A Monte Carlo tool for multi-node reliability evaluation

Abstract

The presently used methods for calculating the probabilities and frequencies in the composite system ignore some important factors like common mode failures, effects of weather, dependent failures and overtrips. This thesis describes the development of a Monte Carlo program (MACS) which can generate contingencies and their probabilities and frequencies including common mode and dependent failures. This program has been further extended to perform the reliability analysis of interconnected power systems. This model is similar to the one used in the industry except that the transmission model used is more sophisticated. The convergence characteristics of this model are compared with another model also used in the industry. N-Area Reliability Program(NARP) is based on the random sampling of generator and transmission line status for each hour. Monte Carlo Approach for Estimating Contingency Statistics along with the Evaluation Subroutine(MACS-ES) advances the generation and transmission using the next event approach. Mathematical analysis and system studies indicate that MACS-ES converges slower than NARP. The simulation time for MACS-ES is generally more than twice the mean duration of loss of load. The CPU time per year of simulation is smaller for MACS-ES. However, the total CPU time for MACS-ES is much longer than NARP. For application testing the IEEE Reliability Test System has been used.