Cyclic Tensile Loading of Point Supports

The following experimental results for cyclic tensile loading of point supports were already presented in detail in an earlier paper [3]. A brief discussion is given here, additionally covering the durability aspects. Regarding the

FIG. 9—Averaged tensile test results for incomplete mixing cases.

Copyright by ASTM Int’l (all rights reserved); Tue May 6 12:07:08 EDT 2014 Downloaded/printed by

FIG. 10—Tensile tests strength for incomplete mixing cases.

performance of the silicone adhesive under cyclic load regimes, a representative application was selected in the form of a planar bonded circular point support

[3] ; see Fig. 12. If the silicone material is free to deform, a significant lateral con­traction of the material appears under tensile loads, such as is observed in case of a dog-bone test specimen used in the section titled “Incomplete Mixing Procedures" during the study of varying mixing ratios. If the silicone adhesive is bonded to a significantly stiffer material as in the case of the point supports, the lateral contraction of the almost incompressible silicone [4] is suppressed at the

FIG. 11—Shear test strength for incomplete mixing cases.

Copyright by ASTM Int’l (all rights reserved); Tue May 6 12:07:08 EDT 2014 Downloaded/printed by

interface leading to complex 3D stress states and an increased effective stiffness of the bonding under tensile loads [5].

This specific behavior is also observed with respect to the failure of the sili­cone material. Under simple one-dimensional loading schemes such as tension or shear, the silicone adhesive suddenly fails, as shown in Fig. 1 and Fig. 2 for dog-bone tests and in Fig. 3 and Fig. 4 for ETAG 002 shear tests. On the con­trary, it is well-known that the failure of point supports under tensile loads shows a more complex pattern [6]. In Fig. 13 the test setup is shown using a spe­cial point support specimen which is made from stainless steel for improved bonding geometry accuracy. Cross-checks with conventional point support specimens bonded to glass samples have demonstrated similar mechanical characteristics. Figure 14 shows the load curve obtained for a 5 mm thick sili­cone adhesive bonded to a 50 mm diameter button, measured at room tempera­ture and a 1 mm/min displacement rate. The load curve features three distinct areas of behavior similar to those previously observed for U-type bonding

Copyright by ASTM Int’l (all rights reserved); Tue May 6 12:07:08 EDT 2014 Downloaded/printed by

FIG. 14—Load curve for bonded point supports under tensile loads.

geometries [7]. The first area is characterized by a high stiffness, visible by the large positive slope at the beginning of the diagram, which indicates a fully functional adhesive material. The second area differs by a significantly reduced slope of the curve which is obviously related to damage mechanisms in the ad­hesive material. Finally, the bonded point support fails due to the propagation of cracks in the adhesive until total separation occurs.

Cyclic loading has been applied to bonded point supports for various load levels in order to check mechanical integrity. In order to avoid problems of a backlash nature evoked by load direction changes, the lower reversal point of the cycles was set to a minimum tensile load of 100 N. Taking the non-linearity of the material behavior into account, the upper reversal point was determined by displacement levels, not by load levels. The upper boundaries (reversal points) were varied between 0.25 mm for the smallest amplitude and 1.5 mm for the largest amplitude leading to cyclic loads between 0.4 MPa and 1 MPa, which are higher than the usual design strength level of 0.14 MPa. Therefore, these tests can also be seen as exploratory tests for the potential future modifi­cations of design stress limits and safety factors. Regarding the load history, 100 cycles were selected as a compromise between the test duration on the one hand, and a sufficiently high number of loadings in order to introduce some fatigue on the other hand. After the last cycle, the specimen was loaded until complete failure.

Two aspects are of great interest with respect to the cyclic loading of point supports: the behavior during cyclic loading and the behavior after cycling. The degradation of the material during the cycles was assessed by the displacements for the lower reversal point and the maximum loads at the upper reversal point. Since only one load rate was investigated during the campaign, it is not possible to differentiate between visco-plastic and visco-elastic effects of the adhesive material in the analysis. The failure behavior of the specimens after the cyclic loading is of special interest with respect to the remaining load bearing capabil­ities after the cyclic loading scheme.

Copyright by ASTM Int’l (all rights reserved); Tue May 6 12:07:08 EDT 2014

Downloaded/printed by

Rochester Institute Of Technology pursuant to License Agreement. No further reproductions authorized.

Figure 15 shows the cyclic loading scheme related to the low amplitude test case of the 0.25 mm upper limit (reversal point) for a bonding geometry of 50 mm in diameter and 7 mm in the adhesive thickness. For a quantitative com­parison of the different load levels, the slopes of the various cycles obtained by secants through the upper and lower reversal points were added to the figure. For low amplitude cycles the slopes do not significantly change for an increasing number of cycles, which obviously implies that there is no significant loss of ma­terial integrity during small load cycles. In Fig. 16, the stress versus displacement cycles for the high amplitude test case are plotted. In contrast to the low ampli­tude case, the slopes significantly decrease for an increasing number of cycles, probably evoked by material damages. In order to allow a quantitative assess­ment, the relative change of loads is plotted for the various test cases in Fig. 17. As expected, the amplitudes are linked to the load reductions. Furthermore, the low and medium amplitudes show an asymptotic behavior, while for the large

Copyright by ASTM Int’l (all rights reserved); Tue May 6 12:07:08 EDT 2014 Downloaded/printed by

FIG. 17—Relative load degradation for various cycle amplitudes.

amplitude test case it is not obvious whether an asymptotic value is obtained for a very high number of cycles.

In order to allow an assessment of the remaining load bearing capability after cyclic loading, Fig. 18 shows the loads experienced during the final test step until failure. In this figure, the cyclic load history is deleted from the curves by con­necting the load history before cyclic loading directly to the load history after the cyclic loading. In this figure, it is quite obvious that the maximum load bearing capability after the loading cycles is significantly reduced for the largest ampli­tude, which is not the case for the lower amplitudes. The step decrease observed for the largest amplitude is related to the reduction of loads during the cyclic load scheme, which is eliminated from this figure. Since there is apparently no recov­ery of the bonding for the high amplitude case, the high load cycles lead to signifi­cant material damage beyond the Mullins effect. The Mullins effect only refers to a softening of the material below the experienced maximum load but does not affect the material behavior beyond the experienced maximum loading.

This test campaign demonstrated that the impact of cyclic loading might have an impact on the ultimate load bearing capability of the adhesive, depend­ing on the dynamic load levels. The test results are in agreement with the

FIG. 18—Overall load curve for various amplitudes (cycles suppressed) [3].

Copyright by ASTM Int’l (all rights reserved); Tue May 6 12:07:08 EDT 2014 Downloaded/printed by

current strength design limit used in actual field applications (0.14 MPa) since no obvious degradation effects have been observed for low stress levels. The effect of dynamic loading on structural integrity is covered today by the exist­ing, quite conservative, and thus quite high, safety factors with respect to design stress values derived from quasi-static tests. The current test results demon­strate that a more precise determination of design limits aimed at higher design stresses, and thus lower safety factors, should also consider the impact of dynamic loading on the performance of the adhesive.