Permanent magnet synchronous motor fault diagnosis method based on deep learning

Abstract

The invention discloses a permanent magnet synchronous motor fault diagnosis method based on deep learning. Main faults of a permanent magnet synchronous motor comprise stator winding open circuit, stator winding short circuit, eccentricity, demagnetization, rotating shaft bending, bearing faults and the like. According to the method, a stator current signal, a vibration signal and a noise signalof a motor are respectively acquired through a current sensor, a vibration sensor and a sound sensor. The method comprises the following steps: 1) preprocessing a signal; 2) extracting the time-domain, frequency-domain and time-frequency characteristics of the signal; 3) training a deep neural network based on characteristic big data; and 4) diagnosing the fault type of the permanent magnet synchronous motor based on deep learning. According to the method, multiple sensors and multiple fault features are fused, and the method has higher applicability than similar methods.

Description

technical field [0001] The invention belongs to the technical field of motor state detection and fault diagnosis, and more specifically, relates to a fault diagnosis method for permanent magnet synchronous motors based on deep learning. Background technique [0002] The permanent magnet synchronous motor has a simple structure, smaller volume, lighter weight, no heating problem of the rotor, large overload capacity, small moment of inertia, and small torque ripple. It also realizes brushless and improves the efficiency of the motor. Therefore, permanent magnet synchronous motors have been widely used in fields such as elevators, electric vehicles and ship electric propulsion. [0003] However, due to the complex operating environment and material life, manufacturing defects, or intermittent operation, PMSM inevitably suffers from various fault types such as stator winding turn-to-turn short circuit, eccentricity, and bearing failure. In addition, due to the replacement of t...

AI Technical Summary

Problems solved by technology

[0004] In the previous research on fault diagnosis methods, domestic research on permanent magnet synchronous motors is still relatively small, and most of them only study the most common inter-turn short-circuit faults of stator windings, and only collect a single signal of stator current or vibration, and rarely use Method of multi-sensor synthesis and multi-feature fusion

In addition, the previous fault diagnosis technology usually only extracts the frequency domain or time-frequency characteristics of the signal for analysis, and technologies such as big data and deep learning are rarely used in the fault diagnosis of permanent magnet synchronous motors.

Images