Effect of boundary conditions

In specimens where secondary concrete failure was of concern bond behaviour was mainly influenced by the radial component of the bond stress triggered by the mechanical interlocking between bar deformations and adhesive. For plain or moderated deformations, bond is mainly influenced by adhesion or by the adhesive shear strength. In case of higher intensity deformations bond is controlled by mechanical interlocking between deformations and adhesive.

Bond characteristics of NSM reinforcements based on advanced test method PhD Thesis by Zsombor K. SZABÓ, supervisor György L. BALÁZS

For G-8-RB reinforcement slip recorded at tension-tension specimens was double as high as measured at the L-shaped specimen. This was expected and it is explained by the larger bond length 300 mm in comparison to 175 mm. The maximal bond stress level was in the range of 12.5 to 17.5 N/mm² for the L-shaped specimen. For the measurement at tension-tension specimens similar values are observed but with higher values over 20 N/mm² at the first segment of the bond length and lower values close to the unbounded length. The high value at the first segment was measured also for the other tested reinforcements. The position first strain gauge was mounted very close to the unbounded zone at 10 mm distance. At the L-shaped specimen this distance was increased to 30 mm.

In the last segment of the bond length relatively low bond strength was measured for the tension-tension specimens at ultimate load.

In this segment failure and bond capacity was influenced by the internal stress developed at the steel bar and by the propagation of the cracks. Stresses were at the steel bar where higher at the specimen ends and decreased towards the middle of the specimen. This was the opposite as for the FRP reinforcement. They were loaded on opposite ends.

In overall the stain measurement was not consistent for the measurements with the tensile-tensile specimen although the strain gauge preparation was similar as for the other specimen. It is believed that the progressive cracking of the concrete disturbed the measurements. Similar disturbance of the strain measurement was observed in tests where the effect of the edge distance was studied.

The measured ultimate load was the highest for the L-shaped specimen. Nevertheless tensile failure of the reinforcement was not observed. Measurement of the transverse deformations showed considerably high values for the tensile-tensile specimens over 1 mm in comparison to the maximum of 0.3 mm observed for the L-shaped specimen.

Failure of the reinforcement G-8-RB was characterised by a longitudinal cracking of the epoxy. A characteristic concrete failure was observed along inclined planes starting from the middle plane of the reinforcement. The failure is shown in Figure 51. The shearing-off the GFRP bars surface deformations observed at the first stage of the bond length is shown in Figure 43.This failure was caused by the large deformations of the GFRP bars. In comparison to the L-shaped specimen failure at the tension-tension specimen was characterised by rather intense concrete cracking especially at the loaded end. Intensive cracking of the concrete was observed also on the anchorage bond length at the first specimen. The propagation of the maximal bond stress along the monitored bond length could not be measured.

For the B-8-SC reinforcement a maximal bond stress level close to 20 N/mm² was recognized for the L-shaped specimen. As for the other specimen this value dropped to 7.5 to 12-5 N/mm² level. At the L-shaped specimen high values were recorded in the first part of the bond length and with low values at the second part. In contrary to this, when the tension-tension setup was used the progressive failure was observed with a good mobilization of the bond capacity along the entire bond length. The slip at the unloaded end was around 1.5 mm for both test setups. Transverse deformations were again high in case of the tension-tension specimen reaching values of above 1 mm as for the other test setup reduced values of under 0.1 mm were measured. Mobilization capacity was also influenced form 76% it dropped to 59% if the tension-tension specimen was used.

B-8-SC reinforcements tested in the tension-tension tests failed by concrete cracking along inclined planes with failure starting from the bottom of the groove. The failure mode was triggered by the splitting of the specimen into two halves along the weakened plane

(Figure 52). In test with the L-shaped specimen shear-off failure of the sand coating was observed followed by splitting of the concrete cover along inclined planes starting from the middle plane of the reinforcement (Figure 44). This example shows a different failure pattern influenced only by the test setup boundary conditions. Similar observations has been made during tests in the gradual development stage of the L-shaped specimens for strip shaped reinforcements

The last tested round cross-section reinforcement in this chapter was a CFRP bar (C-6-SCW). Mobilization capacity was influenced by the boundary conditions. From 84% it dropped to 61% if the tension-tension specimen was used. The maximal bond stress level was not consistent at the L-shaped specimens but for the double pull-out specimen it was in the range of 10 to 15 N/mm². The strain measurement was disturbed for the tests performed with the L-shaped specimen but a high mobilization in the first segment was observed. The shift of the maximal bond stress is not recognizable. In contrary to this at specimens tested with the tension-tension specimen a textbook like distribution of the bond stress can be observed.

The effect of boundary conditions was shown for one strip shaped reinforcement (C3_1). The average value of the mobilization capacity dropped from 96% to 65%. The mean bond strength was 9.2 N/mm² for L-shaped setup and 3.7 N/mm² for the tensile-tensile test setup. The maximal bond stress level was in the range of 15 to 20 N/mm² for one specimen and dropped to 5 to 10 N/mm². The shift of the maximal bond stress towards the unloaded end with a good mobilization along the entire bond length was observed only for the tensile-tensile specimens.

The application of strain gauges to reinforcements in NSM applications is very difficult. Several preparation steps need to be followed with maximal care. The surface deformations have to be removed only on small areas without damaging fibres close to the surface or the continuity of the bond surface. Strain gauges have to be applied on clean surface and the measurement cables have to be covered with a deformable material (a larger diameter plastic tube was usually used) which can accommodate the deformation of the adhesive and the slip of the reinforcement. In the experiment with the L-shaped specimen when the strain gauges were applied to the C-6-SCW reinforcement the protection of the strain gauge cables was replaced with an alternative solution. As a result the measurements were unreliable.

Based on the herein presented research following recommendations can be formulated for experimental study of NSM reinforcements bond properties:

• The distance of the first strain gauge from the loaded end should be of minimal 30 mm, results obtained for strain gauge mounted at 10 mm were found not reliable.

• The application of the strain gauges is very difficult in case of NSM reinforcement. The strain gauges can considerably influence the bond behaviour, therefore their number should be minimized. Distance between consecutive strain gauges is recommended to be at least 50 mm or higher.

• Bond length need to be higher than 175 mm in order to show the progressive failure of NSM reinforcement

• Stiffness of the specimen should be as high as possible in order to avoid the influence of boundary conditions on the test results.

Bond characteristics of NSM reinforcements based on advanced test method PhD Thesis by Zsombor K. SZABÓ, supervisor György L. BALÁZS

• Bond strength measured using the tension-tension specimen was for each of the studied reinforcement lower than measured using the L-shaped specimen. In order to give reliable information on bond capacity the highest value should be used. This value will be considered to be valid for optimal confinement conditions. This value can be afterwards corrected with reduction factors for each particular application.

• Double tension-tension tests are highly labour and cost intensive. The failure at the unmonitored side could be hardly avoided. The wrapping of the specimen on the unmonitored side is recommended.