Dynamic performance of concrete undercut anchors ...

Nuclear Engineering and Design 265 (2013) 1091–1100

Contents lists available at ScienceDirect

Nuclear Engineering and Design journal homepage: www.elsevier.com/locate/nucengdes

Dynamic performance of concrete undercut anchors for Nuclear Power Plants Christoph Mahrenholtz ∗ , Rolf Eligehausen University of Stuttgart, Pfaffenwaldring 4, 70569 Stuttgart, Germany

h i g h l i g h t s

g r a p h i c a l

a b s t r a c t

• Behavior of undercut anchors under

a r t i c l e

i n f o

Article history: Received 4 May 2013 Received in revised form 12 September 2013 Accepted 26 September 2013

Ultimate load Nu

Displacement s Anchor

Load N

dynamic actions simulating earthquakes. • First high frequency load and crack cycling tests on installed concrete anchors ever. • Comprehensive review of anchor qualification for Nuclear Power Plants.

Quasi-static loading rate High loading rate Load N Stiffness k(0.5Nu) Displacement s

Concrete Load transfer by mechanical interlock of undercut

a b s t r a c t Post-installed anchors are widely used for structural and nonstructural connections to concrete. In many countries, concrete anchors used for Nuclear Power Plants have to be qualified to ensure reliable behavior even under extreme conditions. The tests required for qualification of concrete anchors are carried out at quasi-static loading rates well below the rates to be expected for dynamic actions deriving from earthquakes, airplane impacts or explosions. To investigate potentially beneficial effects of high loading rates and cycling frequencies, performance tests on installed undercut anchors were conducted. After introductory notes on anchor technology and a comprehensive literature review, this paper discusses the qualification of anchors for Nuclear Power Plants and the testing carried out to quantify experimentally the effects of dynamic actions on the load–displacement behavior of undercut anchors. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Because of the robust load transfer mechanism by mechanical interlock, two anchor types are in particular suitable to cope with extreme dynamic actions: cast-in-place headed studs (Fig. 1a) are used in reinforced concrete structures since decades. Though the headed stud is an established anchor type of simple make, the required installation of the base plates fitted with the headed studs prior to concrete casting is a significant disadvantage. The position of cast-in base plates is fixed and do not allow design revisions in terms of loads and layout. Further, welding on site is often problematic in view of quality as well as accessibility and

∗ Corresponding author. Present address: Buesingstraße 6, 12161 Berlin, Germany. Tel.: +49 172 6577516. E-mail address: [email protected] (C. Mahrenholtz). 0029-5493/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.nucengdes.2013.09.038

welded connections are hard to be removed later. Post-installed undercut anchors (Fig. 1b) provide maximum flexibility and are suitable for adding new components at a later stage and may be used for versatile retrofit solutions as well. During installation, the expansion elements either extends into an undercut predrilled by a special drilling apparatus or they create the undercut by a self-cutting action of the anchor. Detailed description of the loadbearing behavior of the anchor types can be found in the literature (Eligehausen et al., 2006). Though post-installed anchors have been used in Nuclear Power Plants (NPP) already in the 1960s and 1970s, their use was not regulated in the beginning and the design approaches varied significantly. In 1979, structural failure of supports of safety related piping systems in the US and questions concerning the performance of expansion anchors led to the issuance of the Inspection and Enforcement Bulletin (IEB 79-02, 1979). Thereafter, in 1980, the American Concrete Institute (ACI) codified the design methodology

1092

C. Mahrenholtz, R. Eligehausen / Nuclear Engineering and Design 265 (2013) 1091–1100

Fig. 1. Examples for anchors suitable to sustain extreme actions: (a) cast-in-place headed stud; (b) post-installed undercut anchor.

of anchors based on limited amount of available test data and developed the ACI 349 (ACI 349, 1980) for the design of nuclear safety related concrete structures. Since then, extensive research has been done on anchorages to concrete structures and many clauses have been updated in the latest edition (ACI 349, 2006). The provisions in ACI 349 were reassessed for their applicability to post-installed anchors in the early 1990s and have been harmonized with the provisions of ACI 318 Appendix D, “Anchoring to Concrete” which are incorporated since 2002 (ACI 318, 2002). Nowadays, Appendix D of ACI 318 (ACI 318, 2011) is the most widely accepted provision for the seismic design of fastenings in the US and presumably worldwide. In Europe, the design of concrete structures for NPPs is generally covered by the provisions given in the reinforced concrete design code Eurocode 2 (EN 1992, 2011). Additional provisions for the design of structures for earthquake resistance are given in the Eurocode 8 (EN, 1998:2006) as well as in national codes like DIN 24449 and KTA2201 in Germany. To date, no European code contains anchor design provisions. Therefore, anchor design is generally carried out according to Annex C of the European Technical Approval Guideline ETAG 001 (ETAG 001, 2013) or the CEN Technical Specification CEN/TS 1992-4 of the European Committee for Standardization (CEN/TS 1992-4, 2009). It is expected that the CEN/TS 1992-4 will be transferred into a harmonized standard as Part 4 of the Eurocode 2 until 2014.

Load N

a)

Ultimate load Nu

Quasi-static loading rate High loading rate

Qualified anchors have to be used for safety relevant connections in the US and Europe. The qualification of concrete anchors is conducted according to the qualification guidelines ACI 355.2 (ACI 355.2, 2007) and ETAG 001 (ETAG 001, 2013) in the US and Europe, respectively. These guidelines provide test conditions and acceptance criteria which must be met prior to issuing an Evaluation Service Report (ESR) in the US or a European Technical Approval (ETA). The ACI 355.2 includes seismic approval testing which, however, is based only on load cycling tests with relatively moderate test conditions. Until recently, ETAG 001 did not cover seismic loading. Only in 2013, Annex E of ETAG 001 (ETAG 001, 2013) was published which covers the qualification of post-installed anchors for seismic loading. For the high safety demands of NPPs including extreme loading caused by earthquakes, airplane impacts or explosions, the German Institute for Civil Engineering (DIBt) published the German Guideline for Anchorages in NPP in 1998 (DIBt KKW Leitfaden, 1998). This guideline recognizes the most demanding tension and shear load cycling as well as crack cycling tests worldwide. In general, qualification tests are always carried out at quasi-static loading rates which are considered to be conservative. It is noteworthy that ACI 355.2 qualified anchors have to be used in many countries as their concrete design codes either refer to or originate from Appendix D of ACI 318 (e.g. Chile, Korea, New Zealand, Taiwan). 2. Motivation and research significance As reinforced concrete structures are subjected to dynamic actions such as seismic excitations, concrete members and concrete anchors embedded therein experience different kinds of cyclic loadings at high rates. As known from general material sciences and in particular from concrete testing, high loading rates generally influence the concrete properties in a positive way (Malvar and Ross, 1998; Sharma et al., 2010; Wesche and Krause, 1972; Zielinski, 1982). Accordingly, concrete anchors subjected to dynamic loading are expected to develop higher peak load capacities and smaller displacements at peak load, but also an increased initial stiffness if compared to quasi-static loading (Fig. 2). Any experimentally measured reduction of anchor displacement and increase of anchor capacity as a result of realistic high loading rates would give evidence that the qualification testing at quasi-static rates contains reserves of resistance and would allow relaxing the test conditions or assessment criteria. To this end, monotonic and cyclic tests in shear and tension as well as crack cycling tests were carried out at earthquake relevant rates and frequencies to quantify the dynamic effects on the anchor load and displacement capacities.

Stiffness k(0.5Nu) 3. Literature review

Displacement s b)

Displacement s Anchor

Load N Concrete Load transfer by mechanical interlock of undercut Fig. 2. Schematic to illustrate (a) the influence of loading rate on the load–displacement behavior of (b) concrete anchors (undercut anchor shown as an example).

There is a high probability that anchors are located in cracks caused by external loading or thermal constraints exceeding the low tensile strength of concrete (Bergmeister, 1988; Eligehausen et al., 1986; Lotze, 1987). For shear loads, the influence of cracks on the load–displacement behavior of anchors is relatively small (Fuchs and Eligehausen, 1989; Vintzeleou and Eligehausen, 1991). For headed studs, the reduction of shear capacity due to a 0.4 mm crack is smaller than 10%. For tension loads, however, cracks have a significant impact on the behavior of anchors (Cannon, 1981; Eligehausen and Balogh, 1995). Though undercut anchors and headed studs are less crack sensitive than any other anchor type, e.g. screw or expansion anchors, 0.4 mm cracks reduce the tension capacity by about 25% (Fig. 3). In conclusion, it is important that anchors are tested in realistic crack widths. For reinforced concrete structures responding to seismic loads, large crack widths in the


1093

Fig. 4. Example crack cycling test series according to the German Guideline for Anchorages in NPP with 10 cycles and permanent load at NRk / Mc (Mahrenholtz, 2012). Fig. 3. Influence of crack width on tension load capacity Nu of undercut anchors and headed studs (Eligehausen and Balogh, 1995).

order of 0.8 mm have to be anticipated (Nuti, 2008; Hoehler and Eligehausen, 2008a). The anchor behavior for load cycling was investigated within various studies using individually developed or standard load protocols (ACI 355.2, 2007; CAN3-N289.4-M86, 1986; SEAOSC, 1997). Anchors loaded cyclically in shear are repeatedly bended and therefore affected by low cycle fatigue. The low cycle fatigue strength may be lower than the monotonic capacity. Researchers evaluated different degrees of load reductions depending on the load level and cycle number the anchor is subjected to (Kim et al., 2004; Klingner et al., 1982; Vintzeleou and Eligehausen, 1991). Due to the irreversible deformation in the concrete imposed by the loaded anchor (compaction, spalling of concrete in loading direction) which accumulate during load cycling, the load and stress amplitude in the anchor is not linearly related to the anchor displacement. In consequence, the specific loading pattern has a direct impact on the fatigue strength and anchor displacement behavior during cycling (Guillet, 2011; Rieder, 2009; Usami et al., 1981). In particular, the load level during cycling is critical for the residual load capacity and corresponding displacement (Mahrenholtz, 2012). For anchors loaded in tension, the cyclic load–displacement curve generally follows the monotonic mean curve. High resistance to cycling can generally be assumed even for a large number of load cycles at load levels below peak and/or for limited number of load cycles near peak load (Eibl and Keintzel, 1989a; Hoehler and Eligehausen, 2008b). The influence of the loading pattern on the behavioral characteristic at ultimate load is not significant and the effect of the crack width on ultimate load is the same as for monotonic tests (Eligehausen et al., 2006; Weigler and Lieberum, 1984). Tension cycling with a moderate number of cycles does not significantly influence the peak load and corresponding displacement (Mahrenholtz, 2012). To a lesser extent, the influence of crack cycling on anchor behavior was also studied in the past. Crack width amplitudes are comparatively small for crack cycling caused by a change of the working load, however, cracks may open wide and may close completely when the reinforced concrete structure responds to seismic excitations. The investigated seismic crack cycle regimes are mostly based on the analysis of the number of deformation cycles in structures during earthquakes (Malhotra, 2002; Tang and Deans, 1983). During crack cycling, the anchor experiences substantial displacements (Furche, 1987; Lotze and Faoro, 1988; Rehm and Lehmann, 1982). With ongoing crack cycling, the anchor experience significant increase in displacement. Progressive increase of

anchor displacements during cycling indicates immanent failure (Furche, 1988; Seghezzi, 1985). In most studies, the crack width was cycled between a specific maximum crack width and the crack width present when the concrete specimen is unloaded. Only few researchers compressed the crack to simulate realistically the actual conditions of reinforced concrete structures under seismic loading (Hoehler and Eligehausen, 2008a; Mahrenholtz, 2012). The cumulating anchor displacement during crack cycling and therefore the reduction in embedment and residual load capacity depend on the anchor product and show a large scatter (Fig. 4). Various studies on the effect of the loading rate on the pullout resistance of reinforcing bars can be found (Hjorth, 1979; Vos and Reinhardt, 1982; Weathersby, 2003). Anchor behavior under seismic relevant loading rates is less investigated. Tension tests on undercut anchors at high loading rates gave evidence that the pullout capacity is significantly increased if compared to the tests at quasi-static rates (Eibl and Keintzel, 1989a,b). Also more recent studies on mechanical and adhesive anchors subjected to monotonic tension or shear load showed that high loading rates generally positively affect the ultimate load capacity of cast-inplace or post-installed anchors (Fujikake et al., 2003; Hoehler et al., 2011; Hunziker, 1999; Klingner et al., 1998; Rodriguez et al., 2001; Salim et al., 2005; Solomos and Berra, 2006). Earthquake relevant loading rates may increase the load capacity of undercut anchors and headed studs by about 20% compared to quasi-static loading rates (Fig. 5). However, there is not any published data available showing the effect of loading rate on the initial stiffness and displacements at peak load of anchors. Further, crack cycling, cyclic tension, and cyclic shear tests on anchors at high frequencies were not the subject of earlier studies. 4. Qualification according to the German guideline for anchorages in NPP The experimental study carried out to investigate the influence of loading rate and cycling frequency on anchor behavior is based on the German Guideline for Anchorages in NPP which stipulates special test conditions and assessment criteria for anchors subjected to extreme design loads which occur only once during service life. This includes loads from differential pressure and temperature variations, as well as from extreme earthquakes, aircraft impacts and explosions. The German Guideline for Anchorages in NPP distinguishes three fundamental types of tests which are denominated A, B, and C in the following:

1094


Table 1 Compilation of tests, test conditions and assessment criteria according German Guideline for Anchorages in NPP. Test type

No. of tests

Test conditions Crack widtha

Assessment criteria Anchor loadb

Fu,m

– ncyc = 15

≤50%c ≤50%c

≤20%d ≤20%d

NP = NRk (B.1)/ Mc

ncyc = 15

≤50%c

≤20%d

≥0.8Nu,m (B.1)e ≥0.9Nu,m (A.1)e and 0.7Nu,m (B.1)e ≥0.9Nu,m (A.1)e and 0.7Nu,m (B.1)e

– Vmax = ±VRk (w1 )/ Ms

– ncyc = 15

≤40% –

≤15% ≤10%f

– ≥0.9Vu,m (w1 )g,h

Scatter

Displacement after cycling

CV (s(ncyc ))

s(ncyc )

w2 = 1.5 mm w2 = 1.5 mm

– Nmin = 0.0 kN Nmax = NRk (B.1)/ Mc

A.3 Cyclic crack

n≥5

wmin wmax = w2 = 1.5 mm

w1 = 1.0 mm w1 = 1.0 mm

C.2 Cyclic crack

n≥5

C.3 Cyclic shear

n≥5

a b c d e f g h i j

w1 = 1.0 mm wmin wmax = w1 = 1.0 mm w1 = 1.0 mm

Residual capacity CV(Fu )

n≥5 n≥5

C Anchor displacement tests n≥5 C.1 Cyclic tension

Scatter

CV(s (0.5Fu,m )) A Suitability tests A.1 Monotonic tension A.2 Cyclic tension

B Service condition tests B.1 Monotonic tension n≥5 B.2 Cyclic shear n≥5

No. of cycles

Nmin = 0.0 kN Nmax = NRk (B.1)/ Mc NP = NRk (B.1)/ Mc

ncyc = 10

≤40%c

≤3 mmi,j

ncyc = 10

≤40%c

≤3 mmi,j

Vmax = ±VRk (w1 )/ Ms

ncyc = 15

≤40%c

≤3 mmi,j

Crack width may be reduced if detailed analysis is performed. Characteristic strength as 5% quartile; Mc : partial safety factor for concrete failure = 1.7; Ms : partial safety factor for steel failure = 1.25 (typically). If CV > 30%, number of tests has to be increased to n = 10. If CV > 10%, number of tests has to be increased to n = 10. If requirement not fulfilled, the characteristic strength has to be reduced to NRk,p = NRk,p (B.1) × Nu,m /0.8Nu,m (B.1). If requirement not fulfilled, the characteristic shear load capacity shall be reduced by ˛V = 1/(1 + 0.03(CV − 10%)). If requirement not fulfilled, the characteristic strength has to be reduced to VRk,s = VRk,p (w1 ) × Vu,m /0.9Vu,m (w1 ). If no reference test results in w1 are available, VRk and Vu,m may be either calculated or taken from the ETAG approval tests in 0.3 mm cracks. Criterion effective only for connections which are assumed rigid in the design. If requirement not fulfilled, the characteristic strength has to be reduced to VRk,s = VRk,p (w1 ) × Vu,m /0.9Vu,m (w1 ).

Fig. 5. Influence of loading rate on the peak load Nu of various anchor types (Hoehler et al., 2011).


a)

Support Pull-push-bar

Vmin

Fixation Crack Strong floor

Time t

b)

wmax

Nmax Anchor load N

Actuator for anchor loading Transducer for anchor displacement Anchor Concrete member Transducer for crack width

Crack width w

Tie-down

Crack width w

Vmax

Time t

c)

Actuator for concrete member loading c)

Crack inducers

Anchor load N

NP

wmax wmin

Anchor test location Pilot holes Debonding Fig. 6. (a) Setup shear and tension tests, (b) side elevation of test setup and (c) horizontal section of concrete member used for crack cycling tests.

A. Suitability tests to check the proper functioning of the anchor under extreme conditions. B. Tests for determining the characteristic anchor capacities under service conditions. C. Tests for determining the anchor displacements under service condition. Table 1 provides a comprehensive overview of all required tests, the test conditions and the assessment criteria. For load and crack cycling tests, the assessment criteria at peak load are valid for the monotonic pullout tests which follow the cycling to determine the residual capacity. Also limits of the coefficient of variation (CV) belong to the criteria. The aim of the suitability tests (test series A) is to check the anchor behavior under extreme conditions represented by extreme crack widths. Because the influence of the crack width on the anchor behavior under shear load is secondary, only tests with tension loaded anchors are performed. The results of the service condition tests (test series B) are used to evaluate the characteristic resistance (5% quantile of failure loads) under tension and shear

Crack width w

Reaction frame

wmax Anchor load V

a)

b)

1095

Time t Fig. 7. Load and crack time histories of (a) cyclic shear test, (b) cyclic tension test, and (c) crack cycling test.

loading (VRk and NRk ). The anchor displacement behavior under service condition is determined by test series C. In axial direction, the crack cycling tests are crucial since they generally yield larger axial displacements than cyclic tension tests. In transverse direction, cyclic shear load tests determine the critical transversal displacements. If any test criterion is not met, the tests may be repeated at a lower load level which generally increases the residual load capacity and decreases the anchor displacements but the allowable anchor capacity must be reduced correspondingly. It is important to note that all tests are carried out at quasi-static loading rates on anchors installed in cracks. As pointed out above, earthquakes may cause large crack widths and full crack closure. The German Guideline for Anchorages in NPP does not require full crack closure but stipulates increased maximum crack widths to compensate for the missing negative effect. The assessment criteria given in Table 1 were taken to evaluate the test results presented in Section 6: • Displacement during cycling s(ncyc ): The connection may be assumed as rigid in the design only if the displacement measured in test series C under the design load is not larger than 3 mm. This requirement ensures that all displacement sensitive elements connected to a reinforced concrete structure do not move significantly during an earthquake and that the models for the

1096


b) 80

100 80 60 40 20 0 -20 -40

Load V [kN]

Pullout phase Cycling phase

Medium fload = 1 Hz trise = 0.25 sec Medium

-6 -4 -2

0

Residual capacity

70 60 50 40 30 20 10 0

Pullout phase Medium fload = 1 Hz trise = 0.25 sec

Cycling phase

Medium 0

2 4 6 8 10 12 14 16 18 Displacement s [mm]

1

2

3

4 5 6 7 8 9 Displacement s [mm]

d) 30

70 60 50 40 30 20 10 0

Pullout phase Cycling phase

Residual capacity Medium fcrack = 1 Hz trise = 0.25 sec Medium

Displacement s [mm]

c) 80 Load N [kN]

Load N [kN]

Residual capacity

25 20 15

Load approximately constant

10 5

Displacement increasing

0 0

1

2

3

4 5 6 7 8 9 Displacement s [mm]

Medium

Quasi-static fcrack = 0.01 Hz trise = 180 sec

10 11 12

0

400

800 1200 Time t [sec]

1600

10 11 12

80 70 60 50 40 30 20 10 0 2000

Load N [kN]

a) 120

Fig. 8. Example load–displacement curves for (a) cyclic shear test, (b) cyclic tension test, and (c) crack cycling test as well as (d) load and displacement history of crack cycling test.

structural analysis of the elements supported by anchors (e.g. piping) which usually assume a rigid support are sufficiently accurate. • Stiffness at 50% of the peak load k(0.5Fu,m ) and corresponding coefficient of variation CV(k(0.5Fu,m )): most anchorages consists of several anchors forming one anchor group. To avoid insufficient load distribution and potentially overloading of individual anchors, the stiffness of the anchors should not scatter too much. Therefore, the CV of the stiffness at 50% of the mean residual capacity k(0.5Fu,m ) = 0.5Fu,m /s(0.5Fu,m ) is limited to 50% for test series A and 40% for test series B and C. • Residual capacity Fu : The capacity determined following the load or crack cycling has to meet certain requirements if compared to the monotonic load capacity. Furthermore, the CV of the peak loads is limited. These requirements ensures sufficient load bearing capacity also under adverse conditions and they are the basis for the given partial safety factors . 5. Experimental tests The test setups used for the tests presented in this paper basically fell back on equipment used for qualification testing according to the German Guideline for Anchorages in NPP. The test setups specified therein for monotonic and cyclic shear and tension tests (Fig. 6a) are similar to those described in standard anchor qualification guidelines ACI 355.2 and ETAG 001 and are not further explained in this paper. In contrast, the crack cycling test with a controlled upper and lower crack width is not part of the qualification test program stipulated in ACI 355.2 or ETAG 001. The setup suitable for crack cycling tests according to the German Guideline for Anchorages in NPP required upgrading to allow high speed tests. Furthermore, the cracks were actively closed by a compression force equal to 0.15Ac × fc (Ac = cross section of test specimen, fc = compressive strength of concrete specimen) to simulate realistically the conditions during earthquakes (Hoehler and Eligehausen, 2008a). Therefore, a servo-controlled hydraulic system allowing high testing accuracy at high crack cycling frequencies was employed (Fig. 6b).

Prefabricated concrete members (dimension L/W/H: 700 mm/400 mm/200 mm) were connected to a 1000 kN servocontrolled actuator by means of four high strength rods (Fig. 6c). Two thin metal sheets were embedded as crack inducers at both sides of the center of the concrete members to aid together with predrilled pilot holes the crack formation when the specimen was loaded by the actuator in tension. To ease the crack opening, the high strength tie rods were debonded near the metal sheets. Stirrups made of standard reinforcing bars at both ends took up the bond splitting forces as well as the transverse forces generated when the specimen was loaded in compression for crack closure. The servo-control system used the input signal from the transducers measuring the crack width to control the forces applied on the concrete member. A second actuator mounted on a steel support was used for anchor loading. Reference is made to (Mahrenholtz et al., in preparation) for further details on crack cycling tests. The tested concrete anchor was an undercut anchor which is the recommended anchor type for applications in NPP according to the German Guideline for Anchorages in NPP. The product is prequalified for seismic and other extreme loads according to ACI 355.2 and the German Guideline for Anchorages in NPP. An elaborated test program with variable rates and frequencies is required to quantify the effects of high loading rates and cycling frequencies on the anchor performance. Technically relevant, earthquakes induced frequencies range from 1 Hz to 10 Hz (Eibl and Keintzel, 1989a; Hoehler et al., 2011). The load and crack cycling frequencies fload and fcrack as well as the corresponding rise time trise required to reach the peak load during the final pullout are summarized in Table 2. Basically, the tests were conducted according to the test series C stipulated in the German Guideline Table 2 Definition of rates and frequencies. Rate and frequency

fload

fcrack

trise a

Quasi-static Medium High

0.1 Hz 1 Hz 5 Hz

0.01 Hz 1 Hz 5 Hz

180 s 0.25 s 0.025 s

a

Time to reach peak load.


a)

Normalized with reference to value of quasi-static test [-]

V Vu 0.5Vu,m k(0.5Vu,m)

s(ncyc)

Mode

s

Quasi-static

Medium

High

3xS

3xS

3xS

k(0.5Nu,m)

s

s(ncyc) Quasi-static

Medium

1.25 1.00 0.75 0.50 0.25 s(ncyc)

s(ncyc) Normalized with reference to value of quasi-static test [-]

0.5Nu,m

High

Normalized with reference to value of quasi-static test [-]

0.5Nu,m k(0.5Nu,m)

s(ncyc)

Mode

s

Quasi-static

Medium

High

1 x C, 4 x S

4xS

4xS

CV

Vu

Vu


1.50 1.25 1.00 0.75 0.50 0.25 0.00

c) N Nu

k(0.5Vu)

k(0.5Vu,m) CV(k(0.5Vu,m))

1.75

s(ncyc)

4 x C, 1 x S 3 x C, 2 x S 3 x C, 2 x S

Mode


1.50

0.00

b) N Nu

1.75

1097

s(ncyc)

k(0.5Nu)

CV

k(0.5Nu,m) CV(k(0.5Nu,m))

1.75

Nu

Nu


1.50 1.25 1.00 0.75 0.50 0.25 0.00 s(ncyc)

s(ncyc)

k(0.5Nu)

CV

k(0.5Nu,m) CV(k(0.5Nu,m))

Nu

Nu

Fig. 9. Key data of (a) cyclic shear test, (b) cyclic tension test, and (c) crack cycling test.

for Anchorages in NPP (Table 1), however, technical limits required minor adoptions in cycle number and lower target crack width (Table 3). The test procedures are given as schematic load and crack time histories in Fig. 7. 6. Discussion of test results and evaluation of basic parameters The diagrams in Fig. 8a–d depict example load–displacement curves as well as a load and displacement history of the anchors carried out under the scope of the test program (Table 3). Table 3 Test program. Test type

Cyclic shear

Cyclic tension

Cyclic crack

Anchor size Crack width Anchor load

M10 1.0 mm ±VRk / Ms

M12 1.5 mm ±NRk / Mc

M12 0.5–0.8 mm NRk / Mc

No. of cycles

15

15

10

Rate and frequency

Test repeats


3 3 3

5 5 5

5 5 5

Fig. 8a shows a typical load–displacement curve obtained from a cyclic shear test on an undercut anchor M10 at medium loading rate. The anchor was installed in a hole intercepted by a wide crack of wmax = 1.0 mm and was first subjected to fifteen cycles of alternating shear loads between ±VRk / Ms (=36.0 kN in this case) at a loading frequency of 1 Hz. The hysteresis loops clearly display the large slip, a signature of shear cycling tests, which is caused by the clearances between fixture, anchor, and anchor hole. Once the fifteen cycles were completed, the anchor was subjected to monotonically increasing load until failure due to steel rapture. Fig. 8b shows an example load–displacement curve for a tension test conducted on an undercut anchor M12 at medium loading rate. The crack was opened to a width of wmax = 1.5 mm and the anchor was then subjected to pulsating tension load varying between NRk / Mc (=23.5 kN in this case) and almost 0.0 kN for fifteen cycles at 1 Hz. The hysteresis loops are closely spaced demonstrating the robust behavior of undercut anchors due to the mechanical interlock. On completion of load cycling, the load was monotonically increased until failure which was caused in this case by steel rapture. Fig. 8c represents a typical load–displacement curve obtained from a crack cycling test on an undercut anchor M12 at a medium cycling frequency of 1 Hz. Fig. 8d shows the plots of anchor load and anchor displacement versus time for a quasi-static crack cycling frequency of 0.01 Hz. The anchors were first loaded by a tension load

1098


of N = NRk / Mc (=23.5 kN in this case) in a crack opened to wmin (=0.5 mm). Following, the crack was opened to wmax (=0.8 mm) and again closed to wmin , thus completing one cycle. As the crack opened, the anchor slips out since the stiffness of the system reduces with increasing crack width. The slip was partly reversed when the crack closed, however, the anchor was unable to return to its original position and with each cycle the cumulated anchor displacement increased. After the end of ten cycles of crack opening and closing, the crack was opened to wmax = 0.8 mm and the anchor was monotonically loaded to failure which occurred by steel rapture for the shown case. With reference to the discussion in Section 4, the basic parameters are extracted from the load–displacement curves of the load cycling and crack cycling tests, i.e. the displacement after cycling s(ncyc ) (before releasing the load which relaxes the anchor displacement to s*) as well as the stiffness at 50% of the peak load k(0.5Fu,m ) (i.e. 0.5Fu,m /(s(0.5Fu,m ) − s*)), its corresponding CV(k(0.5Fu,m )) and the peak load Fu (i.e. the residual capacity) of the final pullout after cycling (Fig. 9). For each type of test, the bar diagrams show the mean results of tests (3 shear tests or 5 tension tests, respectively) conducted at quasi-static, medium and high loading rates. To quantify the relative influence of the increasing rates, the values are normalized with reference to the mean of the tests performed at quasi-static loading rates. Fig. 9a shows the summary of results obtained from the cyclic shear tests. Steel failure (mode S) was observed for all anchors. The increase in peak load Vu for increasing loading rates confirms the results of earlier studies. In contrast, the anchor behavior in respect to the displacement after cyclic loading s(ncyc ) and stiffness k(0.5Vu,m ) does not follow a clear trend. However, the scatter of the stiffness CV(k(0.5Vu,m )) clearly increases with increasing loading rates. Fig. 9b summarizes the test data obtained from the cyclic tension tests. The anchors experienced concrete and steel failure (mode C and S). The conclusions are basically the same as for the cyclic shear tests. In particular, higher loading rates neither yielded a smaller displacement after load cycling s(ncyc ) nor a greater stiffness k(0.5Nu,m ). The peak load Nu appears to be unaffected by the loading rate. However, the scatter of the stiffness CV(k(0.5Nu,m )) again increases for increasing loading rates. Fig. 9c compares the key test data determined for the crack cycling tests which ended with concrete and steel failure (mode C and S). For these test series, the stiffness k(0.5Nu,m ) and scatter of the stiffness CV(k(0.5Nu,m )) increase with increasing loading rates. However, such clear trend cannot be identified for the displacement s(ncyc ). It seems that the medium crack cycling frequency led to a significant increase of the anchor displacement after cycling whereas the high frequency resulted in a less pronounced increase. The ultimate load Nu is not significantly influenced by the loading rate. High scatter in the results of post-installed anchors tested in concrete is inevitable. It is attributable to the local concrete properties in the region of the load transfer zone which vary significantly from test to test. They may affect the results to such an extent that the influence of other factors, e.g. loading rate, becomes less significant. The impact of the scatter on the peak loads is also apparent in Figs. 3 and 4. The scatter in displacement data is generally even higher than the scatter in corresponding load data (Mahrenholtz et al., 2011). The scatter in the load and displacement data is visualized best by plotting load–displacement curves of all repeats in one graph. Fig. 10 shows the complete test series for the medium frequency and loading rate as an example. As shown in Fig. 9, the scatter of the displacement behavior increases with increasing frequencies and loading rates. Because of the relatively high scatter of the anchor displacements, the influence of loading rate on the displacements after

Fig. 10. Load–displacement curves at medium frequency and loading rate of (a) cyclic shear test, (b) cyclic tension test, and (c) crack cycling test.

cycling s(ncyc ) and the stiffness at 50% of the peak load k(0.5Fu,m ) cannot be judged with sufficient accuracy on the basis of only three or five test repeats. For an increased number of test repeats, the test results may gain more significance and possibly show a clear correlation between loading rate and displacement data. However, for the currently available test data, a beneficial effect of a high loading rate and frequency cannot be demonstrated for the anchor displacements accumulated during cycling or for the initial stiffness of the load–displacement curve in the residual load test. In contrast, the scatter of the peak loads is relatively small. Therefore, it may be concluded that the residual capacity of the cyclic shear tests (failure caused by steel rapture) increases with increasing loading rates. This agrees with earlier findings. However, the residual capacity after cyclic tension and crack cycling tests (failure caused by steel rapture and concrete breakout) was not much influenced by the loading ratio. 7. Summary and conclusions Post-installed anchors used for safety relevant applications have to be qualified to show their suitability for the assigned application. Particularly in Nuclear Power Plants (NPP), the anchors need to


prove their capability to cope with extreme conditions. The German Guideline for Anchorages in NPP provides guidance for the qualification of anchors used in NPP. The herein required test conditions replicate the extreme demands of dynamic actions to be anticipated for the high safety standards in force. The specified qualification tests are conducted at quasi-static loading rates. In reality, higher loading rates are expected. Therefore, tests on undercut anchors were carried out at dynamic loading rates to check their influence on anchor behavior. However, a significant and consistent decrease of anchor displacement after cycling or increase of initial stiffness for increased loading rates was not observed. Most important, the displacements during load and crack cycling, which are decisive for the determination of characteristic displacements according to the German Guideline for Anchorages in NPP, were not significantly reduced for increased frequencies. The scatter of the displacements increased substantially with increasing frequencies, impeding a reliable quantification of the effects on the basis of a limited number of tests. Based on the test results reported in this paper it is concluded that the design displacements derived from quasi-static qualification tests should not be reduced for seismic applications and a relaxation of the assessment criteria of the concrete anchor qualification guidelines cannot be justified. Acknowledgements The authors would like to thank the RWE Corporation, Germany and the Bhabha Atomic Research Centre (BARC), India for their financial support of this research. The staff of the BARC laboratory and the MPA testing laboratory in Stuttgart, Germany is thanked for the assistance during testing. The most challenging high frequency crack cycling tests were conducted together with Dr. Akanshu Sharma, former scientific officer of BARC and postdoctoral researcher at the University of Stuttgart, Germany. His long standing commitment for the tests as well as his input and consulting for the preparation of this paper are very much appreciated. Opinions, conclusions, and recommendations expressed in this paper are those of the authors and do not necessarily reflect those of the sponsors. References ACI 318, 2002. Building code requirements for structural concrete (ACI 318-02) and commentary (ACI 318R-02). American Concrete Institute, Farmington Hills, MI. ACI 318, 2011. Building code requirements for structural concrete (ACI 318-11) and commentary (ACI 318R-11). American Concrete Institute, Farmington Hills, MI. ACI 349, 1980. Code Requirements for nuclear safety related concrete structures (ACI 349-80) and commentary (ACI 349R-80). American Concrete Institute, Detroit, MI. ACI 349, 2006. Code Requirements for nuclear safety-related concrete structures (ACI 349-06) and commentary (ACI 349R-06). American Concrete Institute, Detroit, MI. ACI 355.2, 2007. Qualification of post-installed mechanical anchors in concrete (ACI 355.2-07) and commentary. American Concrete Institute (ACI), Farmington Hills, MI. Bergmeister, K., 1988. Stochastik in der Befestigungstechnik mit realistischen Einflussgrößen (Stochastic in fastening technology under realistic variables). Universität Innsbruck (Dissertation). CAN3-N289.4-M86, 1986. Testing procedures for seismic qualification of CANDU Nuclear Power Plants. Canadian Standards Association (CSA) (reaffirmed 2003). Cannon, R., 1981. Expansion anchor performance in cracked concrete. ACI Journal 78 (6), 471–479. CEN/TS 1992-4, 2009. Design of fastenings for use in concrete – Part 4-1–4-5. European Committee for Standardization (CEN), Brussels (Technical Specification). DIBt KKW Leitfaden, 1998. Verwendung von Dübeln in Kernkraftwerken und kerntechnischen Anlagen, Leitfaden zur Beurteilung von Dübelbefestigungen bei der Erteilung von Zustimmungen im Einzelfall nach den Landesbauordnung der Bundesländer (Use of anchors in Nuclear Power Plants and nuclear technology installations, guideline for evaluating fastenings for granting permission in individual cases according to the state structure regulations of the federal states of Germany). Deutsches Institut für Bautechnik (DIBt), Berlin (in German).

1099

Eibl, J., Keintzel, E., 1989a. Zur Beanspruchung von Befestigungsmitteln bei dynamischen Lasten (Loading of fastenings under dynamic loads). IMB, Universität Karlsruhe (TH) (in German). Eibl, J., Keintzel, E., 1989b. Verhalten von Dübeln unter hoher Stoß- und Wechselbeanspruchung (Behaviour of anchors under high speed impact and reversed cyclic loads). IMB, Universität Karlsruhe (TH) (in German). Eligehausen, R., Balogh, T., 1995. Behavior of fasteners loaded in tension in cracked reinforced concrete. ACI Structural Journal 92 (May–June (3)), 365–379. Eligehausen, R., Lotze, D., Sawade, G., 1986. Untersuchungen zur Frage der Wahrscheinlichkeit, mit der Dübel in Rissen liegen (Investigation of the probability that anchors are located in cracks. Report No. 1/20-86/17). Institut für Werkstoffe im Bauwesen, Universität Stuttgart (not published; in German). Eligehausen, R., Mallée, R., Silva, J., 2006. Anchorage in concrete construction. Ernst & Sohn, Berlin. EN 1992, 2011. Eurocode 2: design of concrete structures. European Committee for Standardization (CEN), EN 1992:2011. EN 1998, 2006. Eurocode 8: design of structures for earthquake resistance. European Committee for Standardization (CEN), EN 1998:2006. ETAG 001, 2013. Guideline for European technical approval of metal anchors for use in concrete, Parts 1–6. European Organization of Technical Approvals (EOTA), Brussels. Fuchs, W., Eligehausen, R., 1989. Tragsverhalten von Befestigungsmitteln im gerissenen Beton bei Querzugbeanspruchung (Load bearing behaviour of fixing elements under shear loading in cracked concrete. Report No. 1/41-89/15). Institute für Werkstoff im Bauwesen, Universität Stuttgart (not published in German). Fujikake, K., Nakayama, J., Sato, H., Mindess, S., Ishibashi, T., 2003. Chemically bonded anchors subjected to rapid pullout loading. ACI Materials Journal 100 (May–June (3)), 246–252. Furche, J., 1987. Versuchseinrichtung zur Prüfung von in Rissen verankerten Dübeln und erste Versuche an Parallelrisskörpern ([[nl]Test setup for the testing of fasteners in cracks and initial tests in parallel cracks). Institut für Werkstoffe im Bauwesen, Universität Stuttgart (not published in German). Furche, J., 1988. Versuche an Kopfbolzen mit unterschiedlichen Kopfformen bei Verankerungen in sich öffnenden und schließenden Rissen: Part 2 (Tests with headed bolts with various head shapes in opening and closing cracks: Part 2. Report No. 9/5-88/10). Institut für Werkstoffe im Bauwesen, Universität Stuttgart (in German). Guillet, T., 2011. Behavior of metal anchors under combined tension and shear cycling loads. ACI Structural Journal 108 (May–June (3)), 315–323. Hjorth, C., 1979. Ein Beitrag zur Frage der Festigkeiten und des Verbundverhaltens von Stahl und Beton bei hohen Dehngeschwindigkeiten (A contribution to the bond problem of steel and concrete under high strain rates). Technische Universität Braunschweig (Dissertation in German). Hoehler, M., Eligehausen, R., 2008a. Behavior and testing of anchors in simulated seismic cracks. ACI Structural Journal 105 (May–June (3)), 348–357. Hoehler, M., Eligehausen, R., 2008b. Behavior of anchors in cracked concrete under tension cycling at near-ultimate loads. ACI Structural Journal 105 (September–October (5)), 601–608. Hoehler, M., Mahrenholtz, P., Eligehausen, R., 2011. Behavior of anchors in concrete at seismic-relevant loading rates. ACI Structural Journal 108 (March–April (2)), 238–247. Hunziker, P., 1999. Shocktesting of concrete anchor bolts for shock resistant applications in protective structures. In: Proceedings of the 3rd International Conference on Shock & Impact Loads on Structures, Singapore. IEB 79-02, 1979. Pipe support base plate designs using concrete expansion anchor bolts. Inspection and Enforcement Bulletin, USNRC, Revision 2, November 1979. Kim, S.-Y., Yu, C.-S., Yoon, Y.-S., 2004. Sleeve-type expansion anchor behavior in cracked and uncracked concrete. Nuclear Engineering and Design 228, 273– 281. Klingner, R., Gross, J., Lotze, D., Park, H.-G., Rodriguez, M., Zhang, Y.-G., 1998. Anchor bolt behavior and strength during earthquakes. U.S. Nuclear Regulatory Commission, NUREG/CR-5434. Klingner, R., Mendonca, J., Malik, J., 1982. Effect of reinforcing details on the shear resistance of anchor bolts under reversed cyclic loading. ACI Journal Proceedings 79 (January (1)), 471–479. Lotze, D., 1987. Untersuchungen zur Frage der Wahrscheinlichkeit, mit der Dübel in Rissen liegen – Einfluß der Querbewehrung (Investigations on the Probability of Fasteners being Located in Cracks – Influence of Transverse Reinforcement. Bericht Nr. 1/24-87/6). Institut für Werkstoffe im Bauwesen, Universität Stuttgart (not published in German). Lotze, D., Faoro, M., 1988. Rißbreitenentwicklung und Dübelverschiebung bei veränderlicher Bauteilbelastung (Width development and fastener displacement under varying building component loads. Report No. 1/28-88/3). Institut für Werkstoffe im Bauwesen, Universität Stuttgart (in German). Mahrenholtz, C., Silva, J., Eligehausen, R., Hofmann, J., 2013. Testing anchors in cyclic cracks: guidance for testing laboratories on how to generate cracks and cycle crack widths. Concrete International (in preparation). Mahrenholtz, P., 2012. Experimental performance and recommendations for qualification of post-installed anchors for seismic applications. University of Stuttgart (Dissertation). Mahrenholtz, P., Eligehausen, R., Hofmann, J., 2011. Ductility of post-installed anchors. In: Proceedings of the Conference on Advances in Structural Engineering and Mechanics (ASEM11+), Seoul, pp. 1660–1675. Malhotra, P., 2002. Cyclic-demand spectrum. Earthquake Engineering and Structural Dynamics 31 (7), 1441–1457.

1100


Malvar, L., Ross, C., 1998. Review of strain rate effects for concrete in tension. ACI Materials Journal 95 (November–December (6)), 735–739. Nuti, 2008. Consideration on ETAG 001 (Amendment) seismic actions ASPC2. In: EOTA Meeting, 15 December 2009, Rome (not published). Rehm, G., Lehmann, R., 1982. Untersuchungen mit Metallspreizdübeln in der gerissenen Zugzone von Stahlbetonbauteilen (Investigations with metal expansion anchors in cracked reinforced concrete). Fraunhofer Society Research Report V. 1015 (in German). Rieder, A., 2009. Seismic response of post-installed anchors. Institut für konstruktiven Ingenieurbau der Universität für Bodenkultur, Wien (Dissertation). Rodriguez, M., Lotze, D., Gross, J., Zhang, Y.-G., Klingner, R., Graves, H., 2001. Dynamic behavior of tensile anchors to concrete. ACI Structural Journal 98 (July–August), 511–524. Salim, H., Dinan, R., Shull, J., Townsend, P., 2005. Shock load capacity of concrete expansion anchoring systems in uncracked concrete. ASCE Journal of Structural Engineering 131 (August (8)), 1206–1215. SEAOSC, 1997, April. Standard method of cyclic load test for anchors in concrete or grouted masonry. Structural Engineers Association of Southern California (SEAOSC), Whittier, CA. Seghezzi, H., 1985. Wechselbeziehungen zwischen Prüftechnik und Entwicklung von Befestigungselementen (Interactions between testing technology and development of fastening elements). Stuttgart, Presentation 15.01.85. Sharma, A., Oˇzbolt, J., Reddy, G., Vaze, K., Ghosh, A., Kushwawa, H., 2010. Effect of loading rate on load-deflection behavior and failure mode of plain concrete slabs – 3D finite element analysis approach. International Journal of Earth Sciences and Engineering 3 (4), 812–822. Solomos, G., Berra, M., 2006. Testing of anchorages in concrete under dynamic tensile loading. RILEM Materials and Structures 39 (7), 695–706.

Tang, J., Deans, J.,1983. Test criteria and method for seismic qualification of concrete expansion anchors. In: Proceedings of the 4th Canadian Conference on Earthquake Engineering. University of British Columbia, Vancouver, pp. 58–69. Usami, S., Abe, U., Matsuzaki, Y., 1981. Experimental study on the strength of bonded anchors under alternate shear load and combined load. In: Proceedings of the Annual Meeting of the Kantou Branch of the Architectural Institute of Japan. Vintzeleou, E., Eligehausen, R., 1991. Behavior of fasteners under monotonic or cyclic shear displacements. In: Anchors in Concrete – Design and Behavior. ACI Special Publication SP 130, pp. 181–204. Vos, E., Reinhardt, H.-W., 1982. Influence of loading rate on bond behaviour of reinforcing steel and prestressing strands. Matériaux et Construction (RILEM), No. 85. Weathersby, J., 2003. Investigation of bond slip between concrete and steel reinforcement under dynamic loading conditions. Dissertation at The Department of Civil and Environmental Engineering, Louisiana State University. Weigler, H., Lieberum, K.H., 1984. Belastungsprüfungen an Liebig-Einspannankern ultra-plus M16, verankert in kreuzartig gerissenen Betonprobkörpern bei stoßartiger und statischer Belastung (Load tests on Liebig restraining anchor Ultra-plus M16, anchored in concrete specimens with intersecting cracks under impact and static loads. Report). Institute für Massivbau, TH Darmstadt (in German). Wesche, K., Krause, K., 1972. Der Einfluss der Belastungsgeschwindigkeit auf Druckfestigkeit und Elastizitätsmodul von Beton (The influence of the loading rate on the compressive strength and elasticity modulus of concrete). Materialprüfung 14 (7), 212–218 (in German). Zielinski, A., 1982. Fracture of concrete and mortar under uniaxial impact tensile loading. Delft University of Technology (Dissertation).