多参数相似性信息融合的剩余寿命预测

摘要

针对常规基于相似性信息的寿命预测方法的退化指标难以建立的问题,提出一种多参数相似性信息融合的剩余寿命预测方法.该方法并不进行退化指标的建立,而是利用单个原始参数数据直接进行基于相似性的寿命预测,然后将各参数对应的寿命信息进行融合得到设备的剩余寿命.为提高信息融合时的合理性,给出一种以相关性Spearman系数为基础的设备退化敏感性关键参数量化筛选及剩余寿命融合权重分配方法,结合各参数对应的剩余寿命信息,加权融合实现设备的剩余寿命预测.实验结果表明,相比于常规方法,所提方法在剩余寿命预测的准确性及改善预测精度方面更具优势.%In view of the problems of difficulty in constructing Degradation Index (DI) of similarity-based method under multiple degradation parameters,a similarity-based information fusion method for Remaining Useful Life (RUL)prediction was proposed.Without constructing degradation index,multiple degradation parameters were selected to predict RUL based on similarity directly,and the machinery's RUL was predicted by fusing RUL information.To improve the fusion accuracy,a spearman-based method was improved to select degradation sensitive parameter and assign weight,and machinery's RUL was weighted fused with RUL information of key parameters.Experiment result identified the effectiveness of the proposed method in predicting machinery's RUL and improving predicted accuracy.