K-Means Clustering Monitoring Simulation of Operating Status of Ultra-High Voltage Direct Current Primary Equipment
Aiming at the problem that abnormal operation status of UHVDC primary equipment may lead to un-stable operation of power grid,a monitoring method for operation status of UHVDC primary equipment is proposed and simulated.First,a dynamic clustering algorithm based on energy balance is constructed to control the wireless sensor network to collect the operating status data of the UHVDC primary equipment.The k-means algorithm is used to dy-namically cluster,balance the energy of each cluster through the competition function and energy ratio,and divide the nodes in the cluster into nodes of different levels to efficiently collect the operating status data of the primary equip-ment.The improved sparse principal component analysis method is used to select the key features of the primary e-quipment operation status data collected,and finally the operation status data after feature selection is input into the BP neural network after training to realize the operation status monitoring of UHVDC primary equipment.The experi-mental results show that the proposed method has excellent data acquisition capability,and the PR curve and ROC curve in operation state monitoring are more ideal,and the AUC value is higher.
Ultra high voltage direct currentPrimary equipmentOperation status monitoringSparse principal component analysis methodBack Propagation neural network
万方数据
维普
