Identifying the Internal and the External Overvoltage of Distribution Networks Based on Fisher Discriminate Method

It has been testified by running practice to that the internal and the external overvoltage of distribution networks are the main causes of electrical accidents. To capture and analyze the overvoltage in power system, authors have developed a set of overvoltage on-line monitoring system, and it has been running in a 10 kV distribution network for more than 36 months. Thousands of detailed data files have been collected and saved. Based on the brief introduction to the structure and the operational principle of the monitoring system, this paper presents an intelligently identification method of the internal overvoltage and the external overvoltage based on Fisher discriminate method. Firstly, the statistical parameters should be extracted from the original data recorded by the on-line monitoring system. According to the performance of many parameters during the period of the overvoltage occurrence, 8 of them such as time to crest, time to half value of voltage, total duration of overvoltage and etc. are selected as the characteristic parameters of identifying overvoltage types. Secondly, according to the operational data tracked by other monitoring instruments in the field and the daily report, the corresponding reference sample for either overvoltage are brought forward. Then, the identifying function can be constructed by the Fisher discrimination method. Any 8-dimensions data will be projected to 1-dimension. Lastly, the unknown overvoltage data samples are substituted to the identifying function. The distinguished results to 266 samples indicate that the outcomes are accordant, the identifying method along with the identifying function are correct and effective.