Intelligent diagnosis of front-end redundancy for a control system based on physical correlation process

作者：Chen, Yongwei^*; Xie, Yongjing; Zhou, Xinxing; Li, Yonggang

来源：Annals of Nuclear Energy, 2021, 154: 108115.

DOI：10.1016/j.anucene.2020.108115

• Control system
• Fault diagnosis
• Neural network
• Deviation judgment
• Physical correlation

The front-end sensors of the control system are the weak links for a reliable and stable operation of system. Due to insufficient capabilities of the algorithms or methods, there are many control anomalies caused by the failure of the front-end sensors. This paper proposes a front-end redundancy intelligent diagnosis model for control systems, which mainly includes five sub-models: transfinite judgment model, wavelet transform diagnosis model, deviation operation judgment model, process variable neural network learning model, and fault output selection model. The proposed solution can realize the fault diagnosis for the redundant sensors installed on the front end of the control system and thus prevents the abnormal signals of the front end from being input into the control system that usually leads the disturbance within the process system. The work has been carried out using simulation and real working condition approach in order to validate the proposed solution.