Abstract:

The fault diagnosability evaluation can quantify the difficulty of diagnosing faults. A temporal distance measurement method based on dynamic time warping (DTW) is proposed to solve the problems of unobjectivity and inappropriate selection of measurement methods in the fault diagnosability evaluation methods. A diagnosability evaluation model is defined, and four elements of system structure, signal, test and failure mode can be extracted from an actual system, which lays the foundation for the subsequent diagnosability evaluation work. The traditional method of extracting the features from signals for similarity measure is abandoned, and the signals are regarded as time series. Based on the DTW method, the degree of similarity between signals in different states is measured as the evaluation basis for fault diagnosability. Momentum wheel system is used to simulate the proposed method. The experimental results show that the proposed method is objective and effective in diagnostic evaluation.

Key words: fault diagnosability, quantitative evaluation, dynamic time warping, similarity measure, time series