改进的量测一致性的联邦滤波器两级故障检测

The Improvement Two Levels of Fault Detection Algorithm Based on the Consistency between the Measurement of Federal Filter

当联邦滤波器故障为缓变故障或者故障幅值比较小时,传统的基于量测一致性的联邦滤波器故障检测算法由于相应的估计误差和方差同时增大,这使得算法需一定的时间后才能使这种增量大到使故障函数计算结果超出门限,因而会出现警告延迟甚至漏检现象。针对量测一致性算法的缺点,提出了改进的基于量测一致性的联邦滤波器两级故障检测算法。相比传统的基于量测一致性的故障检测算法,该方法增加了故障检测的冗余性,不仅可以区分硬故障和软故障,而且提高了故障检测的可靠性和灵敏度,使得联邦滤波器的故障检测方法更加成熟完善,这对提高整个系统的可靠性具有重要意义。仿真和实验结果表明,该方法准确度高,计算量小,便于工程实现。

On the condition that the federal filter fault for slow varying fault or the fault magnitude is small,due to its corresponding estimation error and variance increased at the same time when using the traditional measurement based on the consistency between the measurement of federated filter for fault detection algorithm,it takes a certain amount of time for algorithm to make this volume increased to reach the fault function results beyond the threshold.Therefore,the alarm will delay even missed.According to the demerit of federated filter,this paper puts forward the improvement two levels of fault detection algorithm based on the consistency between the measurements of federal filter.Compared with the traditional algorithm based on the consistency between the measurements of federal filter,this algorithm increases the redundancy of fault detection.Besides this algorithm not only can distinguish between the hard fault and soft fault,but also could able to improve the reliability and sensitivity of fault detection,which could make fault detection method for federal is more mature and perfect and has a significant influence on improving the reliability of the whole system.Simulation results show this algorithm has features that high accuracy,small amount of calculation,and easing to be realized in engineering.