| RBF多模型神经网络软测量技术在湿法磷酸生产中的应用 |
| |
| 引用本文： | 阚晓旭,金晓明. RBF多模型神经网络软测量技术在湿法磷酸生产中的应用[J]. 化工自动化及仪表, 2006, 33(1): 64-66,81 |
| |
| 作者姓名： | 阚晓旭 金晓明 |
| |
| 作者单位： | 浙江大学,先进控制研究所,工业控制技术国家重点实验室,杭州,310027;浙江大学,先进控制研究所,工业控制技术国家重点实验室,杭州,310027 |
| |
| 摘 要： | 利用多模型的思想进行二水法磷酸装置反应槽SO3浓度软测量建模.首先通过机理分析选出对目标变量SO3浓度影响较大的变量作为辅助变量,并利用神经网络分类思想按目标变量值的不同区间对现场数据进行分类,然后采用多个径向基神经网络建立相应的经验模型.工业装置数据的拟合和预测结果表明:基于神经网络的多模型软测量可以取得更好的效果.
|
| 关 键 词： | 多模型 RBF 神经网络 SO3 湿法磷酸 |
| 文章编号： | 1000-3932（2006）01-0064-03 |
| 收稿时间： | 2005-12-15 |
| 修稿时间： | 2005-12-15 |
Application of Multi-modeling Neural Network Soft-sensing Technique in the Phosphoric Acid by Wet Process |
| |
| KAN Xiao-xu,JIN Xiao-ming. Application of Multi-modeling Neural Network Soft-sensing Technique in the Phosphoric Acid by Wet Process[J]. Control and Instruments In Chemical Industry, 2006, 33(1): 64-66,81 |
| |
| Authors: | KAN Xiao-xu JIN Xiao-ming |
| |
| Affiliation: | National Lab of Industrial Process Control, Institute of Advance Process Control, Zhejiang University, Hangzhou 310027, China |
| |
| Abstract: | The method of multi-modeling is used to build a soft sensor of the SO3 in the phosphoric acid by wet process Firstly, based on the mechanism analysis, the variables which have more influence in object variable SO3 concentration are selected for the auxiliary variables. Then,using the idea of neural network classification, the field data are classified according to the different span of the object variables. Lastly, the RBF neural networks are used to build corresponding empirical models for each classification. The simulation results with industrial data show the soft sensor is satisfied. |
| |
| Keywords: | multi-modeling RBF neural network SO3 phosphoric acid by wet process |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
