基于改进的GA-LSSVM的软测量建模方法  

Soft Sensing Based on Improved GA-LSSVM

作　　者：王强[1,2] 田学民[1] 

机构地区：[1]中国石油大学(华东)信息与控制工程学院,山东青岛266580 [2]东营职业学院工业工程系,山东东营257091

出　　处：《北京理工大学学报》2012年第10期1031-1035,共5页Transactions of Beijing Institute of Technology

基　　金：国家自然科学基金资助项目(51104175);山东省基金资助项目(ZR2011FM014)

摘　　要：针对工业过程中某些重要过程变量难以实现在线测量的问题,提出了一种改进的最小二乘支持向量机(IGA-LSSVM)的软测量建模方法.该方法采用核独立分量分析(KICA)对高维数据进行特征提取,利用改进的最小二乘支持向量机进行建模.该方法既利用了最小二乘支持向量机求解速度快的特点,又利用了自适应遗传算法强大的全局搜索能力,增强了模型的自适应性.用该方法建立柴油凝点的软测量模型,结果表明,基于IGA-LSSVM方法建立的软测量模型具有较高的预测精度和泛化能力.Some variables in industrial process are very difficult to measure on-line. To overcome this problem, a kind of soft sensing based on improved GA-LSSVM （IGA-LSSVM） is proposed in this research. First, KICA was used to extract main features of the data with high dimension patterns, and then an improved GA-LSSVM was established. This model not only utilizes the ability of quickly solving speed of LSSVM, but also the powerful global search performance of adaptive GA. Therefore the adaptability of LSSVM model is improved. The proposed method has been used for building soft sensing of diesel oil solidifying point. The result shows that IGA- LSSVM approach has high precision and good generalization ability.

关 键 词：软测量 核独立元分析(KICA) 遗传算法(GA) 最小二乘支持向量机(LSSVM) 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]