中文摘要：

针对大型设备故障诊断中特征量表现出来的不确定性和非精确性，提出了一种基于灰色关联理论来获取基本概率测度（BPA）的方法。该方法确定各故障特征量的参考样本，然后根据灰色相关性理论求得待诊断样本的相关系数，所取得的相关系数进行归一化处理即得到BPA值，经过Dempster多源组合规则进行多证据融合，得到最后的诊断结果。将所提方法用于电机转子故障诊断，实验结果表明该方法有效。

英文摘要：

To cope with uncertainty and inaccuracy in the fault diagnosis of large equipment, an approach based on grey correlation theory is proposed to obtain the basic probability assignment （BPA）. The method first determines reference samples of each fault mode. Correlation coefficients of samples are then obtained for diagnosis based on the grey correlation theory. After normalization, BPA of each sample is given. All evidences are combined to get the final diagnosis with multi-source Dempster rules. Experimental results show that the method is effective.