| 基于神经网络的过程软测量 |
| |
| 引用本文： | 刘磊,王康斌.基于神经网络的过程软测量[J].天津大学学报(自然科学与工程技术版),1995,28(2):233-236. |
| |
| 作者姓名： | 刘磊 王康斌 |
| |
| 摘 要： | 以神经网络实现对某厂萃取精馏塔产品成分的实时估计，神经网络选用多层前馈网络，具有６个输入结点、１个输出结点、单隐层结构、１０个隐结点。所点数据全部取自生产现场，共取１８组样本数据，其中１５组用于网络的学习，３组用于网络的检验，结构分析表明，年建立的社会网络可以所映实际对象。
|
| 关 键 词： | 软测量 神经网络 数学模型 化工过程 精馏塔 |
SOFT SENSOR BASED ON THE NEURAL NETWORK IN A PROCESS ESTIMATION |
| |
| Liu Lei,Wang Kangbin.SOFT SENSOR BASED ON THE NEURAL NETWORK IN A PROCESS ESTIMATION[J].Journal of Tianjin University(Science and Technology),1995,28(2):233-236. |
| |
| Authors: | Liu Lei Wang Kangbin |
| |
| Institution: | Dept.of Chemical Eng. |
| |
| Abstract: | his paper presents a soft sensor based on the neural network in process estimation.The sensor is used to detect the product components in an extraction distillation columns in realtime.The neural network is a multilayer feedforward network with six input units,a single output unit and one hidden layer architecture that consists of ten hidden units.Eighteen sets of data are taken from the extraction process.Fifteen sets of them are used for training the network and the remaining three used for examining the network.The results illustrate that the neural network is able to discribe the real process very well. |
| |
| Keywords: | soft sensor neural networks mathematical models |
| 本文献已被 CNKI 维普 等数据库收录！ |
