基于模糊灰色关联分析的故障样本集评估方法

Evaluation method for failure sample set based on fuzzy grey relation analysis

为了在测试性验证试验之前获得理想的故障样本集,提出了一种基于模糊灰色关联分析的故障样本评估方法。首先确定了用于评估故障样本的指标,该指标有效地将样本量、样本结构和故障注入等多个特性指标综合考虑,利用模糊层次分析法确定各个指标的权重,然后通过灰色关联分析建立了故障样本集与故障模式集的距离模型,通过该距离模型来描述故障样本集对故障模式集的代表性,应用表明,该方法可以有效评估故障样本集,并可指导故障样本的选取。

An evaluation method for failure sample set based on fuzzy grey relation analysis is proposed in order to obtain perfect failure sample set before testability demonstration experiment. The method effectively fuses multiple factors such as sample quantity, sample structure, fault injection and etc. , and gives the evaluation requirement in- dex. The weight of each factor is obtained based on fuzzy analytic hierarchy process （ FAHP）. Then, the distance model between the failure sample set and the failure mode set is founded, and the distance model can be used to depict the distance between the failure sample set and the failure mode set. Finally, an application example proves that this method can be used to evaluate the failure sample set effeetively and direct the selection of failure sample set.