A Probabilistic Approach for Reliability Quantification of Propulsion Materials

A new probabilistic approach to assess the reliability of brittle materials for high temperature engine components was developed. The probability of failure due the fracture of a randomly oriented damage inherent from manufacturing process was determined by a hybrid probabilistic method (HPM) for computational efficiency and accuracy. The random variables considered by this approach for overall component failure probability include fracture toughness, loads, geometric factor, initial damage size, damage growth rate, damage orientation and damage occurrence in a unit volume. HPM consists of four modules: (1) an efficient reliability algorithm, (2) a surrogate model, (3) a bi-linear interpolation scheme and (4) a random simulation method. Seven cases with various types of loads and failure mechanisms were performed by both the hybrid approach and traditional Monte Carlo simulations (MCS). For all the cases studied, the results by HPM compare very well with that by traditional MCS. HPM also shows great efficiency over traditional Monte Carlo simulations. In addition, this new probabilistic framework allows further improvement in computational efficiency by exploring other efficient simulation methods.