Research Of Bad Data Identification And Correction In Kunming Power Grid Based On GSA Algorithm

Along with the automation degree of power grid become more and more high. The running of power system has become increasingly dependent on the data reliability, and accurate measurement data is the condition of keeping power system operation safe and stable. This paper studies the bad data identification and correction algorithm in power system, and compare with the performance between the algorithm and state estimation in Kunming power grid.At first I studied the state estimation based on the energy management system (EMS) of Kunming power grid dispatch center. The study include the main principle, procedure of algorithm, problems in the actual operation and other aspects of the state estimation system.In this paper, we use the gap statistical algorithms(GSA) to identify the bad data in power systam, and use the genetic-neural network algorithm to revise the bad data. The algorithm of GSA is consists of three parts-BP neural network, K-means clustering algorithm and gap statistical process. First of all, I made a thorough research of these algorithms, which include principles, proceduress, festuresand and other aspects. Then I use MATLAB program to compile these algorithms. What’s more, BP neural network has the characterize of slow convergence, so I use genetic algorithm to optimize the initial structure of neural network and propose to use the genetic neural network algorithm to correct the bad data of power system.At last, in order to verify the algorithm’s performance of the bad data identification and correction. I use the actual operating data of Kunming power grid dispatch center to emulate these algorithms, and compare the simulation result with the result of state estimation system. Point the strengths and weaknesses of the GSA algorithm and genetic-neural network algorithm in power system bad data identification and correction.