石灰回转窑煅烧带温度的软测量方法  

Soft-sensing method for burning zone temperature in lime rotary kiln

作　　者：田中大[1] 张月 毛程程 张超[1] 

机构地区：[1]沈阳工业大学信息科学与工程学院

出　　处：《电子测量与仪器学报》2018年第1期166-173,共8页Journal of Electronic Measurement and Instrumentation

基　　金：辽宁省自然科学基金(20170540686);辽宁省教育厅科学研究项目(LGD2016009)资助

摘　　要：作为石灰回转窑生产过程中的重要参数,煅烧带温度对于石灰生产质量有着极大的影响。由于回转窑结构的特殊性,且煅烧带温度极高,煅烧带温度很难利用传感器直接测量。针对石灰回转窑煅烧带温度的测量问题,将石灰回转窑煅烧带温度作为研究对象,提出了一种软测量方法。利用生产现场易测的数据建立回转窑煅烧带温度的软测量模型。首先引入核主成分分析对采集的可测数据进行主成分的提取,减少预测模型输入变量之间的耦合与相互干扰,同时减少了建模的复杂度。然后通过最小二乘支持向量机对石灰回转窑煅烧带温度建模。最小二乘支持向量机预测性能与模型参数有着很大的关系,为了提高预测精度,利用具有良好优化性能的和声搜索算法对最小二乘支持向量机预测模型中的参数进行优化。仿真实验结果表明,提出的回转窑煅烧带温度软测量方法具有较高的预测精度,减少了预测误差,预测值较好的反映了回转窑煅烧带温度的变化趋势。同时,减少了建模的复杂度。提出的石灰回转窑煅烧带温度软测量方法是有效的。As an important parameter in the production process of lime rotary kiln,the burning zone temperature has a great influence on the quality of lime production. As a result of the special structure of the rotary kiln and the high temperature of the burning zone,it is difficult to measure the temperature of the burning zone directly by sensor. The burning zone temperature of lime rotary kiln is studied,and a soft-sensing method is proposed. This soft-sensing method of burning zone temperature in lime rotary kiln is established by using the data measured at the production field. Firstly,kernel principal component analysis is introduced and data of principal component is extracted. The coupling and mutual interference between the input variables of the prediction model,the complexity of the modeling are reduced. Then,the burning zone temperature prediction model is built by least squares support vector machine algorithm. The prediction performance of least squares support vector machine is closely related to the model parameters. In order to improve the prediction accuracy,the harmony search algorithm with good optimization performance is used to optimize the parameters of least squares support vector machine prediction model. The simulation results show that the soft-sensing method of burning zone temperature in the lime rotary kiln has a higher prediction accuracy and less prediction error. The prediction value reflects the trend of the burning zone temperature of the rotary kiln. At the same time,the complexity of modeling is reduced. The proposed soft-sensing method of burning zone temperature in lime rotary kiln is effective.

关 键 词：石灰回转窑 煅烧带温度 核主成分分析 最小二乘支持向量机 和声搜索算法 软测量 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TQ177.26[自动化与计算机技术—控制科学与工程]