论文部分内容阅读
为了定量分析故障分离的难易程度,提出基于改进距离相似度的故障可分离性分析方法。以基于等价空间的故障诊断残差产生器为例,根据不同故障引起残差向量在概率分布上的差异建立改进故障分离条件,利用残差向量的距离相似度和方向相似度评价故障分离的难易程度,提出基于改进距离相似度的故障可分离性定量评价指标。通过仿真试验分析某固定翼无人机纵向飞行控制系统的故障可分离性。结果表明:该方法能够准确判断故障分离条件,定量分析系统进行故障分离的难易程度,与现有方法相比,改进故障分离条件更为直观,基于改进距离相似度的故障可分离性评价指标能够实现残差向量在距离差异和方向差异的综合评价。
In order to quantitatively analyze the difficulty of fault separation, a method of fault separability analysis based on improved distance similarity is proposed. Taking the fault diagnosis residual generator based on equivalent space as an example, an improved fault separation condition is established according to the difference in probability distribution of the residual vector caused by different faults. The distance similarity and the direction similarity of the residual vector are used to evaluate the fault separation The degree of difficulty, the quantitative evaluation index of fault separability based on improved distance similarity is proposed. Through the simulation test analysis of a fixed-wing UAV longitudinal flight control system fault separability. The results show that this method can accurately determine the conditions of fault separation and quantitatively analyze the system ease of fault separation. Compared with the existing methods, this method can improve the fault separation conditions more intuitively. Based on the improved distance similarity, the fault separability evaluation index It is possible to achieve a comprehensive evaluation of the residual vector in the distance difference and the direction difference.