSIM.

0029-5493/97/$17.00 © 1997 Published by Elsevier Science S.A. All PII S0029-5493(96)01298-8

NRC), the overall objective of the research de- scribed in this paper is to verify, by testing, the adequacy of the assumption used in US nuclear power plant designs that the behavior and strength of anchor bolts (cast in place, expansion and undercut) and their supporting concrete un- der seismic loads do not differ significantly from those for static conditions. That objective deter- mined many of the research decisions that are discussed in this paper.

At the start of the project, the researchers and the USNRC agreed that the research should con- centrate on the behavior, under dynamic loads, of anchors generally used in existing nuclear plants, particularly for equipment mounting. Based on information furnished by the USNRC, it was agreed to test a particular wedge-type expansion anchor (called here expansion anchor and expan-

rights reserved.

24 M. Rodriguez et al./Nuclear Engineering and Design 168 (1997) 23-34

sion anchor II), a particular undercut anchor (U/C anchor 1), and also some anchors placed in cored holes using cementitious grout (grouted an- chor). Later in the testing program, other anchors were added (U/C anchor 2 and sleeve anchor).

3. OveraH organization of research program

The research program, with a scheduled dura- tion of 3 years, comprises five tasks:

task l: static and dynamic behavior of single tensile anchors;

task 2: static and dynamic behavior of multiple tensile anchors;

task 3: static and dynamic behavior of near- edge anchors;

task 4: static and dynamic behavior of multiple- anchor connections;

task 5: submission of final report.

At the end of each task, an interim report is to be submitted to the USNRC. At the present time, work is essentially complete on the tasks 1 and 2 and is beginning on task 3. Each task is divided into subtasks. The subtasks comprising tasks 1 and 2 are discussed in more detail in later sections of this paper.

4. Materials, equipment and test methods

extend the test results to other concrete strengths. Concrete with soft limestone aggregate generally represents the most conservative case. However, some concrete with river gravel aggregate of medium hardness is also being used.

4.2. Anchors

On the basis of surveys of existing anchors in nuclear applications, the USNRC was primarily interested in documenting the behavior of selected wedge-type expansion anchors, of selected under- cut anchors and also of anchors in cementitious grout. The testing program for tasks 1 and 2 originally emphasized one wedge-type expansion anchor (referred to here as expansion anchor), with some tests on one undercut anchor (U/C anchor 1), and other tests on anchors in one type of cementitious grout. As described later in this paper, several anchors were added: a variant on the expansion anchor (expansion anchor II), an- other undercut anchor (U/C anchor 2) and a heavy-duty sleeve-type expansion anchor (sleeve anchor 1). On the basis of current use in nuclear applications, it was decided to test anchors rang- ing in diameter from -38 to 1 in, with emphasis on ¼ in diameter. In September 1993, the technical community on anchorage to concrete was advised in general terms of the research plans by means of a letter disseminated to American Concrete Insti- tute (ACI) Committee 349 (Nuclear Structures) and ACI Committee 355 (Anchorage to Con- crete).

4.3. Embedment depths

4.1. Concrete strength and characteristics

Preliminary USNRC data indicated that most concrete found in existing nuclear plants has a compressive strength between 4000 and 5500 lbf in-2 (28 and 38 MPa). Some may be as weak as 3000 lbf in -2 (21 MPa). The testing program therefore emphasizes concrete with compressive strengths between 4000 and 5500 lbf in 2 (28 and 38 MPa) and includes some tests in 3000 lbf in-2 (21 MPa) concrete. Analytical research (studies with non-linear fracture-type finite elements) will be verified using test results and then used to

For nuclear applications, the overall anchor design objective is to have failure governed by yield and fracture of anchor steel. Because the dynamic behavior of the anchor steel itself is relatively well understood, in tasks 1 and 2 of this testing program the embedments were shallow enough that behavior would be governed by pull- through (anchor through sleeve), pull-out (anchor through concrete) or concrete cone break-out. Information on dynamic capacity as governed by those failure modes can be used to ensure ductile behavior in later phases of the testing program, and also in practice.

M. Rodriguez et al./Nuclear Engineering and Design 168 (1997) 23 34 25

4.4. Anchor installation

All anchors were installed according to manu- facturer's instructions. To duplicate the effects of loss of pre-stress due to concrete relaxation, an- chors were torqued to manufacturer's specifica- tions, and then they were loosened and re-torqued to half the specified value.

4.5. Load&g

Most static loads in tasks 1 and 2 were applied using a hand-controlled electrical pump, applying displacement at a constant rate. For dynamic loads, two quite distinct loading patterns were used. For the initial dynamic testing covered in tasks 1 and 2, it was desired to find the effect of earthquake-type dynamic loading on anchor ca- pacity as governed by factors other than steel failure. Equipment response to strong earthquake depends on the earthquake and also on the dy- namic characteristics of the equipment. Unlike fatigue loading, earthquake response usually con- sists of relatively few reversed cycles of load, at frequencies of 3 Hz or less. It was initially consid- ered appropriate to subject anchors to such pulses. However, for purposes of this test pro- gram, it was necessary to load the anchors dy- namically to failure, as governed by mechanisms other than steel yield and fracture.

Previous research (Collins et al., 1989) had shown that anchors loaded by triangular pulses would not fail under low cycle fatigue unless the load level exceeded the static failure load. It would therefore be necessary to subject the an- chors to a dynamic triangular load pulse whose magnitude would need to exceed the anchor ca- pacity (which would not be known in advance). Under these circumstances, it was reasoned that it would make no difference whether the pulse were a triangle or simply an increasing ramp load, since the test would be ended in any event by anchor failure. As a result, the dynamic load selected for tasks 1 and 2 was a ramp load to failure. As shown in Fig. 1, the rise time of this load (about 0.1 s) was set to correspond to that of typical earthquake response. For subsequent phases of the testing program, a reversed cyclic loading

history representing earthquake response will be used.

4.6. Reinforcement

Most specimens were designed without rein- forcement in the area that would be affected by the concrete break-out cone. Specimens in series labeled 'reinforced concrete' had a curtain of grade 8 bars (13 ram) spaced at 8 in (200 mm) in each direction, placed with 1½ in (38 mm) and intended to simulate reinforcement in a heavily reinforced wall.

4. 7. Procedures used for cracked concrete testing

Because this investigation is concerned with seismic behavior, it was necessary to investigate the effects of concrete cracking on anchor behav- ior under dynamic loads. For this purpose, the European Union of Agr~ment (1992) criterion of a crack with a surface width of 0.3 mm was selected. Several procedures are now accepted for imposing this type of crack. 1. A large block of concrete can be subjected to

tension, producing cracks. This procedure re- quires a reaction frame, and the resulting cracks are not uniform in width.

2. A block of concrete can be cast with metal plates on it. These act as crack formers when the block is placed in tension. This procedure is relatively complex and only permits cracks to form where the plates are placed.

3. Cracks can be formed using split tubes placed on both sides of the planned anchor location. The split tubes are expanded by wedges to

"O 1- 0.9,

0 .-I 0.8- E o7 :~ 0.6 E • - 0.5 K 0.4

0.3 "0 0.2 m 0 0.1 "J 0

/ i i ' Anchor Capacity < Maximum Load

l . . . . . . Pulse Load

Ramp Load

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Time, sec

Fig. 1. Ramp-type dynamic loading used in tasks ! and 2.

26 M. Rodriguez et a l . / Nuclear Engineering and Design 168 (1997) 23-34

open the crack to the desired width. This proce- dure is relatively simple and has the additional advantage of permitting cracks to be located anywhere on the surface of the concrete.

It was decided to use the third method, with the following test sequence.

(a) The crack was started using wedges and was reduced to hairline width. (b) The anchor was installed at the location of the crack. (c) The crack was widened to 0.3 mm at the surface. (d) The anchor was tested. The crack width was monitored, but was not controlled.

5. Detailed information on task 1 testing

5.1. Original outline for task 1 testing (August 1993)

The original outline for task 1 testing (August 1993) was as follows: 1. static and dynamic behavior of single tensile

anchors: (a) static tensile tests of single anchors in unreinforced uncracked concrete; (b) dynamic tensile tests of single anchors in unreinforced uncracked concrete; (c) static tensile tests of single anchors in reinforced uncracked concrete; (d) dynamic tensile tests of single anchors in reinforced uncracked concrete; (e) static tensile tests of single anchors in unreinforced cracked concrete; (f) dynamic tensile tests of single anchors in unreinforced cracked concrete; (g) static tensile tests of single anchors in reinforced cracked concrete; (h) dynamic tensile tests of single anchors in reinforced cracked concrete,

Because the original outline for task 1 was superseded in several respects (as explained in the next sections), the original test matrix is not given here. The final test matrix is given in Table 1.

5.2. Comparison of original and current expansion anchors

It had been known from the beginning of the test program that the wedge-type expansion an- chor emphasized in the original task 1 testing program was no longer commercially available in its original form (that found in many nuclear power plants). The current version of the anchor (identified here as expansion anchor II) was thought to behave quite differently from the pre- vious version.

As a result, the static and dynamic perfor- mances of the original expansion anchor were compared with those of the current expansion anchor II. Efforts were made to select embedment depths such that all anchors would experience cone break-out failure.

Results obtained from comparison testing showed that the original expansion anchors of 43- in, installed at embedments equal to or greater than the manufacturer's standard embedment of 43 in (121 mm), failed by pull-through of the anchor rather than by concrete break-out. Be- cause the expansion anchor I! has an improved performance in this regard, it was concluded that test results with the expansion anchor II at those embedments should not be assumed to apply to the original expansion anchor.

According to USNRC estimates, more than three quarters of the original expansion anchors were installed at embedments equal to or greater than the manufacturer's standard embedment. Therefore, in order to satisfy the specific objec- tives of this research project, it was believed nec- essary to conduct at least some static and dynamic tests on original expansion anchors, in- stalled at the manufacturer's standard embed- ment. If those tests indicated that the pull-through capacity of the original expansion anchors was not decreased under dynamic loading conditions, then subsequent tests would be conducted at the smaller embedment depths associated with con- crete cone break-out, down to the manufacturer's minimum embedment of 3¼ in (83 mm). At those embedment depths, the behavior of the expansion anchor II was expected to be the same as that of the original expansion anchor. This would also be verified by test.

M. Rodriguez et al./Nuclear Engineering and Design 168 (1997) 23 34 27

Table 1 Final test matrix for task 1

Series Description Concrete Anchors tested Number of an- Number of strength chors tests

1-0 Static comparison tests of ex- 4700 lbf in 2 Expansion anchor 3 Expan- 5 5 pansion anchor and expansion limestone sion anchor II, 3 anchor II at s tandard embed- merit

Dynamic comparison tests of 4700 lbf in 2 5 5 Expansion anchor and Expan- limestone sion anchor II

Static comparison tests of ex- 4700 lbf in -2 5 5 pansion anchor and expansion limestone anchor II at min imum embed- ment

Dynamic comparison tests of ex- 4700 lbf in -2 5 5 pansion anchor and expansion limestone anchor II at min imum embed- ment

expansion anchor, ~ expan- sion anchor II, 3

expansion anchor, 3 expan- sion anchor lI, 3

expansion anchor, 3 expan- 3 sion anchor II, a

1-1 Static tensile tests of single an- chors in unreinforced concrete

4700 lbf in -2 expansion anchor II, 3 ¼ 10 10

3000 lbf in -2 expansion anchor II, ~ 5 5 limestone

3 4700 lbf in 2 expansion anchor II, a 5 5 river gravel

3000 lbf in -2 expansion anchor II, 3 5 5 river gravel

4700 lbf in 2 U/C anchor 1, 3 5 5 limestone 4700 lbf in 2 Grouted anchor, ¼ 5 5 limestone

1-2 Dynamic tensile tests of single anchors in unreinforced concrete

4700 lbf in -2 expansion anchor II, 3, 3 10 10

3000 lbf in -2 expansion anchor II, ¼ 5 5 limestone

4700 lbf in -2 expansion anchor II, j 5 5 river gravel

3000 lbf in -2 expansion anchor II, ~ 5 5 river gravel

4700 lbf in 2 U/C anchor I, 3 5 5 limestone 4700 lbf in -2 Grouted anchor I, ~ 5 5 limestone

1-3

1-4

Static tensile tests o f single an- chors in reinforced concrete

Dynamic tensile tests of single anchors in reinforced concrete

4700 lbf in--" expansion anchor 11, ~ 5 5 limestone

4700 lbf in e expansion anchor II, J 5 5 limestone

28 M. Rodriguez et al . / Nuclear Engineering and Design 168 (1997) 23-34

Table 1 (continued)

Series Description Concrete Anchors tested Number of an- Number of strength chors tests

I-5 Additional static tensile tests of 4700 lbf in -~ U/C anchor 1, 3 5 5 single anchors in unreinforced limestone concrete

4700 lbf in -2 U/C anchor 2, ~ 5 5 limestone

4700 lbf in -2 Sleeve anchor, M10, M20 10 10 limestone

4700 lbf in -2 U/C anchor 1, ¼ 5 5 river gravel

3000 lbf in -2 U/C anchor 1, ¼ 5 5 limestone

3 1-6 Additional dynamic tensile tests of 4700 lbf in -2 U/C anchor 1, ~ 5 5 single anchors in unreinforced limestone concrete

4700 lbf in -2 Sleeve anchor M10, M20 10 10 limestone

4700 lbf in 2 U/C anchor 1, -] 5 5 river gravel

3000 lbf in 2 U/C anchor 1, ¼ 5 5 limestone

1-7 Static tensile tests of single an- 4700 lbf in -2 expansion anchor II, ¼ 5 5 chors in unreinforced, cracked limestone concrete

4700 lbf in -2 U/C anchor 1, 3, ¼ l0 10 limestone

3 4700 lbf in -2 U/C anchor 2, ~ 5 5 limestone

4700 lbf in -2 Sleeve anchor, MI0, M20 10 10 limestone

4700 lbf in 2 Grouted anchor, 3 5 5 limestone

1-8 Dynamic tensile tests of single an- 4700 lbf in 2 expansion anchor lI, J 5 5 chors in nnreinforeed, cracked limestone concrete

4700 lbf in -2 U/C anchor 1, 3, 3 10 10 limestone 4700 lbf in -2 U/C anchor 2, ¼ 5 5 limestone

4700 lbf in 2 Sleeve, Anchor, MI0, M20 10 10 limestone

4700 lbf in 2 Grouted anchor ¼ 5 5 limestone

M. Rodriguez et al./ Nuclear Engineering and Design 168 (1997) 23 34 29

5.3. Verification of the behavhgral differences between the original expansion anchor and expansion anchor H (December 1993)

As part of series 0 of task 1 (see Table 1), 20 static comparison tests were carried out using the original expansion anchor and the expansion anchor II,

3 both of 3 in diameter, and embedded in 4700 lbf in- 2 (32.4 MPa) limestone aggregate concrete. In each case, two embedment depths were used: the manufacturer's standard embedment of 43 in (121 ram), and the minimum embedment of 3¼ in (83 mm). All measured capacities were normalized by the square root of the measured compressive strength of the concrete used.

Results showed that under static load, and using the standard embedment, expansion anchor II does not fail in the same way as the original expansion anchor. At the minimum embedment, the expan- sion anchor II does fail in the same way as the original expansion anchor, but at a higher load and a much smaller displacement. The reason for this increased capacity is probably the improvement in the wedge mechanism of the expansion anchor II, which significantly reduces pull-through.

On the basis of these results, it was clear that subsequent test results with expansion anchor II could not be assumed to apply to the original expansion anchor. It was expected that this would be particularly true in cracked concrete. Given the availability of original expansion anchors, it was proposed to conduct 20 additional dynamic com- parison tests in cracked concrete, and then to conduct no further tests with the original expansion anchor.

Results would then be available from about 60 tests involving direct comparison of the two anchor types, including 30 tests with the original expansion anchor. Using those tests, limited conclusions could be derived regarding the performance of the origi- nal expansion anchor under dynamic loads, includ- ing the effects of cracked concrete. Those results would satisfy the minimum objectives of this pro- ject with respect to that type of existing anchor. However, it would not be possible to include the original expansion anchor in the other phases of the testing program, owing to the lack of anchors. As a result, the rest of the test program would apply

~113/4~4,75 I1113/4~4.00 U 113/4~3+25 j

38.61 38.65 38.17 4O

10

4700 I~ 4700 ps~ 3000 p~ 3000 psi Umestone River Lime~one River

Grav(d Gravel Concrete Strength and Aggregate Type

Fig. 2. Effect of embedment depth, concrete strength and aggregate type on cone break-out capacity of expansion an- chor II, uncracked concrete.

only to the expansion anchor II and to undercut and grouted anchors.

5.4. Extension of task 1 testing to include other anchor types

Because of the above finding, and because of some of the results obtained with wedge-type expansion anchors in the task 1 testing, it was decided to expand the research objectives to include the performance of modern anchors designed spe- cifically to function in cracked concrete. A torque- controlled sleeve anchor with follow-up expansion capability (sleeve anchor) was added to the testing program, as was another undercut anchor (U/C anchor 2).

5.5. Final task 1 test matrix (250 tests)

The final task 1 test matrix, modified as discussed above, and comprising about 250 tests, is shown in Table 1.

6. Typical results from static and dynamic tensile tests on wedge-type expansion anchors in uncracked concrete (January 1994)

6.1. Effect of embedment depth, concrete strength and aggregate type on tensile capacity of expansion anchors in uncracked concrete

Fig. 2 shows the tensile capacities (normalized 0.5 1.5 by fc he ), for expansion anchor II, for different

30 M. Rodriguez et al . / Nuclear Engineering and Design 168 (1997) 23-34

concrete strengths and aggregate types. Each bar represents the average of five tests. In Fig. 2 and similar figures, the legend has the following mean- ing: II3/4@4.75 refers to expansion anchor II of 3 in (19 mm) diameter and 4.75 in (121 ram) embed- ment; II3/4@4.00 refers to expansion anchor II of 3 in diameter and 4.00 in (102 mm) embedment; II3/4@3.25 refers to expansion anchor II of 3 in diameter and 3.25 in (83 mm) embedment; II3/ 8@2.25 refers to expansion anchor II of 3 in (9.5 mm) diameter and 2.25 in (57 ram) embedment.

The figure shows that, as the embedment depth decreases, the normalized capacity increases. The reason for this is that, at deeper embedments, anchor capacity (not normalized) increase, and the anchor is more prone to pull-through and pull-out. Such anchors slip until the embedment depth is small enough to produce a cone break- out, at an embedment depth considerably smaller than the original value. Because capacity is nor- malized using the original embedment, the nor- malized capacity is artificially small. The graph also suggests that normalized tensile capacity is not significantly affected by concrete compressive strength or aggregate type. These point are dis- cussed in more detail below.

6.2. Effect of loading rate (static vs. dynamic') on tensile capacity of wedge-type expansion anchors

Fig. 3 shows the tensile capacity, normalized by 0.5 1.5 (35.05f~ he ), for the original expansion anchors

and for expansion anchor II. The figure shows that the tensile capacity of the original expansion anchor is reduced by about 10% under dynamic

~. 0.94 0.99 0.97 o - ~ -

Expansion Expansion Anchor Anchor II

Fig. 3. Effect of loading rate (static vs. dynamic) on tensile capacity of original expansion anchor and expansion anchor If, uncracked concrete.

~. 50 ,r C PIC C C ~ 4 o p ~ C P/C ~ _ PIC !

30 P (';

2O

0r314~4.75 0r3/403.25 113/4@4,75 11314~3,25 113/8@2.25

Fig. 4. Effect of loading rate (static vs. dynamic) on failure mode of wedge-type expansion anchors, uncracked concrete.

loading. The tensile capacity of the expansion anchor II is about the same. Results are obtained from all concrete strengths, aggregate types, an- chor diameters and embedment depths.

6.3. Effect of loading rate (static vs. dynamic) on failure mode of wedge-type expansion anchors

Fig. 4 shows the tensile capacity (normalized by f°Sh~5), for the original expansion anchor and expansion anchor II, both of 3 in diameter, as a function of embedment depth. Nominal concrete strength was constant 4700 lbf in-2 (32.4 MPa), and limestone aggregate was used.

The graph shows that, at an embedment depth of 4.75 in (121 mm), the original expansion an- chor fails by pull-through under static and dy- namic loads. Dynamic capacity is less than static capacity. As the embedment depth is decreased to 3.25 in (83 ram), the statically loaded anchor fails by cone breakout. While the capacity of the dy- namically loaded anchor is about the same, the failure mode for some of the five replicates is pull-through. Expansion anchor II has a greater tendency toward cone failure rather than puU- through. In general, the capacities of both expan- sion anchors are slightly less under dynamic load than under static load.

6.4. Effect of loaditig rate (static vs. dynamic) on tensile cone break-out capacity of undercut and grouted anchors

Fig. 5 shows the effect of loading rate on the tensile capacity, as governed by concrete cone break-out, of the U/C anchor 1 and the grouted anchor. All capacities are normalized by f°cShl5.

M. Rodriguez et al./ Nuclear Engineering and Design 168 (1997) 23-34 31

6O ~.. 50 o .0 40 ~: 30

20

.? lO

U/C Anchor I Grouted

Fig. 5. Effect of loading rate (static vs. dynamic) on tensile cone break-out capacity of U/C 1 and grouted anchors, un- cracked concrete.

Results show an increase in normalized capacity of about 30%.

6.5. Effect o f reinforcement on tensile capacity

Fig. 6 shows the effect of reinforcement on the normalized tensile capacity of anchors loaded statically. All capacities are normalized h- ,~0.5,, 1.5 - Y J c H e •

All anchors and all concrete strengths are in- cluded. Effects of reinforcement are seen to be negligible. T h i s observation applies to reinforce- ment placed parallel to the free surface of the specimen, perpendicular to the axis of the anchor.

6.6. Effect o f aggregate type on tensile capacity

Fig. 7 shows the effect of aggregate type on the normalized tensile capacity of all anchors. For both concrete strengths, changing from soft lime- stone aggregate to medium-hard river gravel ag- gregate has no significant effect on tensile capacity.

4O

35 I 30 25

• = 26

~ lO 5

0

36.1 35.21

All Anchors

Fig. 6. Effect of reinforcement on tensile cone break-out resistance of all anchors, uncracked concrete, static loading.

38.65 38A7 I [] Limestone 1

• River Gravel 36.t 35.3

3000 psi 4700 psi

Concrete Strength

Fig. 7. Effect of aggregate type on normalized tensile capacity of all anchors, uncracked concrete, static loading.

7. Preliminary conclusions from static and dynamic tests on wedge anchors in uncracked concrete (January 1994)

The following conclusions were presented in general form at meetings of ACI Committee 349 (Nuclear Structures) and ACI Committee 355 (Anchorage to Concrete) in March 1994. 1. Wedge-type expansion anchors have approxi-

mately the same capacity under dynamic load, as under static load. However, this average is deceptive, because it results from combining two distinct failure modes.

(a) For wedge-type expansion anchors, if failure is by concrete cone break-out, dy- namic capacity exceeds static capacity. Con- crete cone break-out capacity under static loads is predicted with a coefficient of varia- tion of about 10%, by the CC method (for- merly known as the kappa method) (Comit6 Euro-international du Beton, 1991) ( P =

0 5 1 5 35fo he ). The multiplicative constant (ide- ally, 35) increases as the effective depth decreases. For wedge anchors, the effective embedment he is measured from the free surface of the concrete to the point of the clip in contact with the concrete. Because of pull-through of the mandrel, estimation of the effective embedment when the cone is produced is difficult for wedge anchors. (b) For wedge-type expansion anchors, if failure is by pull-through, dynamic capacity is less than static capacity,

2. Dynamic loading worsens the performance of wedge-type expansion anchors. It increases the

32 M. Rodriguez et al./Nuclear Engineering and Design 168 (1997) 23 34

tendency for failure by pull-through and pull- out, rather than by concrete cone break-out. Evidently, dynamic loading decreases the coeffi- cient of friction between the cone and the clip (steel to steel), and between the clip and the concrete (steel to concrete).

3. At embedments less than those required to produce steel failure, grouted anchors and U/C anchors fail by concrete cone break-out, without pull-out. The concrete cone break-out capacity of grouted anchors and of U/C anchors is predicted with a coefficient of variation of less than 10%, by the CC method (Comit6 Euro-In- ternational du Beton, 1991) (p = 4 2 f c0.Shel.5).

4. Dynamic loading increases the capacity of grouted anchors and U/C anchors. Under dy- namic loads, the concrete cone break-out capac- ity of grouted anchors and U/C anchors is about 30% greater than the capacity under static loads.

5. For all anchors tested so far, performance in concrete with limestone aggregate is not very different from performance in concrete with river gravel aggregate.

6. For all anchors tested so far, heavy reinforce- ment (grade 8 bars at 8 in, 11 in cover (13 mm bars at 200 mm, with 38 mm cover) placed parallel to the surface of the concrete has no appreciable effect on performance.

8. Successful start of cracked concrete testing (March 1994)

In March 1994, cracked-concrete testing was successfully commenced using the procedure dis- cussed above. Implementation of this testing was facilitated through the loan of equipment and experienced technicians from the Hilti Corpora- tion. Results from cracked-concrete testing will be presented in the near future.

9. Detailed information on task 2 testing

9.1. Original outline for task 2 testing (August 1993)

The original outline for task 2 testing (August 1993) was as follows:

1. Static and dynamic behavior of multiple an- chors:

(a) static tensile tests on multiple anchors; (b) dynamic tensile tests on multiple an- chors; (c) overlapping cone vs. linear interpola- tion approach; (d) static tensile tests of multiple anchors in reinforced concrete; (e) dynamic tensile tests of multiple an- chors in reinforced concrete; (f) submission of preliminary task 2 report to USNRC.

9.2. Modifications to original task 2 testing

The above task 2 testing program had been based on the original project proposal, prepared in 1992. Since that time, the overlapping cone vs. linear interpolation issue has been significantly clarified. It was believed possible to use the linear interpolation procedure with little additional tech- nical justification. It was also proposed to elimi- nate the reinforced-concrete tests, since task 1 testing had shown no significant influence of rein- forcement.

It was proposed that, in addition to a reduced number of multiple-anchor tests, the rest of Task 2 could be used to conduct a series of single-an- chor tests aimed at filling in the gaps in our knowledge of the basic load-deflection behavior of ductile anchors loaded in pure tension (already almost done), pure shear (already almost done), and combined tension and shear (not well known). This information would be useful to calibrate engineering models of load-deflection behavior of single anchors. Those models are necessary to predict the behavior of complex con- nections, and to estimate the design requirements for ductile behavior in such connections.

Based on the above information, in March 1994 the revised test matrix shown in Table 2 was approved for task 2. This test matrix included two series (series 2-1 and 2-2) intended to fulfill the original objective of resolving the cone vs. CC method controversy, but now with emphasis on dynamic loads only. It also contained several test series (series 2-3, 2-4 and 2-5) intended to investi- gate the basic load-deflection behavior of deeply

M. Rodriguez et al./Nuclear Engineering and Design 168 (1997) 23-34 33

"F.

.=o

5

e~

~ ~ 0 O ~ 0

7 e-

~ r-- E

8

,£ .=_

e -

~ E

r ~

_ o o o o _ o _ o o _ o _ o o _ 2

d .-= .-=

~ e a P. N N N "~

.-= ~ .-= 7" - -

S ~ S 8 .o N ~

= . ~ " ° a ..o a - ~

.~=~ -~ ~ ~ ~,~

,-4 ,-4 ,-4 ,.-4 ,-4

34 M. Rodriguez et al./ Nuclear Engineering and Design 168 (1997) 23-34

embedded anchors for use in developing computer programs for predicting the response of multiple- anchor connections. Finally, it contained one test series at shallower embedments (series 2-6), in- tended to help to estimate the behavior of existing anchors designed under previous assumptions, such as the 45 ° cone. The revised task 2 test matrix involves 223 anchors.

the overall load-displacement behavior of an- chors loaded in combinations of axial force and shear. These will be used to predict the behavior of multiple-anchor connections with complex loading conditions.

Analytical results will be presented in the near future.

9.3. Results from task 2 testing Acknowledgements

At the present time, results from task 2 testing are being reviewed by the researchers and by the USNRC. They will be discussed in generic form at technical committee meetings associated with anchorage and will be released in detailed form as soon as possible.

The work described here is sponsored by the U S N R C under Contract NRC-03-92-025 ("An- chor bolt behavior and strength during earth- quakes"). The Contract is administered by the Commission's Office of Research, and the Techni- cal Contact is Herman L. Graves l i e

9.4. Analytical work to date

At the present time, analytical work is proceed- ing from three different view points. 1. Work is continuing on smeared-cracking mod-

els to predict the load-displacement behavior of single anchors far f rom free edges. Reason- able correspondence has been obtained be- tween analytical load-displacement predictions, and experimental results.

2. Work has begun on discrete-cracking models. Results will be compared with those from smeared-cracking models.

3. Work has begun on models intended to predict

References

Fastenings to reinforced concrete and masonry structures: state-of-the-art report, Comit6 Euro-International du Be- ton, Bulletin D'Information 206 and 207, August 1991.

D.M. Collins, R.A. Cook, R.E. Klinger and D. Polyzois, Load-deflection behavior of cast-in-place and retrofit con- crete anchors subjected to static, fatigue, and impact tensile loads, Research Report CTR 1126-1, Center for Trans- portation Research, The University of Texas at Austin, February 1989.

European Union of Agr6ment, UEAtc Technical Guide on Anchors for Use in Cracked and Non-cracked Concrete, M.O.A.T. 49, London, 1992.