Efficient and accurate planetary gearbox fault diagnosis is the key to enhance the reliability and security of wind tur-bines. Therefore, an intelligent and integrated approach based on deep belief networks (DBNs), improved logistic Sigmoid (Isigmoid) units and impulsive features is pro-posed in this paper. Vanishing gradient problem is an in-herent drawback of conventional Sigmoid units, which usu-ally occurs in the back-propagation process, resulting in that the learning considerably slows down and the classi-fication rate is reduced. To solve this problem, Isigmoid units are designed to combine the merits of unsaturation from Leaky Rectified Linear (LReL) units. The results of handwritten digit recognition experiments show the supe-riority of Isigmoid over Sigmoid on convergence speed and classification accuracy. Since impulses contain much useful fault information, especially for early failures, an integrated approach using the optimized Morlet wavelet transform, kurtosis index and soft-thresholding is applied to extract impulsive components from original signals to improve the diagnosis accuracy. Then the features ex-tracted from original signals and impulsive signals are employed to train and test the DBNs with Isigmoid, Sigmoid and LReL units for comparison. Finally, the results of plan-etary gearbox fault diagnosis show that Isigmoid has high-er comprehensive performance than conventional sigmoid and LReL.
It is significant for diagnosing the early faults of wind turbine planetary gearboxes. However, the weak transient feature generated by the early gear failure is often submerged in other vibration signals and noise, thus this study discusses a new automatic sparse representation method for detecting weak transients. An improved Morlet wavelet which strictly satisfies the admissibility condition is constructed, and it is highly beneficial to sparsely separating signal component. By the use of the improved Morlet wavelet dictionary and Fourier dictionary, the transient component can be extracted by solving a sparse problem. Aiming at the drawback of the kurtosis and squared envelope spectrum (SES) negentropy in evaluating the similarity between the extracted transient and the original transient, an improved kurtosis index is designed, which can be used to search the optimal sparse parameters. Then by the use of the improved kurtosis index and iterative thresholding method, an adaptively iterative thresholding shrinkage algorithm is proposed to detect the repetitive transients from the planetary gearbox signal. The comparative results in simulation and experiments show that the proposed method can more effectively and accurately perform the fault diagnosis of wind turbine planetary gearboxes than the infogram method and the resonance-based signal decomposition method.