Repository logo

Longitudinal shear resistance of steel and concrete composite beams



Change log


Hicks, Stephen James 


In this study the longitudinal shear resistance of steel and concrete composite beams using conventional headed stud connectors has been investigated through the testing of 42 push-out specimens. In addition, the applicability of the existing design codes and the standard specimen used in assessing the strength of stud connectors, has also been examined.

Push tests are typically employed to find the experimental resistance of stud connectors and in the past have been used to develop and calibrate the equations in the existing design codes. From the results of the new push tests, an additional parameter which affects the experimental strength of headed stud connectors has been identified as the generation of frictional forces developing at the interface between the base of the specimen and the reaction floor. The existence of such forces, which can significantly influence the apparent shear capacity developed, appears not to have been identified previously. In addition, detailing at the base of specimens which incorporate decking also appears to affect the ability of the studs to transfer shear.

The new experimental results, presented here, as well as those found in the current literature, indicate that the existing reduction factor method employed in current design is not entirely appropriate. Failure of the new specimens with transverse decking was characterised by the formation of cones of concrete around the stud connectors. The shear transfer resistance of the studs was strongly influenced by the geometry of these surfaces, which in turn was influenced by the deck geometry and the arrangement of connectors. Therefore, three new generalised concrete plasticity theory expressions based on idealised postulated shear cone geometry's have been developed. After calibrating these models with experimental evidence to determine the appropriate effectiveness factor which should be applied to the concrete strength in the theoretical expressions, the new models have shown promising correlation with the new experimental results as well as data from 90 push tests conducted elsewhere. In addition to having the advantage of being based on observed failure forms, all of the new cone models tended to yield more reliable predictions than the existing design codes.

Two specific longitudinal splitting modes were experienced in the new tests, which were strongly influenced by the arrangement of the stud connectors. Also, the current code assumption that decking in the transverse orientation participates in the resistance to splitting, appears to be not entirely justified by the results of the new tests. In some circumstances the current design methods gave slightly unsafe predictions. Therefore. five new expressions based on concrete plasticity theory, have been developed for this form of failure. After subsequent calibration to determine the required effectiveness factor. these models showed better correlation than the existing design methods when compared with both the new results, and with those of beam and push tests reported in the literature.

From the work presented in this dissertation, it is concluded that some of the scatter observed in previous calibration studies on push test results may be due to the significant (and previously unaccounted for) amounts of frictional force arising in different laboratories between the base of the specimen and the reaction floor. Also, a tighter standardisation of the current code of practice push specimen would be beneficial, particularly for cases when deck-slabs are employed. With respect to the theoretical estimation of the strength of stud connectors welded through decking, the three new shear cone models appear to form a more rational basis for design in comparison with the existing reduction factor methods. Finally, the estimation of the longitudinal splitting strength using the newly developed concrete plasticity expressions appear to yield much more reliable results than existing methods, and are therefore recommended.






Doctor of Philosophy (PhD)

Awarding Institution

University of Cambridge