Multi-status-variable Perception-based Online Monitoring of Protective Relays
Due to inferiorities such as incorrect determination of fault nodes and low monitoring accuracy of equipment sta-tus with respect to the currently prevailing monitoring methods,a multi-status-variable perception-based online monitoring method for protective relays was designed.Through the constraint relationship between relay status informa-tion,the sampling value component vectors of different nodes were calculated to obtain equipment status information.In a calibrated environment,oscillations were eliminated by providing external damping,and the factors of detail in the ampli-tude were compared to obtain the final denoised status signal.The SSD monitoring model was established to generate the prior frames of different layer scales,and the loss function was used to match the prior frames of the fault images.The prediction frame of the highest priority was obtained,and the detection results of relay status were locked,so as to com-plete the monitoring.The experimental results showed that when the spatial correlation index was constantly lower than the time window of the threshold,the tested node was determined to be a faulty node.The monitoring accuracy of the ex-perimental group was always above 95%,showing satisfactory selectivity and accuracy in monitoring abnormalities.
万方数据
维普
