LSSTM Algorithm-based Automatic Fault Location for Distribution Networks
Currently prevailing distribution network fault location methods suffer in practice high error rate of faulty zone determination and high RMSE.The present work studied an LSSTM algorithm-based fault location method.The method was designed to determine the perception target based on fault characteristics of distribution networks,to use wireless sensors to perceive fault state vectors,to sparsely filter fault state signal by adopting SF algorithm,and to learn and train fault state features by employing LSSTM algorithm,thereby determining the faulty zone.The proposed method achieved in verification experiment an error rate of faulty zone determination within 2%,and a root mean square error of less than 0.2.