Fault diagnosis method for wind driven generator set gearbox

Abstract

The invention discloses a fault diagnosis method for a wind driven generator set gearbox. The method mainly comprises the following steps of: firstly, by using an acceleration sensor, acquiring various fault types of vibration signals of the gearbox; then whitening the vibration signals to reduce a correlation between data samples; directly extracting features from original vibration signals by using a sparse filtering algorithm, wherein the algorithm can automatically complete the extraction as long as the number of features to be learned is set so as to avoid a manual feature extraction method that requires a large amount of signal processing knowledge and technology and to save time and effort; finally, classifying a gearbox failure type by using support vector regression. The support vector regression can directly classify multiple categories, does not need to build multiple binary classifiers, and has greater advantages in processing multi-class problems compared with a support vector machine. The fault diagnosis method for the wind driven generator set gearbox based on sparse filtering and support vector regression can improve the efficiency and accuracy of wind driven generator set gearbox fault diagnosis.

Description

technical field [0001] The invention relates to the technical field of mechanical fault diagnosis, in particular to a fault diagnosis method for a gearbox of a wind turbine based on sparse filtering and support vector regression. Background technique [0002] A wind turbine consists of wind rotors, low-speed shafts, gearboxes, high-speed shafts, generators, towers and other components. As an intermediate device for speed conversion of wind turbines, the gearbox is crucial to the normal operation of the entire unit. The failure of the gearbox will often cause the shutdown of the entire wind turbine, directly affecting the performance and safety of the wind turbine. Since the wind turbine gearbox usually operates at a height of tens of meters, it is more prone to failure and more difficult to repair than other rotating machinery. The investigation report on the operation quality of wind farm equipment shows that gearbox failure is the equipment that causes the longest downtim...

AI Technical Summary

Problems solved by technology

This method can realize the classification of faults. However, artificially extracting fault features from raw data requires a lot of prior knowledge and skills in signal processing, and selecting sensitive features also requires rich expert experience. The degree of feature extraction will depend on It will directly affect the accuracy of recognition, so manual feature extraction and selection is a very important and time-consuming work

Images