A Method For Detecting Abnormal States Of Power Equipment Based On Self-organizing Mapping Neural Network

To improve the utilization of equipment information and alert potential equipment failures in power equipment anomaly detection methods, a device status data anomaly detection method based on self-organizing neural networks and time series is proposed. Combining big data analysis technology and equipment evaluation technology, using time series autoregressive models and self-organizing mapping neural networks (SOM) to discretize continuous power equipment data into a single sequence. Calculate the transition probability of state variables on the timeline and quickly detect data anomalies through transition probability and clustering methods. The experimental results show that the proposed method can quickly and effectively detect abnormal states of power equipment.