System design method for minor fault detection and position of electrical traction system

Abstract

The invention discloses a system design method for minor fault detection and position of an electrical traction system. The method comprises the following steps: 1) establishing an off-line data model: collecting sensor steady-state operation data of the electrical traction system and carrying out pretreatment on the off-line data; calculating values of a principal component and a residual component in different layers of subspaces of the data obtained after pretreatment and a load vector; determining and calculating a performance index and a probability density function of a principle component analysis method and a fault detection threshold; and 2) on-line fault diagnosis: processing online data; calculating a performance index of the principal component and the residual component in different layers of subspaces; and through the obtained fault detection threshold, constructing a probability matrix and carrying out fault diagnosis through Bayesian reasoning. The method can carry outeffective multi-feature description on tiny signals of an electrical drive device before occurrence of a fault, and can also carry out real-time online fault diagnosis under the condition that the model and parameters of the electrical drive system are unknown.

Description

technical field [0001] The invention belongs to the field of fault detection and fault location, in particular to a system design method for small fault detection and location of an electric traction system. Background technique [0002] Today, AC electric drive systems are widely used in the field of transportation, such as high-speed trains, subways, electric vehicles, etc. In order to improve the safety and reliability of the system, fault detection and diagnosis (FDD) of AC electric drive system has been one of the main issues of research and practical activities in the past two decades. [0003] In general, there are three main types of faults in AC electric drive systems: mechanical faults, electrical faults, and sensor faults. Electric traction systems are generally equipped with different types of sensors, such as: voltage sensors, current sensors, speed sensors, etc. Current sensors and speed sensors are mainly used for closed-loop control, and voltage sensors are...

AI Technical Summary

Problems solved by technology

The accuracy of traditional fault diagnosis methods based on observers and odd-even space methods depends on the accuracy of the established nonlinear system model. Generally, when the model-based method is used to model the actual AC electric drive system, some linear Hypothetically, this is the main cause of false positives and false negatives in troubleshooting

Images