引用本文:刘瑞兰, 牟盛静, 苏宏业, 褚健.基于支持向量机和粒子群算法的软测量建模[J].控制理论与应用,2006,23(6):895~899.[点击复制]
LIU Rui-lan, MU Sheng-jing, SU Hong-ye, CHU Jian.Modeling soft sensor based on support vector machine and particle swarm optimization algorithms[J].Control Theory and Technology,2006,23(6):895~899.[点击复制]
基于支持向量机和粒子群算法的软测量建模
Modeling soft sensor based on support vector machine and particle swarm optimization algorithms
摘要点击 1512  全文点击 1781  投稿时间：2005-04-19  修订日期：2005-12-12
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  
  2006,23(6):895-899
中文关键词  支持向量机  特征样本  粒子群优化算法  PTA氧化过程  软测量
英文关键词  support vector machine  feature subset  particle swarm optimization algorithm  PTA(purified terephthalic acid)oxidation process  soft sensor
基金项目  国家863计划资助项目(2001AA413020); 国家杰出青年科学基金资助项目(60025308).
作者单位
刘瑞兰, 牟盛静, 苏宏业, 褚健 南京邮电大学自动化学院, 江苏南京210003
新加坡高性能计算研究所, 新加坡117528
浙江大学先进控制研究所, 工业控制技术国家重点实验室, 浙江杭州310027 
中文摘要
      针对PX氧化过程中的4-CBA浓度的估计问题, 提出了基于支持向量机和粒子群算法来估计机理模型参数的方法. 用支持向量机回归来提取特征样本, 这些少量的特征样本估计机理模型参数可以减少计算时间, 同时避免了人工随机试凑法选择训练样本的盲目性. 采用粒子群算法来估计非线性机理模型的参数, 可以避免传统方法对初始点和样本的依赖. 工业实例表明, 本文提出的方法是有效的．
英文摘要
      The estimation of 4-CBA (carboxybenzaldchydc) concentration in industrial PTA (purified terephthalic acid) oxidation process is of fundamental importance in process monitoring, advanced control and optimization. The support vector machine (SVM) and particle swarm optimization (PSO) algorithms are used to estimate the parameters of the first principle model. The training set for estimating the parameters is the feature subset selected by SVM regression algorithm, which overcomes the drawback of the trail-and-error method. Parameter estimation method based on the PSO algorithm is also used to avoid dependence on initial parameters and training samples. By use of real industrial data, the simulation results show that the presented method is effective for modeling the soft sensor of 4-CBA concentration in industrial PTA oxidation process.