摘要
为了解决稀土萃取分离过程元素组分含量在线检测的难题,提出了稀土萃取过程组分含量的一种最小二乘支持向量机(LS-SVM)软测量方法。利用量子粒子群算法来优化LS-SVM的参数及核函数参数。仿真结果表明,所提出的软测量方法是有效的,比已有的神经网络软测量方法能更好的实现稀土萃取过程中元素组分含量的在线估计。
In order to solve a difficult problem that was how to detect some important process variables online at real time, a method of soft sensor was proposed based on LS-SVM in this paper. A mixed kernel function was built, and the Quantum Particle Swarm Optimization （QPSO） algorithm was proposed to select the parameters of LS-SVM and the mixed kernel function. This method contributed to the distinct improvement of precision and generalization ability of the soft sensor model based on LS-SVM. The results showed that the proposed method had good approximation and well generalization ability. Compared with a method base on neural network, the method based on LS-SVM was more effective to realize online prediction of the component content in the rare earth extraction process.
出处
《中国稀土学报》
CAS
CSCD
北大核心
2009年第1期132-136,共5页
Journal of the Chinese Society of Rare Earths
基金
江苏省自然科学基金项目(BK2007210)资助
关键词
稀土萃取
软测量
最小二乘支持向量机
量子粒子群优化
rare earth extraction
soft-sensor
least squares support vector machines
quantum particle swarm optimization
