A Fault Diagnosis Method for Electric Drive System

Abstract

The invention discloses a fault diagnosis method for an electric drive system. In the method, the collected motor phase current is respectively passed through a Clark / Park transformation, a filter, a Park vector processor, a normalization processor, an average value calculator, a second An extreme value difference processor, an absolute value calculator, an absolute value averaging processor, and a second extreme value difference processor perform conversion and processing, and use the average current extreme value difference to judge the fault criterion and the absolute value of the average current extreme value difference judgment The fault criterion is the fault diagnosis criterion, so as to realize the real-time detection and location of the electric drive system fault. This method does not need to add additional sensors, is simple and easy to implement, and can diagnose and locate open-circuit faults of a single power tube or multiple power tubes in real time. When the system load changes suddenly or the speed changes, it does not need to adjust the fault judgment threshold. Robustness Strong, not easy to misdiagnose.

Description

technical field [0001] The invention relates to the technical field of fault diagnosis of an electric drive system, in particular to a fault diagnosis method of an electric drive system. Background technique [0002] With the development of multi-electric, all-electric aircraft and hybrid and pure electric vehicles, the motor drive system has ushered in new development opportunities and challenges. In addition to requiring high power density and high efficiency, it also needs to have high output performance and high reliability. performance, which has become the key to the motor drive system. However, when the electric drive system fails, the motor will run asymmetrically, and the output torque will pulsate, resulting in large mechanical noise, which will lead to the decline of the overall performance of the system, especially the greatly reduced output power, or even malfunction, seriously endangering the The safety of the system, in order to avoid or reduce the serious ha...

AI Technical Summary

Problems solved by technology

The average current Park vector method is one of the better methods. Professor A.J.Marques Cardoso proposed the current Park vector method for inverter fault diagnosis, but this method requires a very complicated pattern recognition algorithm and is not suitable for integration into the drive controller.

Subsequently, the average current Park vector method was proposed. This method is carried out in the d-q coordinate system and can quickly judge and determine which IGBT is faulty.

In order to overcome these shortcomings, relevant scholars have proposed the normalized average current method and the normalized average current absolute value method. The judgment threshold is too dependent, and there is still a problem of misjudgment when there is a sudden change in load and speed or when multiple power tubes are simultaneously open-circuited.

Images