Fault Cell Mapping Analysis Methodology of Complex System

This methodology is to realize high real-time and high accuracy effect on complex system fault analysis. This methodology selects abnormal information as input space. As to input space, fault cells can be reached through fuzzy decomposition in two-dimensional space. As to output space, this paper introduces respectively two different methods. One method is the same as fuzzy decomposition of input space. The other is depended on fault type characteristics to partition artificially and experientially output space. Designing reasonably logic mapping deduction models, this paper presents fault cells-cells mapping and fault cells-space mapping analysis models. Taking transformer fault diagnosis as experimental example, it proves this methodology can be well in decreasing dramatically many useless interferential input variables, mining and explaining key characteristics and inner relation of abnormal information. The experimental result is also validated to be satisfied to solve the contradiction between high real-time and high accuracy.