复杂动态系统的实际非完全失效故障的可诊断性评估

发布时间：2018-06-27 10:35

本文选题：动态系统 + 故障可诊断性　；参考：《自动化学报》2017年11期

【摘要】：针对缺乏有效的非完全失效故障(Loss of effect,LOE)可诊断性量化分析方法的现状,本文提出了一种基于距离相似度的系统非完全失效故障的实际可诊断性评价方法.通过将状态空间描述的动态系统转换为时间堆栈动态模型,使故障的可诊断性评估分析问题转化为多元分布的相似度问题.给出系统非完全失效故障可检测性与可隔离性的相关定义,并对故障的可诊断性进行量化.通过求取最小二乘解计算最小巴氏距离,增大了算法适用范围.最后,通过仿真实例验证评价方法的有效性,并通过所提出的可诊断性评估算法求取非完全失效故障的最大可诊断效能系数.
[Abstract]:In view of the lack of an effective quantifiable diagnostic analysis method for incomplete failure (loss of effect loe), a practical diagnostic evaluation method for incomplete failure of systems based on distance similarity is proposed in this paper. By transforming the dynamic system described by state space into a time stack dynamic model, the problem of fault diagnosability evaluation and analysis is transformed into the similarity problem of multivariate distribution. The definitions of detectability and isolability of incomplete failure faults are given, and the diagnosability of faults is quantified. The least square solution is used to calculate the minimum pasteurian distance, which increases the applicable range of the algorithm. Finally, the effectiveness of the evaluation method is verified by a simulation example, and the maximum diagnostic effectiveness coefficient of incomplete failure fault is obtained by the proposed diagnostic evaluation algorithm.
【作者单位】：北京控制工程研究所;北京空间飞行器总体设计部;空间智能控制技术重点实验室;
【基金】：国家自然科学基金项目(61690215,61640304) 国家杰出青年科学基金项目(61525301)资助~~
【分类号】：TP277

【相似文献】