Training method of anomaly identification model of vehicle network electrical coupling data

Abstract

The invention provides a training method for an anomaly identification model of vehicle network electrical coupling data. The training method comprises the steps of obtaining a training data set of vehicle network electrical coupling, wherein the training data set comprises labeled abnormal data; and inputting the training data set into a pre-established anomaly identification model for training to obtain a trained anomaly identification model, wherein the anomaly identification model comprises a feature extraction unit used for extracting low-level features and a residual block unit used for extracting high-level features. According to the method, voltage data or current data can be labeled, a training data set labeled with abnormal data is input into an anomaly identification model and training is completed, a vehicle network electrical coupling anomaly identification model based on data driving is obtained, vehicle network electrical coupling abnormal data is identified according to the waveform of a signal, the vehicle network electrical coupling state can be monitored in real time, and the actual engineering requirements are met.

Description

technical field [0001] The present application relates to the technical field of power system operation analysis, and in particular, relates to a training method for an abnormality identification model of vehicle-network electrical coupling data. Background technique [0002] In the field of electrified railways, there is a complex electrical coupling relationship between electric locomotives, EMUs, etc. and the traction power supply system, which is related to the safety, reliability, and stability of electrified railway operations. There are many types of anomalies caused by the deterioration of the electrical coupling relationship between trains and networks. Different types of anomalies have adverse effects on different aspects of electrified railway operations, threatening the safe and stable operation of trains, causing economic losses, and so on. [0003] At present, vehicle-network coupling anomalies mainly rely on experienced technicians for manual identification, w...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the embodiments of the present application is to provide a training method for anomaly identification model of vehicle-network electrical coupling data, so as to solve the problem that the traditional model-driven vehicle-network electrical coupling anomaly identification method cannot adapt to the situation of frequent occurrence of multiple abnormal types in practice

Images