A Novel Stacked Auto Encoders Sparse Filter Rotating Component Comprehensive Diagnosis Network for Extracting Domain Invariant Features

Abstract

In recent years, the method of deep learning has been widely used in the field of fault diagnosis of mechanical equipment due to its strong feature extraction and other advantages such as high efficiency, portability, and so on. However, at present, most kinds of intelligent fault diagnosis algorithms mainly focus on the diagnosis of a single fault component, and few intelligent diagnosis models can simultaneously carry out comprehensive fault diagnosis for a rotating system composed of a shaft, bearing, gear, and so on. In order to solve this problem, a novel stacked auto encoders sparse filter rotating component comprehensive diagnosis network (SAFC) was proposed to extract domain invariant features of various health conditions at different speeds. The model clusters domain invariant features at different speeds through the self-coding network, and then classifies fault types of various parts through sparse filtering. The SAFC model was validated by the vibration data collected, and the results show that this model has higher diagnostic performance than other models.