摘要
软测量技术是解决现代复杂工业过程中较难甚至无法由硬件在线测量参数的实时估计问题的有效手段。本文介绍了基于回归支持向量机(SVR)算法的基本原理,并以非线性、时变、大滞后的PTA氧化过程为研究对象,使用SVR算法对4-CBA含量进行了预测。结果表明,与传统预测方法相比,采用SVR算法的预测模型,具有精确度高,泛化能力强等优点,是用于PTA氧化过程中4-CBA含量预测的一种有效的方法,具有很好的应用价值。
In the modem complex industrial processes, some variables are very hard to be measured or even cannot be measured on-line by existing instruments and sensors. Soft-sensing technology can effectively solve the problem of realtime estimation. The basic principle of Support Vector Regression Machine （SVR） algorithm was introduced in this paper. For the process of PTA oxidation with non-linear, time-varying, large time delay characteristics, the 4-CBA Content was predicted by SVR. The predicted result was compared with that of traditional forecasting methods. The comparison result shows that the SVR has better integrative performance, high precision and generalization ability, so it is an effective method for being used in forecasting of 4-CBA content of PTA oxidation process and has a very good application value.
出处
《自动化博览》
2010年第3期76-78,共3页
Automation Panorama1
关键词
软测量
数学模型
4-CBA
回归支持向量机
Soft-sensing
Mathematical model
4-CBA
Support Vector Regression Machine
