Effect of correlation between different ply failures on reliability of laminate composites

Abstract

The present study focuses on the correlation between different ply failures in laminate composites and its effect on the laminate reliability. First order reliability method (FORM) is employed to compute the failure probability of each ply, and the the correlation between different ply failures is derived from the linearized limit state function of each ply. Significant correlation between different ply failures is observed, and the magnitude of the correlation depends on the uncertainty level of stochastic variables. The correlation is demonstrated laying heavy influence on the laminate reliability. It is clarified that the max algorithm (adopted by previous studies) may provide a heavy underestimation on the laminate failure probability, depending on the magnitude of the correlation coefficient between different ply failures. On the contrary, the sum algorithm (adopted by previous studies) is found providing huge overestimation of the laminate failure probability due to its complete neglect of the correlation between different ply failures. To well account for the correlation between different ply failures in composite re liability evaluation, the integration of multidimensional normal distribution (IMND) is recommended to be employed.