Reliability Evaluation Of Generation And Transmission System Based On Improved Monte Carlo Method

The power system reliability evaluation methods can be divided into deterministic methods and probabilistic methods. As the latter methods considering the stochastic characteristics of events so that the risk of the power system can be reflected more accurately. So the probabilistic methods have been widely used. The probabilistic methods can also be divided into two kinds: analytical method and Monte Carlo method. The state space of the composite generation-transmission system which involves a lot of components and contains all kinds of random factors is enormous. So, it is almost impossible to enumerate each failure state by the analytical method. The Monte Carlo is not limited by system scale so that it is more suitable for that system. The disadvantage of Monte Carlo method is that the evaluation time is closely related to the ca evaluation precision. In order to obtaining a high precision result, it often takes a lot of computing time. The convergence rate of reliability indicators is slow. At present, most of the reliability evaluation softwares of the power system improve the convergence rate of Monte Carlo method by using variance reduction technology. The common variance reduction methods include importance sampling method, variable control method, antithetic method, scattered sampling method and so on. Each of these methods has its own characteristics and can accelerate the convergence rate to a certain extent. But these methods have a common disadvantage, poor adaptability to the system.To solve this problem, this paper presents an improved sampling method named scattering-roulette method which is the combination of scattered sampling method and splitting-roulette method. This method divides the power system into important regions and non-important regions according to the contribution for reliability index.Using the roulette way to sample in the non-important regions and using scattering technology to sample in the important regions, the sampling time and the generation of random numbers can be reduced. So, the evaluation efficiency can be improved.This method reduces the sampling variance and judges the system state at the same time. It shows the good adaptability to the system. Based on the method proposed in this paper, the IEEE-RTS79 and the MRTS have been evaluated. Contrasting with the traditional Monte Carlo method and splitting-roulette method, the method proposed in this paper is proved to be accuracy, efficient and has a good adaptability to different systems.In this paper, the original reliability evaluation software is used to make reliability modeling for a provincial power grid which contains 83 generators, 71 transformers, 549 bus lines, 1173 lines and 523 loads. The real power grid is evaluated by 3 kinds of methods and the assessment results confirm the improved method's applicability in engineering.