Abstract: Product performance degradation assessment is the key stage of product health operation and residual life assessment.An assessment approach of product performance degradation based on outlier detection was proposed to achieve fast and real-time assessment.The correlation between performance monitoring parameters was determined based on the neighborhood rough set theory and functional and structural relevance.Then,the degradation index of performance parameters was constructed according to the knowledge contained in the correlations among them.Considering that the performance state representation of product function module was impacted by multiple performance parameters,the performance parameters were weighted by principal component analysis.The degradation score of a product function module was obtained by considering multiple performance parameters.Taking the wind turbine as an example,the effectiveness of the proposed approach was verified by comparative analysis.

Key words: outlier detection, neighborhood rough set, correlation of performance parameters, degradation assessment, mechanical product

摘要： 产品性能退化评估是产品健康运作和剩余寿命评估的关键步骤,为了实现快速、实时评估,提出一种基于离群点检测的产品性能退化评估方法。首先基于邻域粗糙集理论和功能结构关联性确定性能监测参数之间的相关关系;然后根据性能参数相关关系所包含的知识逐步构建性能监测参数的退化指标;由于产品功能模块的性能状态可能表征于多个性能参数,利用主成分分析法对性能监测参数赋权,综合考虑多个性能参数得到产品功能模块的退化得分;最后,以风力发电机组为例,通过对比分析验证了所提方法的有效性。

关键词: 离群点检测, 邻域粗糙集, 性能参数相关关系, 退化评估, 机械产品