Abstract
The reliability of components is the basis for reliability evaluation of the power system. Wrong reliability parameters of components will inevitably lead to erroneous evaluation results, which will influence the accuracy of power system planning. With the popularization and installation of intelligent electric meters, relatively accurate system/node reliability index can be easily obtained. This paper proposes a method to solve inverse problem of reliability evaluation (IPRE), which can overcome the disadvantages of traditional numerical solution methods that can only obtain local optimal solutions and overly depend on initial values. First, using the non-sequential Monte Carlo method (NSMC), the analytical expressions of system indexs and variable reliability parameters are established. Based on analytical expressions, a large number of training data samples can be generated. Secondly, extreme learning machine algorithm (ELM) is proposed to obtain the approximate value of unknown component parameters. Finally, three cases prove the accuracy and feasibility of the proposed method.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call