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.

控制系统的前端传感器是系统可靠，稳定运行的薄弱环节。由于算法或方法的能力不足，前端传感器的故障会导致许多控制异常。本文提出了一种控制系统的前端冗余智能诊断模型，主要包括五个子模型：超限判断模型，小波变换诊断模型，偏差操作判断模型，过程变量神经网络学习模型和故障输出选择模型。所提出的解决方案可以实现对安装在控制系统前端的冗余传感器的故障诊断，从而防止前端的异常信号输入到控制系统中，该信号通常会导致过程系统内的干扰。