Transformer fault diagnosis method and system based on multi-source data fusion

• Main Authors: ZHONG YOUPING, LIU JIA, XU BO, YANGFAN, LI YINGHAN, WEI YIJUN
• Format: Patent
• Language: Chinese; English
• Published: 07.06.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention belongs to the technical field of transformer fault diagnosis, and discloses a transformer fault diagnosis method and system based on multi-source data fusion, and the method comprises the steps: collecting oil chromatographic analysis data, a high-voltage bushing infrared detection map, a discharge ultrasonic detection map and an ultrahigh frequency partial discharge detection map, and carrying out the feature extraction respectively; splicing to obtain a multi-source fusion feature vector; optimizing the weight of each layer of RBM network of the deep belief network by using a black widow optimization algorithm, inputting the multi-source fusion feature vector into the optimized deep belief network for training, and performing transformer fault diagnosis through the trained deep belief network; according to the method, the transformer fault diagnosis precision is improved, the deep belief network is optimized through the black widow optimization algorithm, and the problem that the deep belief