基于最小二乘支持向量机的电站锅炉吸热面积灰监测研究  

Research of Monitoring Ash Fouling on Utility Boiler Heat Transfer Surface Based on LSSVM

作　　者：刘正峰[1] 王景成[1] 史元浩[1] 

机构地区：[1]上海交通大学自动化系系统控制与信息处理教育部重点实验室,上海200240

出　　处：《化工自动化及仪表》2014年第3期262-266,共5页Control and Instruments in Chemical Industry

基　　金：国家自然科学基金资助项目(61174059;61233004;60934007);国家"973"计划项目(2013CB035406);上海市经信委重大技术装备研制专项资助项目(ZB-ZBYZ-01112634);上海市经信委引进技术与创新项目(12GA-31)

摘　　要：基于最小二乘支持向量机和粒子群算法,实现了对电站锅炉低温过热器灰污状态的在线监测。首先,定义清洁系数表征锅炉受热面的灰污状况,利用电站数据采集系统采集到机组实际运行数据,采用最小二乘支持向量机建立锅炉低温过热器系统的清洁吸热量模型;其次,选取机组正常运行工况下的数据,采用该模型预测出受热面的清洁吸热量;再次,根据热力学平衡计算出受热面的实际吸热量和清洁系数,从而实现了对锅炉受热面积灰状况的在线监测。以某电站300MW机组锅炉实际生产数据进行仿真验证,结果表明:所提方法能够对锅炉主要对流受热面的灰污状况进行有效的监测,模型预测精度较高、鲁棒性较强。Based on the least square support vector machine （LSSVM） and the particle swarm optimization （PSO）,the on-line monitoring of ash fouling on utility boiler＇s low-temperature superheater was implemented.Firstly,having the clean coefficient defined to characterize the ash fouling level,and the power station＇s data acquisition system （DAS） used to collect the operating data of the unit as well as the LSSVM applied to build the clean heat absorption model; then,having the thermal data of power plant selected to calculate the clean heat absorption capacity based on the proposed model; finally,according to the energy balance equation,having the real heat absorption capacity and the heat coefficient calculated to realize on-line monitoring of the ash fouling on boiler heat transfer surface.Verifying the proposed model with the operation data from a 300MW generating unit proves the effectiveness and high accuracy and strong robustness of the proposed PSOLSSVM model in predicting the clean heat absorption capacity.

关 键 词：电站锅炉 低温受热面 最小二乘支持向量机 粒子群算法 积灰监测 

分 类 号：TH865[机械工程—仪器科学与技术]