An automatic ridge estimation method for tacholess order tracking of wind turbine gearbox under varying-speed operation

Abstract The most widely used method for estimating the instantaneous shaft speed of a wind turbine gearbox is based on time-frequency ridge estimation. However, this method requires expert knowledge to determine the target component and the appropriate objective function to ensure the accuracy of the ridge estimation, which significantly reduces its applicability in practical applications. In this paper, an end-to-end automatic ridge estimation (ARE) method is proposed to address these limitations. First, based on the inherent harmonic relationship in the vibration signal of the wind turbine gearbox, a selection index for automatically determining the target ridge with clear physical meaning is proposed. Second, in order to be able to perform the estimation along the target ridge trajectory, a sub-region division strategy is proposed, which realizes the fusion of the one-step optimization strategy and the dynamic path optimization strategy, mitigates the interference of other components and improves the computational efficiency. Finally, a parameter-free objective function is introduced, which considers both the amplitude and smoothness of the ridge, getting rid of the parameter mismatch problem in the penalty term of the traditional method. To verify the performance of the ARE method, we collected the vibration signals of the wind turbine gearbox on site. Experimental results show that the ARE method achieves ARE and outperforms other methods in terms of accuracy and computational efficiency.