锅炉水总碱度的软测量方法  

Prediction of Total Alkalinity in Boiler Water Based on Soft Sensor

作　　者：吉训生 郁燕东 袁彪[2] 邓宏康[2] 

机构地区：[1]江南大学轻工过程先进控制教育部重点实验室,江苏无锡214122 [2]江苏省特种设备安全监督检验研究院,江苏南京210006

出　　处：《仪表技术与传感器》2013年第4期68-73,共6页Instrument Technique and Sensor

基　　金：江苏省科技支撑项目(BE2010748)

摘　　要：锅炉是工业企业最重要的热交换装置,通过对锅炉水总碱度的监测与预测,可以改变锅炉水质的结垢倾向,减少锅炉事故的发生,对提高企业的生产效率有着重要的意义。软测量技术广泛应用于工业过程,其核心是建立一个可靠的软测量模型。为了实现对锅炉水总碱度的预测,基于FIA系统测定的混合碱溶液的浓度和电压值,文章提出使用PLS和LSSVM的软测量技术来预测锅炉水中NaOH和Na2CO3的浓度,比较二者在预测精度和建模时间上的优劣,在此基础上建立PLS-LSSVM模型,试验和仿真结果表明PLS-LSSVM模型的预测方法将NaOH的平均相对误差从17.48%降低到了4.54%,而Na2CO3的预测平均相对误差从17.52%降低到了3.64%,其预测效果可以更好地满足工业现场的需求。The boiler has been the most important heat exchanger in industrial enterprises.We can change the scaling tendency and reduce the occurrence of the boiler accident by monitoring and foresting the total alkalinity of the boiler water.It is significant to improve the production efficiency for the enterprises.Soft sensors have been widely used in industrial process,the core of which is the establishment of a reliable soft sensor model.In order to achieve this goal,based on the concentration of the mixture and voltage value tested by FIA system,the PLS and LS-SVM models were established to predict the concentration of NaOH and Na2CO3.Finally,with the comparison of both methods from the modeling accuracy and time,a forecasting modeling method based on PLS-LSSVM approach was put forward.The simulation and experiment results show that the average relative error of the predictions by PLS-LSSVM method of NaOH decreased from 17.48% to 4.54%,while the average relative error of Na2CO3 reduced from 17.52% to 3.64%.The prediction can better meet the needs of industry sites.

关 键 词：锅炉水总碱度 偏最小二乘 最小二乘支持向量机 软测量 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]