基于多信号流图和相似性度量的故障可诊断性评价方法

Fault diagnosability evaluation method based on multi-signal flow graph and similarity measure

针对目前基于定性模型无法对电子系统故障可诊断性进行定量评价的问题,将定性模型与数据驱动方法相结合,提出一种基于多信号流图和相似性度量的故障可诊断性评价方法。首先,根据系统的组成结构建立多信号流图模型并得到故障-测试相关性矩阵,基于故障-测试相关性矩阵提出了故障可诊断性评价准则。其次,提取测试信号的小波包Shannon熵作为特征向量,以欧式距离作为相似性度量指标,将故障可诊断性定量评价问题转换为不同故障模式下测试信号特征向量的相似性度量问题。然后,构建故障可诊断性评价矩阵,依据故障可诊断性评价矩阵提出了系统可诊断性指标。最后,通过仿真分析验证了所提方法的有效性。结果表明,所提方法在不构建数学模型的前提下,能够实现电子系统的故障可诊断性的定量评价。

Aiming at the problem that the fault diagnosability of electronic system cannot be quantitatively evaluated based on qualitative model, the paper combines qualitative model with data-driven methods, and a method of fault diagnosability evaluation based on multi-signal flow graph and similarity measure is proposed. Firstly, the multi-signal flow graph model is established according to the composition of the system and the fault-test correlation matrix is obtained. Based on the fault-test correlation matrix, the fault diagnosability evaluation criteria is proposed. Secondly, the Shannon entropy of the wavelet packet of the test signal is extracted as the feature vector, and the Euclidean distance is used as the similarity measurement index. The problem of fault diagnosability quantitative evaluation is transformed into the similarity measure problem of the feature vector of the test signal under different fault modes. Then, the fault diagnosability evaluation matrix is constructed, and the system diagnosability indicators is proposed according to the fault diagnosability evaluation matrix. Finally, the effectiveness of the proposed method is verified by simulation analysis, and the results show that the method proposed in the paper can realize the quantitative evaluation of the fault diagnosability of electronic system without constructing mathematical model.