Reliability Evaluation Of Power Transmission System Based On The Improved Importance Sampling Algorithm

With the development of the economic, electricity consumers in all kinds of sectors have a increasingly high demand of the power supply reliability. That’s because negative impact of power failure is getting bigger and bigger. To ensure safe and reliable power supply is of vital importance. Although reliability evaluation is very complex now for the structure of power system getting much more complex than before, we need to finish it to find out the reference for the programming, operation and improvement of power system.Among all kinds of reliability evaluation methods, Monte Carlo method is the most widely used one. This method is especially suitable for large-scale power system reliability assessment and calculation. That’s because this method’s sampling frequency has nothing with power system scale. The biggest problem for the method, however, is that in order to obtain high precision reliability index needs a large number of sampling calculation. The contradiction between accuracy and calculation speed greatly limits the convergence speed of reliability evaluation. With the development of components’ reliability, the reliability of the transmission system is becoming better and better. But it also brings increasing difficulties to the convergence of reliability assessment at the same time. Aiming at the problem of Monte Carlo method, scholars have put forward various sampling algorithms, in order to speed up the convergence rate by reducing variance. Among those methods, the importance sampling method is considered as the most efficient one. This method is quite effective in power generation system,which has a higher failure rate. While in transmission system with low failure rate and good reliability, the convergence effect of the traditional algorithm, that obtain optimal multiplier k through iterative, is not obvious.In order to solve the problem above, the value and correction method of optimal multiplier A; was improved in this paper. The idea of dividing sampling into two steps, pre sampling and formal sampling, was put forward. Firstly, determine the value of k by pre sampling of small samples, then calculate system reliability indices by formal sampling. Reliability calculation for transmission part of the reliability testing system RBTS and IEEE-RTS was done, by applying Conventional Monte Carlo method, importance sampling method and improved importance sampling method respectively. Analysis and comparison of the calculation results and the calculation precision, verified the accuracy and efficiency of improved method. In order to test the applicability of improved method, taking a practical power system M in a certain area as the research object, the reliability indexes of the whole grid and each node were calculated by the above three methods. This paper compared calculation results and schedule of the three methods and completed the verification. In the last, combining with M network topological structure and component parameter characteristics, it found out the reliability weak link in the area, proposed solutions to improve system reliability, and verified the effectiveness of these solutions.