基于多元分析和模糊规则的系统特征冗余度排序算法

Specification Redundancy Ordering Algorithm Based on Multivariable Statistics and Fuzzy Rules

针对参数性故障分布具有很强随机性的特点 ,将特征观测值看作随机变量的取样 ,引入多元统计理论的分析方法 ,对多种状态下的特征样值进行分析筛选 ,进而对初始的检测特征集合进行冗余度排序 ,可以大大压缩特征参数数量 ,减少故障诊断的时间 .综合时间效率和灵敏度两个参考指标 ,运用模糊推理的方法对检测特征进行了总排序 ,并采用算法对一个实用电路进行了分析 .实验结果表明 。

The specification diagnosis dominates the analog fault diagnosis, that is, inner faulty condition can be shown by output characteristic of the circuit. But the method needs high requirements to the selection of the specifications. In this paper, an algorithm of specification ordering based on the multivariable statistics and fuzzy rules was presented. Considering the time-consumption and sensitivities, this paper uses the way of fuzzy rules to give an order to the original specification set. The simulation shows that this algorithm is effective in analysis of circuits.