一种摄动粒子滤波故障检测方法

发布时间：2018-03-10 09:30

本文选题：随机摄动　切入点：粒子滤波　出处：《电机与控制学报》2017年11期 　论文类型：期刊论文

【摘要】：粒子滤波算法在故障检测应用中面临的两个难题是退化问题和难以跟踪突变状态。针对上述问题,将随机摄动再采样方法和强跟踪滤波算法引入粒子滤波,提出了一种摄动粒子滤波故障检测方法,旨在于解决粒子滤波的退化问题并提高算法对突变状态的跟踪能力,从而提高故障检测方法对故障的检测准确度。通过强跟踪滤波更新粒子,来提高算法跟踪突变状态的能力;当出现退化现象时,采用随机摄动再采样方法,对粒子集中的最优粒子迭加一个随机摄动量,用摄动粒子替换粒子集中的退化粒子,解决退化问题。仿真结果显示该算法能及时、准确地检测系统故障。
[Abstract]:Two difficult problems in the application of particle filter in fault detection are the degradation problem and the difficulty in tracking the abrupt state. In view of the above problems, the random perturbation resampling method and the strong tracking filter algorithm are introduced into the particle filter. A perturbed particle filter fault detection method is proposed to solve the degradation problem of particle filter and to improve the ability of the algorithm to track the abrupt state. In order to improve the accuracy of fault detection, the ability of the algorithm to track the mutation state is improved by strong tracking filter, and the random perturbation resampling method is used in the event of degradation. A random perturbation momentum is added to the optimal particle in the particle set and the degenerate particle in the particle set is replaced by the perturbed particle. The simulation results show that the algorithm can detect the fault of the system in time and accurately.
【作者单位】：火箭军工程大学控制工程系;火箭军工程大学发射工程系;
【基金】：国家自然科学基金(61104223,61573366)
【分类号】：TN713

【相似文献】