Fault Diagnosis based on Domain Adaptive Multi-task learning Convolutional Neural Network

Abstract

In modern industry driven by machines, effective fault diagnosis of mechanical equipment is the basis for its healthy and safe operation. Current fault diagnosis methods are mostly based on the assumption of independent and identical distribution (IID). However, the characteristics of the same fault are different under different working conditions and not all fault data under all working conditions are available, so using the fault data obtained under certain working conditions to solve the fault diagnosis under all working conditions is of great significance to ensure the reliability and safety of the system. In order to solve these problems, a domain adaptive fault diagnosis method based on convolutional neural network (DAMCNN) is proposed in this paper which consists of three parts: encoder, decoder and tag generator. These three parts are combined to form the supervised and unsupervised learning parts of the model. Then, the multitask learning method is used to minimize the cost function of the source domain and the target domain. Finally, a model for data fault diagnosis in the target domain is obtained.