Design of GIS Ultrasonic Partial Discharge Detection Signal Recognition Algorithm Based on Deep Q-Learning Network
In view of the low accuracy of gas insulated switdgear(GIS)equipment defect identification,a GIS ultrasonic partial discharge detection signal identification algorithm based on deep Q-learning is designed.The algorithm uses the ultrasonic of GIS equipment as the input data,decomposes the ultrasonic into multiple product components using the local mean decomposi-tion(LMD)algorithm,and uses it as the input of the deep DQN model to learn and train the DQN model.The DQN model af-ter training is used to realize the detection and recognition of partial discharge faults of GIS equipment.The results of numerical examples show that compared with deep belief network(DBN)and DQN algorithms,the average recognition accuracy of the proposed algorithm can reach 89.75％,which has significant advantages.In the actual operation and monitoring of GIS equip-ment in power grid,it is found that corona discharge and suspension discharge account for a large proportion,so the monitoring and prevention of these two kinds of partial discharges should be strengthened.
deep Q-learning networkpartial dischargeultrasonicGIS
万方数据
维普
