Intelligent Fault Diagnosis Method for Coupling Rotating Machinery Based on Deep Convolutional Neural Network

In the industrial process, coupling faults between different components tend to continually happen due to the complications of mechanical equipments. Compared to the single fault, coupling faults have complex characteristics and features, especially when they have the similar patterns. In this paper, a fault diagnosis method of rotating machinery is proposed to diagnose coupling machinery vibration signals based on deep convolutional neural network. The proposed algorithm coalesces both feature extraction and pattern classification into a single and compact learning network which can load raw signals without pretreatment. Through the wind turbine drivetrain diagnostics simulator (WTDS) platform, vibration signals are collected and the effectiveness of the proposed method is validated. The results demonstrate that the detection accuracy of the proposed method is higher than other intelligent fault diagnosis algorithms.