Wind turbine generator gearbox fault diagnosis method based on vibration signal blind source separation and sparse component analysis

Abstract

The invention discloses a wind turbine generator gearbox fault diagnosis method based on vibration signal blind source separation and sparse component analysis. The wind turbine generator gearbox fault diagnosis method mainly comprises three algorithms, which are an empirical mode decomposition (EMD), singular value decomposition (SVD) and K-means clustering (K-Means) based source signal number estimation algorithm, a fuzzy C-means clustering (FCM) based aliasing matrix estimation method, and a minimized l1-norm based source signal estimation and diagnosis algorithm. An idea of blind source separation (BSS) and sparse component analysis (SCA) is introduced into the whole algorithm process. The wind turbine generation gearbox fault diagnosis method takes simulated signals and actual vibration signals as a test object, detailed algorithm description is provided, and the effectiveness of the algorithms in the aspects of signal processing and fault analysis is verified through a series of experiments.

Description

technical field [0001] The invention belongs to the field of vibration signal processing and fault diagnosis methods, and in particular relates to a vibration signal processing step and a fault diagnosis method for a gearbox of a wind power generating set. Background technique [0002] The expensive operation and maintenance costs of wind turbines is one of the important factors hindering the rapid development of the wind power industry. With the continuous increase of wind power single unit capacity, the volume and hub height of wind turbines are also increasing, and the stress on the internal transmission system is more complicated. Accidents caused by faults of wind turbines often occur, resulting in huge economic losses. . The gearbox is located in the nacelle of the wind turbine. It is the main component of the transmission power of the wind turbine and an important hub connecting the main shaft and the generator. It has the characteristics of compact structure, high t...

AI Technical Summary

Problems solved by technology

However, the existing blind source separation technology still has certain limitations: on the one hand, most of the current blind source separation algorithms are for overdetermined or positive definite situations, that is, the number of observed signals is required to be greater than or equal to the number of source signals, and for underdetermined situations ( Research on Blind Source Separation when the Number of Observation Signals is Smaller than the Number of Source Signals

However, in many cases, this condition is not satisfied, resulting in a certain aliasing of the separated source signal, which affects the final diagnosis result

On the other hand, the real gearbox fault signal has the characteristics of non-stationary and nonlinear, and the existing methods usually perform poorly in dealing with such signals.

Images