摘要
针对故障呈现渐发特性的复杂装备健康状态预测问题,提出基于多距离形态相似度评估(M-DSSE)和相关向量机(RVM)的预测方法。在提取装备状态特征信息的基础上,采用M-DSSE方法对装备的健康状态进行评估,计算得到装备的健康指数;运用RVM回归模型对装备的健康指数进行预测,实现对装备健康趋势的预知,为最终的预知维修提供重要技术支撑。在某航空机电设备上的应用结果表明,该方法可以有效解决复杂装备健康状态评估与预测的问题,结果与实际情况相吻合。
Aiming at the health state prognostics problem of complicated equipments, of which the fault shows gradual failure, a new prognostics method based on the multiple features extraction was presented compounding multiple distance and shape simi- larity （M-DSSE） and relevance vector machine （RVM）. The M-DSSE method was developed to evaluate the health index of equipments and the RVM regressive model was used to predict the performance trend based on the health index information, which greatly supported the predictive maintenance. The application on a certain aerial eleetromeehanieal device suggests the approach is capable to solve the health state prognostics problem of complicated equipments, and the results are consistent with the practical situation.
出处
《计算机工程与设计》
CSCD
北大核心
2014年第11期3997-4002,4039,共7页
Computer Engineering and Design
基金
总装预研基金项目(9140A27020212JB 14311)
关键词
健康指数
多距离形态相似度
相关向量机
状态评估
状态预测
health index
multiple distance and shape similarity （M-DSS）
relevance vector machine （RVM）
state evaluation
state prognostics
