文献类型：期刊文献

中文题名：基于数据驱动的大型化工装置健康评估与故障源诊断

英文题名：Healthy evaluation and fault source diagnosis for process equipment systems based on data-driven

作者：曹稼斌[1];郑建荣[1];夏春明[1];张亮[1]

机构：[1]华东理工大学机械与动力工程学院,上海200237

年份：2013

卷号：19

期号：5

起止页码：1085

中文期刊名：计算机集成制造系统

外文期刊名：Computer Integrated Manufacturing Systems

收录：CSTPCD;;EI(收录号：20132916508673);Scopus;北大核心:【北大核心2011】；CSCD:【CSCD2013_2014】；

基金：国家自然科学基金资助项目(60974061);上海市引进技术吸收与创新计划资助项目(产-30-方向-28)~~

语种：中文

中文关键词：过程装置;健康评估;控制回路;控制阀非线性故障;故障源定位;氯乙烯装置

外文关键词：process equipment; health evaluation; control loop; nonlinearity measures for control valve; fault source localization; vinyl chloride production plant

摘要：针对大型在役化工过程装置亚健康问题,提出一套基于实际操作过程参数及其内在联系分析,进行大型在役化工过程装置定量健康评估与故障源诊断的方法。该方法依据最小方差控制方法评估控制回路性能,基于双相干频谱非线性分析评估控制阀的性能,实现装置健康状态定量评估;采用功率谱独立源成分进行系统波动源分析,进而通过格兰杰因果检验确定波动故障传播路径和根源。给出有关理论和方法,并据此对大型氯乙烯装置进行健康评估和故障源定位,获得装置性能定量评估结果,确定异常控制回路、故障源及其传播路径。现场验证结果表明,评估结果能够较好地反映装置的实际情况。
Aiming at the subhealthy of large in-service chemical process equipment,a method of healthy quantitative evaluation and fault source diagnosis for large process equipment systems was proposed based on practical operation process parameters and their inherent relationship analysis.The performances of control loop were evaluated by using minimum variance control,and performance of control valves were assessed by using bicoherence-based nonlinearity measures so as to realize healthy quantitative evaluation for equipment.Fault separation and oscillatory source localization were evaluated by special independent component analysis,and oscillation propagation analysis was defined based on Granger causality test.The theories and methods were introduced and used to evaluate a large-scale vinyl chloride production plant.The results of quantitative evaluation for the plant loop were obtained,and the oscillation disturbance and propagation were determined.The field survey demonstrated that the results could reflect practical situations of the plant.