Influence of Random Variable Dimension on the Fast Numerical Integration Method of Aero Engine Rotor Disk Failure Risk Analysis
Abstract In the risk assessment of turbine rotor disks, the probability of failure of a certain disk type (after N flight cycles) is a vital criterion for estimating whether the disk is safe to use. Monte Carlo simulation (MCS) is often used to calculate the failure probability but is costly because it requires a large sample size. The numerical integration (NI) algorithm has been proven more efficient than MCS in conditions entailing three random variables. However, the previous studies on the NI method have not dealt with the influence of random variable dimension on calculation efficiency. Hence, this study aims to summarize the influence of variable dimensions on the time cost of a fastintegration algorithm. The time cost increases exponentially with the number of variables in the NI method. This conclusion provides a reference for the selection of probability algorithms involving multiple variables. The findings are expected to be of interest to the practice of efficient security design that considers multivariable conditions.