基于参考模型的线性时不变系统的故障诊断

摘要

Considering the idea of reference model adaptation, a new method for fault detection and isolation based on reference model is proposed. It firstly obtains system states through unknown input observer (UIO), then generates residual signal using the observing states and the states of reference model, finally estimates the fault signal and the formulation of estimating error is given. The simulation result indicates its validity.%首次将模型参考自适应思想引入控制系统的故障诊断，提出了一种基于参考模型的新型故障诊断方法。该方法首先利用带有未知输入状态观测器观测出系统状态，然后通过观测状态和参考模型的状态构造出残差信号，再利用最小二乘法估计出故障的幅值大小，并给出了估计误差的计算公式。仿真结果表明了该方法的有效性。