Abnormity detection method for traction substation

Abstract

The invention provides a traction substation anomaly detection method, which comprises the steps of firstly establishing an anomaly detection data set, calculating a background condition cluster of the data set, performing depth feature inner distance information extraction on an input image by constructing a distance feature extraction model, and then constructing an anomaly detection model, the detection model is trained by using the distance features, the output of the model is an abnormal score chart corresponding to the input image, finally, binarization and statistical analysis are performed on the score chart to obtain an abnormal detection result, including whether an abnormality occurs and the position where the abnormality occurs, and meanwhile, online updating is performed on the detection model to obtain an abnormal detection result. And the device can adapt to substation environment change. Based on the technical scheme of the invention, the requirement of abnormal detection of the traction substation can be effectively met.

Description

technical field [0001] The invention relates to the technical field of intelligent vision and intelligent systems, in particular to a method for detecting abnormalities in traction substations. Background technique [0002] As the key power supply facility of the high-speed rail traction power supply system, the traction substation mainly undertakes the task of transforming the electric energy of the power system to supply the high-speed rail EMU, and its operation safety issues have become increasingly prominent. Abnormalities that occur in traction substations are accidental and uncertain. If these abnormalities cannot be found in time and corresponding measures are taken, it will pose a great threat to the operation safety of the substation, and even cause serious safety accidents. The cause of abnormalities in traction substations may be due to factors such as equipment, environment, and human factors. If these substation abnormalities or events can be detected in time b...

AI Technical Summary

Problems solved by technology

Most of the existing anomaly detection methods need to determine the anomaly type and a large amount of labeled data in advance to train the model to exert its detection ability, which is contradictory to the actual situation of the substation, and it is difficult to deal with the anomaly of the substation. existential complexity

Images