Composite pressure tanks have received increased attention across a number of civilian applications due to their lightweight and high strength. Traditionally, the design of composite vessels is based on deterministic analysis. However, the design of these structures is challenging and involves several kinds of uncertainties. In fact, different computational investigations have been carried out but no studies provide a resolution for small failure probability evaluation of composite pressure tanks. The aim of this study is to establish a computational framework to investigate small failure probability levels of composite tanks using the Subset Simulation method (SS). The model was developed in two steps, first, the development of limit state functions for hoop and helical layers using netting analysis, and afterwards, a probabilistic computation with six random variables. To quantify the effect of the randomness of different parameters on the
structural reliability of composite tanks, a sensitivity analysis was performed using different values of coefficients of variation (COV). It was observed from the results that SS has the ability and the accuracy required to evaluate small failure probabilities which are commonly encountered in composite tank applications. In addition, the hoop strength, the internal pressure, and the thickness of the composite are the major design variables that have a great impact on the
structural reliability of the axially symmetric composite tank whereas the fiber winding angle has little effect. Moreover, high COV values drastically reduce the safety zone, which could eventually lead to the burst failure of the composite pressure tank. Furthermore, this study implements a reliability-based design from the perspective of hoop and helical composite layer thicknesses, thus providing a rational assessment of the risk of structural failure.
