本文提出了一种基于MBPLS故障检测方法,与其他的PLS不同的是,本方法重在结合设备的工作工况,从系统级、部件级进行剖析,深层次挖掘设备运行数据之间的关联关系,结合PLS算法优势,实现更优的设备故障检测。在本文中,先通过研究设备各作业...本文提出了一种基于MBPLS故障检测方法,与其他的PLS不同的是,本方法重在结合设备的工作工况,从系统级、部件级进行剖析,深层次挖掘设备运行数据之间的关联关系,结合PLS算法优势,实现更优的设备故障检测。在本文中,先通过研究设备各作业工况下相关变量关系,再对各变量进行分组,从系统角度和局部角度分别展开数据分析,可以提高故障检测效率。仿真结果表明,基于设备工况进行分块的MBPLS故障检测方法包含了更多的负载信息并且提高了诊断性能。In this paper, a PLS fault detection method based on working condition classification is proposed. Different from other PLS, this method focuses on the working condition of the equipment, analyzes from the system level and component level, deeply excavates the correlation between the operating data of the equipment, and combines the advantages of PLS algorithm to achieve better equipment fault detection. In this paper, the fault detection efficiency can be improved by first studying the relationship between the relevant variables under each operating condition of the equipment, and then grouping each variable to carry out data analysis from the system perspective and local perspective. The simulation results show that the MBPLS fault detection method based on equipment working conditions contains more load information and improves the diagnostic performance.展开更多
文摘本文提出了一种基于MBPLS故障检测方法,与其他的PLS不同的是,本方法重在结合设备的工作工况,从系统级、部件级进行剖析,深层次挖掘设备运行数据之间的关联关系,结合PLS算法优势,实现更优的设备故障检测。在本文中,先通过研究设备各作业工况下相关变量关系,再对各变量进行分组,从系统角度和局部角度分别展开数据分析,可以提高故障检测效率。仿真结果表明,基于设备工况进行分块的MBPLS故障检测方法包含了更多的负载信息并且提高了诊断性能。In this paper, a PLS fault detection method based on working condition classification is proposed. Different from other PLS, this method focuses on the working condition of the equipment, analyzes from the system level and component level, deeply excavates the correlation between the operating data of the equipment, and combines the advantages of PLS algorithm to achieve better equipment fault detection. In this paper, the fault detection efficiency can be improved by first studying the relationship between the relevant variables under each operating condition of the equipment, and then grouping each variable to carry out data analysis from the system perspective and local perspective. The simulation results show that the MBPLS fault detection method based on equipment working conditions contains more load information and improves the diagnostic performance.
