面向设备性能退化监测的分步参数估计策略OA北大核心CSTPCD
A stepwise parameter estimation strategy for equipment performance degradation monitoring
为通过预防性定期维护提升化工装置的精细化管理水平,探讨了关键设备的在线性能监测原理,针对设备性能参数可估计性差的问题,提出了一种基于机理模型分析设备性能退变的分步参数估计策略,根据灵敏度分析确定参数处理规则.首先以工业装置的严格机理模型为基础给出在线的粗粒度模型候选项,以历史数据拟合误差最小为目标完成初步参数估计,给出最优模型和各参数初值.然后基于灵敏度分析矩阵特征给模型参数排序,以表征性能退变的时变参数为界,前后不同顺位的定常参数采用不同策略依次处理.最后以滑动时间窗方式估计时变参数,实现设备性能监测.以反应釜夹套换热器的传热性能监测为例,按照分步策略获得设备性能参数的退变,剥离操作变量的影响,结果与工厂操作周期符合.表示该策略可以用于设备性能监测.
Periodic preventive maintenance is essential to chemical plant management and can be realized through online performance monitoring of key equipment.A stepwise parameter estimation strategy based on sensitivity analysis was proposed to solve the parameter estimability problem in the mechanism model analysis of equipment performance degradation.First,the online coarse-grained model candidates were given based on the rigorous mechanism model,and the optimal model and initial parameter values were obtained through preliminary parameter estimation using historical data fitting.Then,the parameters were sorted based on the sensitivity analysis,and the constant parameters of different orders defined by the time-varying parameters representing performance degradation were processed sequentially using different strategies.Finally,the time-varying parameters were estimated by sliding time windows to realize equipment performance monitoring.Taking the monitoring of the heat transfer performance of a jacketed heat exchanger in a reactor as an example,the degeneration of equipment performance parameters was obtained through the stepwise strategy,and the influence of operating variables was stripped away.The results were consistent with the factory operation cycle,indicating that the strategy could be used for performance monitoring.
沈雪婷;李红阳;陈伟锋;陶建国;祝铃钰
浙江工业大学 化学工程学院,浙江 杭州 310014浙江鸿盛化工有限公司,浙江 绍兴 312369浙江工业大学 信息工程学院,浙江 杭州 310023
化学工程
工业数据设备性能退化参数估计灵敏度矩阵
industrial dataequipment performance degradationparameter estimationsensitivity matrix
《高校化学工程学报》 2024 (002)
294-301 / 8
浙江省'尖兵''领雁'研发攻关计划(2024C01028);浙江省自然科学基金重点项目(LZ21B060001).
评论