Residual strength prediction of composite materials: Random spectrum loading

The goal of this project is to identify if and how load order impacts residual strength in an E-glass/vinyl ester composite laminate subjected to variable amplitude fatigue loading. This paper presents results for constant amplitude loading data which, are used to fit parameters for a phenomenological model that can then applied to the spectrum loading cases. The residual strength distribution shape, in addition to median values, is modeled using Weibull statistics. Three cases are run experimentally and modeled for a 735,641 cycle spectrum containing 22 stress levels. The first two are ordered block loading, from highest stress to lowest and from lowest stress to the highest. In both cases, the model predicts the resulting residual strength distribution very accurately. A final case where the entire spectrum was randomized produced unexpected results with every specimen failing after 200,000-400,000 cycles while the model predicts identical residual strength when compared with the block loading case. This work points to a dire need for focus on developing a better understanding of load order impacts in design of composite structures based on constant amplitude fatigue tests.