基于支持向量机软测量技术的应用

Application of Soft-sensing Technology based on Support Vector Machine

软测量技术在工业过程控制中得到广泛的应用。在软测量建模过程中,基于支持向量机的算法能较好地解决小样本、非线性、高维数、局部极小点等问题。在简单介绍最小二乘支持向量机算法的基础上,提出了一种新的改进算法———多输入多输出最小二乘支持向量机算法,将其应用到丙烯腈收率的预测模型中,并且与传统的神经网络算法以及多输入单输出最小二乘支持向量机算法进行建模比较。结果表明,这种算法可以在付出轻微代价的基础上,实现多输入多输出模型的软测量,并取得良好的效果。

The technology of soft-sensing has been widely used in industrial process control. In model establishment of soft-sensing, the prob- lems of small sample, non-linearity, high dimensions and local minimal value can be well solved by support vector machine algorithm. On the basis of least square support vector machine algorithm, a new improved algorithm of multi- in multi- out LS- SVM algorithm is proposed and used in prediction model of acrylonitrile output rate. In addition, this method is compared to traditional network algorithm and multi-in, single -out LS-SVM algorithm. The results show that based on low cost, excellent effects are achieved for implementing seft-sensing of multi-in multi-out model.