Reliability evaluation of composite power systems using sequential simulation with Latin Hypercube Sampling

Monte Carlo (MC) sequential simulation is capable of providing system chronological information. It is a very useful tool for power system reliability analysis and planning, especially for systems involving time-dependent sources. In this paper, a new sequential simulation approach is proposed for reliability evaluation of composite power systems. The main idea is to apply Latin Hypercube sampling (LHS) to generate the time duration of each system state in order to facilitate simulation convergence. LHS can yield system chronologies of higher representativeness than the conventional random sampling using the same number of sample states. The LHS-based sequential simulation approach is performed on IEEE Reliability Test System (IEEE RTS). It is performances are examined for several cases with different types of down time distributions of system components. Reliability indices including loss of load probability, expected unsupplied power, loss of load frequency and loss of load duration are estimated. The case results demonstrate that the LHS-based sequential simulation approach outperforms the conventional approach by reducing the variances of simulated reliability indices.