
Research on Fault Diagnosis Method of Wind Turbine Bearing Based on Deep belief Network
Author(s) -
Liang Wang,
Yuanyuan Ma,
Xiaoming Rui
Publication year - 2019
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/677/3/032025
Subject(s) - bearing (navigation) , turbine , fault (geology) , deep belief network , wind power , short time fourier transform , vibration , signal (programming language) , condition monitoring , frequency domain , computer science , engineering , time domain , control theory (sociology) , artificial intelligence , artificial neural network , fourier transform , acoustics , computer vision , aerospace engineering , geology , fourier analysis , mathematical analysis , physics , electrical engineering , mathematics , control (management) , seismology , programming language
Bearings of wind turbines have become one of the components with high failure rate in wind turbines because of their bad operating environment. In this paper, a fault diagnosis model based on deep belief network is proposed for bearing fault diagnosis of wind turbine. The time-frequency spectrum of wind turbine bearing vibration data after short-time Fourier transform (STFT) is used as the input of fault diagnosis model, and the output is the identification code of various fault types of wind turbine bearing. Compared with the deep belief network diagnosis model based on the time domain signal input to the vibration data of wind turbine bearings, the deep belief network fault diagnosis model based on the short-time Fourier transform of the input signal has higher recognition accuracy. Based on the vibration data of different working conditions and rotating speeds, the model can automatically find fault features and identify the faults of rolling elements, inner rings and outer rings of rolling bearings at different locations, thus avoiding expert experience and feature engineering, making the model more versatile and generalizable and potential for efficient on-site rolling bearing fault diagnosis.