On-line Fault Diagnose of Distribution System Based on Modified Rough Sets Reduction Algorithm
Jing Dai · Proceedings of the CSEE · 2007
A rough sets reduction algorithm based on genetic algorithm is presented to the puzzle problem of key data acquirement in the real complex distribution system fault diagnosis with thousands of data.By this approach,we can get right diagnosis conclusion with less information.The genetic algorithm effect of genetic parameters to the evolutionary process is analyzed.Furthermore the fitness function,punishing function and punishing factor are emphasized to study.The reduction method with utilization of the capability of searching for global optimum of genetic algorithm achieves better reduction result compared with classical rough sets reduction algorithm.Example of America PGE distribution power system with 69 nodes shows that the attribute reduction by genetic algorithm accelerates the evolutionary process and avoids premature convergence effectively for the system with 202 attributes and 319 records.According to the example,it shows that this method makes the feasibility of fault diagnosis in complex distribution system with thousands of data.