摘要
研究使用BP神经网络来完成化工过程的建模。一般的化工过程建模问题是非线性的问题。很多情形下 ,与不同相有关量的变化速率数量级相差很大 ,从而问题是刚性的。这时 ,用数值方法难于对过程的变化精确求解。由于BP网络能实现任何非线性的连续映照 ,故适于处理复杂化工建模问题。将BP神经网络用于精馏塔的温度计算 ,结果令人满意。
A study on the modeling of chemical processes with BP neural networks is presented.Usually,the modeling problem of a chemical process is a nonlinear problem.In many cases,the quantitative differences between the changing rates of variables related to different phases are quite great,thus the problems have a high stiffness,and it is difficult to solve the modeling problem accurately with the common digital calculation methods.It is showed that BP neural networks could perform any continuous nonlinear mapping.Hence they are suitable for handling complicated dynamic problem.In the paper a BP neural network is used to calculate the temperatures of the output flows of a fractionation tower.The results were satisfactory.
出处
《江苏石油化工学院学报》
2000年第2期45-47,共3页
Journal of Jiangsu Institute of Petrochemical Technology
关键词
BP神经网络
化工过程
建模
BP neural networks
chemical process
modeling
