基于衰减记忆思想的故障容错方法  

A fault-tolerant method based on fading memory idea

作　　者：高超 赵修斌[1] 庞春雷[1] 王勇[1] 张闯[1] GAO Chao;ZHAO Xiubin;PANG Chunlei;WANG Yong;ZHANG Chuang(Information and Navigation College,Air Force Engineering University,Xi'an 710077,China)

机构地区：[1]空军工程大学信息与导航学院

出　　处：《测绘通报》2019年第5期93-97,共5页Bulletin of Surveying and Mapping

基　　金：国家自然科学基金(61601506)

摘　　要：残差χ2法难以准确检测小幅值突变故障,改进的序贯概率比检测(SPRT)方法通过迭代递推,能够检测到故障,但存在延时且对故障结束不敏感。基于此,本文给出了区分故障大小的方法,针对小幅值故障提出了一种基于衰减记忆思想的故障容错方法:当改进的SPRT方法检测到故障后,对系统进行隔离重构,用不包含故障信息的数据进行量测更新,利用衰减因数减小包含故障信息的陈旧滤波值的权重,以提高滤波精度,同时构造"伪正常"状态,使该方法能够及时判断故障结束。仿真结果表明,本文提出的方法适用于故障多发的情形,能够提高滤波精度,增强系统可靠性。The residual chi-square test method is difficult to detect and isolate accurately,the improved SPRT( sequential probability ratio test) method can detect the fault by iterative recurrence,but there is a delay and it is not sensitive to the end of the fault. Based on this,this paper analyzes the method of differentiating the size of the fault amplitude,and proposes a fault-tolerant method based on the fading memory idea. When the improved SPRT method detects the fault,the system is isolated and reconstructed. The filter is updated with the data that does not contain the fault information,and the fading factor is used to reduce the weight of the fault information so as to improve the filtering accuracy. Meanwhile,constructing the pseudo normal state,so that the method can judge the end of the fault in time. Simulation results show that the proposed method is suitable for multiple faults,which can improve filtering accuracy and enhance system reliability.

关 键 词：故障幅值 衰减记忆 “伪正常”状态 故障结束判定 

分 类 号：P228[天文地球—大地测量学与测量工程]