Wind turbine bearing fault diagnosis based on sparse representation of condition monitoring signals

Abstract

Sparse representation of massive condition monitoring signals is an effective approach to save the cost for data storage and transmission in fault diagnosis of wind turbines. This paper explores a sparse representation method over a new dictionary designed particularly for bearing vibration signals of wind turbines operating under varying-speed conditions. First, the time-varying shaft rotating frequency (SRF) of the wind turbine contained in the vibration signal(s) is estimated from a generator current signal recorded synchronously with the vibration signal. Then, a new dictionary is designed, in which the basic functions, called time-varying cosine packet, change with the SRF. Finally, a sparse coefficient spectrum (SCS) of the vibration envelope signal is constructed by the sparse coefficients of the signal projection on the new dictionary. The merit of the proposed sparse representation method is that the dictionary designed is adaptive to the variations of major frequencies of the vibration signals. The nonzero sparse coefficients over the dictionary designed represent the major order components contained in the vibration signal. Therefore, the possible bearing fault characteristic orders can be identified from the SCS of the vibration envelope signal. Laboratory and field test results show that the sparse coefficients obtained by the new sparse representation method are suitable for the representations of the bearing vibration signals and the SCS is effective for bearing fault diagnosis of direct-drive wind turbines.