基于多传感器粒子权重优化的序贯最大似然比故障诊断算法  

Sequential maximum likelihood ratio test for fault diagnosis based on multi-sensor particle weight optimization

作　　者：胡振涛[1] 付春玲[2] 刘宇[1] 

机构地区：[1]河南大学图像处理与模式识别研究所,河南开封475004 [2]河南大学基础实验教学中心,河南开封475004

出　　处：《光电子．激光》2013年第8期1549-1556,共8页Journal of Optoelectronics·Laser

基　　金：国家自然科学基金(6092119);河南省教育厅自然科学基金(13A413066);河南省科技厅基础前沿(132300410148;132300410278)资助项目

摘　　要：针对多传感器量测下非线性系统故障诊断问题,给出了一种基于多传感器粒子权重优化的序贯最大似然比检验——故障诊断(SMLR-MWOPF)算法。首先,为改善量测随机噪声对粒子权重度量稳定性不良影响,结合多源信息融合技术实现对多传感器量测中冗余和互补信息的提取和利用,设计一种粒子权重优化策略;通过减小粒子权重方差以改善粒子权重的可靠性和稳定性,进而实现滤波器估计精度的提升。其次,结合序贯概率比检验(SPRT)和交互式多模型(IMM)框架构建了一种基于残差检测的在线序贯最大似然比检验方法;另外,在结构上合理简化输入交互和输出交互环节,以降低粒子滤波(PF)在IMM模型结构实现中的计算复杂度。理论分析和仿真实验验证了算法的可行性和有效性。Aiming at the nonlinear system fault diagnosis in multi-sensor observatlons,a novel sequennal maximum likelihood ratio test for fault diagnosis based on multi-sensor particle weight optimization is proposed. Firstly,to improve the adverse effects on the stability of particle weight measurement caused by random noise,the particle weight optimization strategy is designed by multi-source information fusion technology to fully extract and exploit the redundancy and complementary information from multi-sensor observations. Its implementation principle is to improve the reliability and stability of particle weight by the decline of particle weight variance, and tO promote the estimation precision. Secondly,combined with the sequential probability ratio test and the interacting multi-model, a novel online sequential maximum likelihood ratio method based on residual test is presented. In addition, considering the decline of compu- tation complexity in the combination process of particle filter and interacting multi-model,the input inter- action step and the output step are reasonably simplified in the construction of algorithm. The theoretical analysis and experimental results show the feasibility and efficiency of the proposed algorithm.

关 键 词：故障诊断 多源信息融合 似然比检验 粒子滤波(PF) 粒子权重优化 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]