03.35: Rivet shear strength in long connections

Abstract

ABSTRACTRivet shear limit states are based on historical test data. Several different grades of rivets and plate material were used, which were accommodated by transforming the test data into a non‐dimensional form. Because an exact solution for rivet shear in a connection requires solving non‐linear simultaneous equations, a more practical empirical solution is proposed that eliminates this requirement. A statistical analysis was performed using the existing partial factor, of 1.25 to demonstrate that an acceptable reliability, β, close to or slightly greater than the desired 4.0 is obtained for all connection lengths when a length reduction factor R2 (βLF) of 0.90 is used, thus eliminating the need for a variable connection length reduction factor, βLF. The data was also analyzed for the case with no length reduction factor R2 (βLF) = 1.0. The statistical analysis of the test data indicates that there is no justification for a variable shear strength reduction due to connection length provided there is only a limited amount of second‐order effects in combination with gross and net areas equal to or greater than the specification limits.The assembled research data, based on rivet strengths of 392 N/mm2 and 523 N/mm2, were compared with the Eurocode 3 (EN 1993) provisions to determine compatibility. It was determined that the research test data did not support the Eurocode 3 provisions for a variable shear strength for rivets in connections longer than a specified basic length based on rivet diameter.This investigation identified a riveted connection condition wherein with appropriate connection strength and quasi‐stiffness, there would be no need to reduce the rivet shear strength with increasing connection length. Typically it was found that this condition was achieved when both current Eurocode 3 checks for the limit state of yielding on the gross section and the limit state of fracture on the net section were satisfied.