Application of Improved Latin Hypercube Sampling in Reliability Assessment of Generation and Transmission Systems

An improved Latin hypercube sampling method is proposed to solve the problems of low computational efficiency and the correlation between computational time, amount of computation, and evaluation accuracy in the sequential Monte Carlo simulation for reliability evaluation of generation and transmission systems. The optimal multiplier is used to modify the approximate optimal parameter distribution of the system, and then sequential Latin hypercube sampling is performed to obtain the state duration time series of elements and systems. This method further reduces sample variance and improves sampling efficiency on the basis of conventional Latin hypercube sampling, solves the correlation problem between the computational efficiency and evaluation accuracy of the sequential Monte Carlo simulation effectively. Apply this method to evaluate the reliability of IEEE-RTS 79 test systems and high reliability systems, the results show that the proposed method has faster convergence speed, less computational effort, and less computational time, and has good applicability in high reliability systems.