基于最小二乘支持向量机的航煤干点软测量应用研究

传统支持向量机是近几年发展起来的一种基于统计学习理论的学习机器,在非线性函数回归估计方面有许多应用。最小二乘支持向量机用等式约束代替传统支持向量机方法中的不等式约束,利用求解一组线性方程得出对象模型,避免了求解二次规划问题。本文采用最小二乘支持向量机解决了航空煤油干点的在线估计问题,结果表明,最小二乘支持向量机学习速度快、精度高,是一种软测量建模的有效方法。在相同样本条件下,比RBF网络具有较好的模型逼近性和泛化性能,比传统支持向量机可节省大量的计算时间。

Dry point soft sensing application to aviation kerosene based on LS-SVM regression

this article is concerned with a classical problem of how to estimate the distillation composition on-line. A novel methodology based on LS-SVM (least squares support vector machine) regression is proposed to develop the distillation composition soft sensor. SVM (support vector machine) is a newly developed method based on statistical learning theory, and is widely used in non-linear function regression. By solving a set of linear equations instead of quadratic programming,LS-SVM has a better performance of non-linear modeling than SVM regression. The soft sensor in this paper is applied to predict the dry point of aviation kerosene in the distillation column. The simulation results show that LS-SVM has a good ability in non-linear modeling with high learning speed and accuracy. Using the same samples,our algorithm has better abilities of model approach and generalization than RBF neural networks method,and less running time than traditional SVM regression method.