基于广义回归神经网络的传感器故障诊断研究

Research on GRNN-based sensor fault diagnosis

针对诊断传感器偏置故障与漂移故障的难点问题,提出了一种基于广义回归神经网络(GRNN)的传感器故障诊断方法。该方法充分利用控制系统闭环回路测控信息,建立一组多输入单输出GRNN观测器,通过将观测器输出与传感器实际输出相比较获取残差序列,获得基于残差序列的传感器偏置故障和漂移故障的辨识策略,实现控制系统传感器故障在线诊断。仿真结果表明:该方法可以快速准确地检测和分离传感器故障,辨识传感器故障类型、故障大小以及故障发生的时间。

Aimed at solving the challenging problem of diagnosis for sensor bias and drift faults, a novel approach of sensor fault diagnosis based on generalized regression neural network （GRNN） is proposed. The method makes full use of the closed-loop monitoring information in control system to establish multi-input and single-output GRNN observers. By comparing the outputs of the GRNN observers and the actual values of sensors, the identification strategy based on the sequence of remain residuals for sensors bias fault and drift fault is acquired and sensors on-line fault diagnosis in control system is carried out. The simulation results indicate that the approach can quickly and accurately detect, isolate and identify fault.