Shear breakout capacity of various fastening systems in concrete elements

The results of a test program carried out to investigate the difference in the concrete breakout capacity of various fastening systems installed parallel to the edge of a concrete member, and loaded in shear towards the edge of the concrete member, are presented and analysed. According to the current design recommendations, the calculated concrete breakout capacity of anchor channels yields to a lower failure capacity compared to headed anchors. Welded anchors (headed anchors welded to a steel plate) and bonded anchors are also tested. Based on the test results, various fastening systems in concrete elements are compared.


Introduction
Significant developments have been made over the past decades in the field of fastening technology.The main purpose of fasteners is to connect different types of structural and non-structural elements, i.e. to enable connection between concrete members and the steel structure.There are different types of fasteners, such as headed anchors, chemical and mechanical post-installed anchors, and anchor channels.Anchors can be used as single fasteners or as a group of anchors.Anchor channels consist of at least two anchors.The main advantage of anchor channels is the flexibility of installation, i.e. the structure can be fixed at different positions.A typical application in which anchor channels are installed close to the edges of a concrete member is the fastening of curtain wall facades (see Figure 1).Due to the small edge distance, the steel capacity of the fastening system cannot fully be utilized and so the controlling failure mode is the concrete edge breakout.Currently, according to the design standards (DIN EN 1992-4 [2] and AC232 [3]), the calculated concrete breakout capacity of anchor channels yields to lower resistance compared to headed anchors for the same edge distance and member thickness.Main reasons for the reduced capacity of anchor channels are listed as follows: -In design, an uneven load distribution to the anchors is assumed for anchor channels, whereas for headed anchors the shear load is evenly distributed to the anchors.
-For anchor channels, the critical spacing (s cr,V ), the critical member thickness (h cr,V ) and the critical corner distance (c cr,V ) to ensure full breakout body, are assumed to be much larger compared to headed anchors.In the design of headed anchors s cr,V = 3c 1 , h cr,V = 1.5c 1 and c cr,V = 1.5c 1 , while the following is valid for anchor channels s cr,V = 4c 1 + 2b ch , h cr,V = 2c 1 + 2h ch and c cr,V = 2c 1 + b ch (c 1 = edge distance, h ch = channel height, b ch = channel width).
-The basic equation for calculating the concrete edge breakout resistance of one anchor is different for anchor channels and headed anchors.Experimental investigations in uncracked concrete were performed with headed anchors, welded anchors, bonded anchors, and anchor channels, in order to understand the difference in concrete edge breakout capacity between fastening systems consisting of various types of anchors.Hence, in all tests, the edge distance c, the number of anchors in a group n, and the anchor spacing s of the fastening system, as well as the concrete strength, were kept constant.For the tested parameters (n = 3, c = 100 mm, s = 150 mm), according to DIN EN 1992-4 [2], the fastening with anchor channels leads to 30% higher utilization in design compared to the fastening with headed anchors for the same acting shear load.On the other hand, according to CEN/TS 1992-4 [4], headed anchors and anchor channels exhibit a comparable utilization in design.Therefore, the main motivation for performing the experiments presented in this paper was to improve the current design models.A comprehensive review of current design models is given in Eligehausen et al. [5].It is important to note that a direct comparison of calculated resistances is not possible since the resistance of the entire group is calculated for headed anchors, whereas the most unfavourable anchor is verified for anchor channels.More details can be found in Grosser et al. [6].Welded anchors (headed anchors welded to a steel plate embedded flush to the concrete surface) and bonded anchors were also tested for comparison purposes.Test specimens and test setup are described in detail in the first part of the paper.Experimental results are summarized in the second part.Finally, conclusions relating to current design are drawn based on test results.

Experimental tests
A total of 20 tests were performed at the Faculty of Civil Engineering in Rijeka, Croatia.The main aim of the tests was to investigate the difference in concrete shear breakout capacity of different fastening systems arranged close to concrete edges.Hence, to make the results comparable, all tests were performed in concrete slabs of the same mix composition, age, and concrete strength.In addition, the same geometrical parameters, such as edge distance, number of anchors, and anchor spacing, were used.

Test specimens
Four different fastening systems, headed anchors (Figure 2.a), welded anchors (Figure 2.b), anchor channels (Figure 2.c), and bonded anchors (Figure 2.d), were tested.The first three types of fasteners (headed anchors, welded anchors, and anchor channels) were cast in place, while bonded anchors were cut to the length of 160 mm from M16 threaded rods and postinstalled in the hardened concrete slab with the adhesive Hilti HIT-HY 200-A (Figure 2.d).More details can be found in [7].A total of five concrete slabs were cast: four measuring 1600 mm × 1600 mm × 300 mm and one measuring 1600 mm × 1600 mm × 200 mm.The specimens were cast in a wooden formwork.All tested slabs were reinforced with the wire-mesh Q 131 embedded near the top and bottom of concrete slabs for handling purposes (see Figure 3).Due to the wire position (approximately 200 mm away from the edges of the slab), the reinforcement did not affect the outcome of test results.All slabs were cast horizontally and compacted using a vibrator.The concrete specimens were stored in the laboratory of the Faculty of Civil Engineering, Rijeka, and wrapped with plastic sheets according to HRN EN 12390-2 [8] for 28 days.Afterwards, the plastic sheets were removed and concrete slabs were stored at an ambient temperature until the day of the testing.
Detailed installation parameters for each fastening system in concrete are summarized in Table 1.Thus, it may be seen that four types of fastening systems were tested in 300 mm thick concrete slabs, while anchor channels were also tested in a 200 mm thick concrete slab.A schematic view of test specimens is given in Figure 4 for a group of headed anchors and for anchor channels.In all tests, the fastening systems were arranged parallel to the edge of the concrete slabs and subjected to shear load acting perpendicular to the edge.Each type of fastening system contained 3 anchors that are horizontally spaced at 150 mm intervals.All fastening systems were installed with the edge distance of 100 mm.Headed  anchors, welded anchors and bonded anchors were installed with an embedment depth of 102 mm, while the embedment depth of anchor channels amounted to 106 mm.

Material properties
The normal weight low strength concrete (strength class C20/25) was used for all specimens.The concrete was produced according to requirements specified in

Shear breakout capacity of various fastening systems in concrete elements
Annex A [11].An average cube compressive strength amounted to 33.61 MPa (results ranged from 31.28 MPa to 36.97 MPa).The cube compressive strength was converted into the strength of cylinders using relation given in ETAG 001, Annex A, Eq. 2.1a [11].Fifteen specimens in total were tested.All results are summarized in Table 2.

Test setup and instrumentation
The tests were carried out in the laboratory of the Faculty of Civil Engineering in Rijeka on a strong floor equipped with a ZWICK ROELL hydraulic actuator, type LH 0250-100, with a 250 kN load cell.The photo of the typical test setup and measuring instrumentation is given in Figure 5.To ensure a fixed position of the concrete slab during testing, a special steel frame was designed and manufactured.The steel frame was placed in front of the hydraulic actuator, mounted to a strong floor with steel rods (M20, 8.8 CS) and supported at the front side to avoid movement during testing.The concrete slabs were placed on top of the steel frame and mounted as shown in Figure 5, to avoid upward displacement during testing.A servo hydraulic actuator was used to apply shear load by means of a 30 mm thick loading plate, which was connected to the actuator with a threaded M20 rod.Depending on the fastening system tested, the loading plate was connected directly to the anchors (Figure 6.a) or channel bolts (Figure 6.b), or connected with M20 adapter bolts in case of welded anchors (Figure 6.c).A torque moment of 60 Nm was applied to each fastening system.To reduce friction, a Teflon sheet 2 mm in thickness was placed between concrete and the loading plate.The horizontal displacement in the direction of shear load was measured using displacement transducers type LD 320-50 OMEGA.Two LVDTs were glued to the concrete surface behind the outermost anchors (see Figure 5).The anchor displacement was determined by averaging two LVDTs measurements.The anchor shear load was measured using the load cell type BPS-TL0250.10.00 (max load 250 kN) (see Figure 5).
The tests were performed at room temperature.The age of concrete at the time of testing was approximately 2 months.For all tests, the spacing of concrete support blocks was defined to ensure sufficient distance so as to enable complete

Load-displacement curves
The measured ultimate loads for each test series are summarized in Table 3, and the measured load-displacement curves for each test series are plotted in Figure 7.Note that slip can occur in this type of testing, as can be seen in the measured load-displacement curves.It can lead to different peak displacements and curve shapes (compare CS3-1 and CS5-3 in Figure 7).The reasons can vary, e.g. it can be due to slip of the loading plate, activation of the anchors and anchor channels that can be nonsymmetric, local failure and crushing of concrete, and frictional effects.Test results show that the ultimate strength of headed anchors and anchor channels is approximately the same in the 300 mm thick concrete slabs.Compared to headed anchors (CS1) and anchor channels (CS3), the tests with welded anchors (CS2) exhibit approximately 10% higher ultimate strength results.Furthermore, it is interesting to note that in the case of bonded anchors (CS5), the measured ultimate strength is also by about 10% higher compared to headed anchors (CS1).Generally, the failure mode is of brittle type and is due to the failure of concrete in tension.In terms of ductility, it can be observed that headed anchors (CS1) and bonded anchors (CS5) exhibit relatively brittle post-peak response whereas the post-peak

Comparison of test results
To allow a direct comparison between the fastening systems, all test results were normalized to a concrete compressive strength measured on cylinder f c = 25 MPa (Figure 9.a).Therefore, in order to account for the different concrete strengths, the peak loads were scaled with the factor (25/f cm ) 0.5 , where f cm is the mean cylinder strength of concrete, as given in Table 2.This kind of scaling is due to the fact that the breakout resistance in design is taken as the square root function of the concrete compressive strength.The related failure loads (the ratio of ultimate strength response of welded anchors (CS2) and especially that of anchor channels (CS3, CS4) is more ductile.Note that, from the fracture mechanics point of view, the concrete cone failure of fasteners belongs to the category of negative geometry (the stress intensity factor decreases with an increase in crack length) where concrete fracture energy is more relevant for the resistance than the tensile strength of concrete.Compared to the positive geometry (e.g.threepoint bending), their response is more ductile.However, compared to the failure of reinforced concrete, the fasteners are generally less ductile.

Failure patterns
The concrete edge failure was observed in all tests.Typical breakout patterns for each type of fastening system are shown in Figure 8.As expected, the breakout pattern is similar in all tests.However, it can be observed that the crack starts from the steel plate or channel profile in case of welded anchors (CS2    1).Therefore, the welded anchors can be considered as an anchor channel with infinite rigidity.Current design standards (DIN EN 1992-4 [2] and AC232 [3]) take into account the influence of member thickness in case of anchor channels through reduction factor ψ ch,h,V .In case of the tested anchor channels in a 200 mm thick concrete slab with 100 mm edge distance, Eq. ( 1) leads to the approximately 15% reduction of the calculated concrete breakout strength.
gdje je h cr,v = 2c 1 + 2h ch (1) According to DIN EN 1992-4 [2] there should be no difference between headed anchors and bonded anchors.However, tests with bonded anchors show a 10% higher ultimate strength in shear compared to headed anchors.

Conclusions
The main aim of the research presented in the paper was to investigate the difference in concrete breakout strength of anchor channels and headed anchors in uncracked concrete.
According to DIN EN 1992-4 [2], the design of anchor channels for the concrete edge failure mode leads, for the investigated parameters, to 30% higher utilization compared to the design of headed anchors, whereas according to CEN/TS 1992-4 [4] headed anchors and anchor channels lead to comparable utilization in design.Experimental test results show that the shear breakout strength of anchor channels is comparable to headed anchors for the same parameters.Planning engineers are now faced with the question of how a cost-efficient solution for anchor channels can be obtained with the design according to DIN EN 1992-4 [2] and AC232 [3].The discrepancy between headed anchors and anchor channels needs to be corrected in the current design codes.However, further experimental and numerical studies are required in this respect.

Figure 1 .
Figure 1.Application of anchor channel in facade construction[1]

Figure 2 .
Figure 2. Fastening systems subjected to testing: a) headed anchors in CS1; b) welded anchors in CS2; c) anchor channels in CS3 and CS4; d) bonded anchors in CS5

Figure 6 .Figure 5 .
Figure 6.Connection between fastening system and loading plate for: a) headed anchors; b) anchor channels; c) welded anchors

Figure 7 .
Figure 7. Load-displacement curves of fastening systems: a) headed anchors in CS1 (h = 300 mm); b) welded anchors in CS2 (h = 300 mm); c) anchor channels in CS3 (h = 300 mm); d) anchor channels in CS4 (h = 200 mm); e) bonded anchors in CS5 (h = 300 mm) ) and anchor channels (CS3 and CS4), which leads to a smaller breakout body due to the smaller net edge distance (c` = c -b ch /2).During the testing, it was observed that in case of welded anchors further concrete cracking appears from the back edge of the steel plate (red lines in Figure8.b)soon after initial concrete cone crack formation (black lines in Figure8.b).A full concrete breakout cone developed for all fastening systems tested in 300 mm thick concrete slabs.In case of anchor channels tested in a 200 mm thick concrete slab, the concrete breakout cone is truncated by the lower edge of the concrete slab, which explains reduced capacity observed in test series CS4.

Test specimen (top and side view): headed anchors in CS1 (left); anchor channels in CS3 (right), all dimensions in mm
[8]]EN 2016- 1:2006 and HRN 1128:2007.A single type of concrete mixture was used for all concrete slabs.The concrete consistency S2 was applied, and the maximum aggregate size of 16 mm was used.The concrete compressive strength was determined in accordance with HRN EN 12504-1:2009[9]and HRN EN 12390-3:2009[10]at the Faculty of Civil Engineering, Rijeka.Three cubes 150 mm in side length were made during casting of each slab.The concrete cubes were cured in a water tank for 28 days in accordance with HRN EN 12390-2[8], and were then cured in air until testing.The compressive strength measured at the time of testing ranged between 31.22 MPa and 44.78 MPa (average value: 38.33 MPa).However, to achieve the best representation of concrete strength, strength test specimens should be cured under the conditions identical to those applied for test slabs.Therefore, 3 concrete cylinders measuring 100 mm in diameter and 100 mm in height were taken from each concrete slab after completion of the respective test.The compressive strength measured on concrete specimens was converted into the cube strength using Eq.2.1e of ETAG 001, Figure 4.

Table 2 . Concrete compressive strength measured at day of testing on cored cylinders
Natalija Bede, Philipp Grosser, Joško Ožbolt development of failure cone.Tests were repeated four times per test series.Shear load was applied by controlling hydraulic actuator displacement at the constant displacement rate of 0.02 mm/s, and so the peak load was reached in approximately 3 to 5 minutes (quasi-static tests).During each test, anchor loads and displacement measurements were recorded with the sampling rate of 100 Hz and collected via a data acquisition system (National Instruments).

Table 3 . Summary of test results
Natalija Bede, Philipp Grosser, Joško Ožbolt V u,test,25 to mean ultimate strength calculated from the test series CS1 V u,m,test,25 (CS1) ) are shown in Figure9.b. Anchor channels tested in a 300 mm thick concrete slab (CS3) show a slightly lower ultimate strength compared to headed anchors (CS1).This can be attributed to the smaller anchor diameter for anchor channels.