Research on Power System External Fault Monitoring Techniques Based on Deep Learning

With the advancement of technology, the Global Positioning System (GPS) has played a pivotal role in various domains. Among these, China's BeiDou Navigation Satellite System has emerged as one of the four major global satellite navigation systems. To address the security concerns related to transmission lines, this study integrates the principles of BeiDou positioning, deep learning theories, and the interchange of intranet and internet data to develop an intelligent security and external damage monitoring technique. The technique employs mechanical equipment object recognition based on Mask-RCNN and YOLO. Experimental results indicate that the MEMark-RCNN technique achieves high detection precision and accuracy in various scenarios, with an average precision of 98.75% and an average accuracy of 98.6%. The research suggests that intelligent recognition techniques have broad application prospects in the design of external damage monitoring systems.