DFIG Bearing Fault Diagnosis Based on Time-Frequency Domain Analysis

Abstract

In the context of global energy transition, doubly-fed wind turbines have become mainstream, and rolling bearings, as their core components, are prone to faults (accounting for over 30% of total wind turbine faults) due to harsh working conditions, leading to unplanned shutdowns and high maintenance costs. Traditional fault diagnosis methods have shortcomings such as poor noise resistance and insufficient accuracy, making early precise diagnosis an urgent need. To solve these problems, this paper proposes a doubly-fed wind turbine bearing fault diagnosis technology based on joint time-frequency domain analysis, which includes vibration signal collection, preprocessing, time-frequency domain feature extraction, fault characteristic frequency matching, and fault location/severity determination. Taking the FAG 6332-M bearing at the drive end of a 1788r/min doubly-fed wind turbine as the research object, the diagnosis results show that the method can accurately identify the outer ring fault, with a diagnostic accuracy of 100%. It can effectively avoid unplanned shutdowns and reduce maintenance costs by more than 40%.