Back-propagation Based Contribution for nonlinear fault diagnosis

This paper proposes a novel fault diagnosis method by means of back-propagation based contribution (BBC) for nonlinear process. As a method based on the deep learning model, BBC can deal with the nonlinear problem in process monitoring by utilizing the nonlinear features extracted by auto-encoder (AE). Moreover, the smearing effect is an important factor affecting the performance of fault diagnosis. In order to solve this problem, BBC utilizes the basic idea of reconstruction based contribution (RBC), and describes the propagation of fault information by back-propagation (BP) algorithm. The validity of the proposed method is tested and verified by a nonlinear numerical example and the Tennessee Eastman benchmark process.