基于未知输入观测器的故障检测诊断降虚警研究

Decreasing False Alarm of Fault Detection and Diagnosis Based on Unknown Input Observer

对基于模型的故障检测与诊断中的虚警问题进行了研究,系统地分析了其发生机理,给出了一种基于未知输入观测器生成残差的方法,该方法能够在确保故障检测率要求的同时,使虚警率最小,仿真结果验证了方法的有效性。

The problem of false alarm in the fault detection and diagnosis based on model is researched.The mechanism of the arising of false alarm is analyzed,a kind of residual formulation method,based on the UIO(unknown input observer),is proposed.It is verified that the method can make the false alarm ratio(FAR) minimal,as well as ensure the false diagnosis rate(FDR).Mathematical simulation results show that the way is effective.