基于双权值神经网络的煤气炉系统优化

A Gas Furnace Optimized Application Based on Double-weight Neural Network

一般的化工过程的输入输出系统具有高度的非线性特性,双权值神经网络具有较强的逼近与泛化能力,用工厂的实测数据,用双权值神经网络很好地逼近了一个煤气炉系统。利用遗传算法全局寻优的特点,找到了系统的若干组可行输入,并根据能耗评估,得到系统的最优输入。利用双权值神经网络及遗传算法为系统优化提供了一套解决方案。

Input-output system of gas furnace for the chemical system has highly nonlinear characteristics. Double-weight neural network has strong approximation and generalization ability. With double-weight neural network approximation to the gas system,use global optimization characteristics of the genetic algorithm（ GA） to seek the feasible input of the system,and according to the input energy evaluation,a solution for factory to reduce energy consumption has been offered by getting the optimal system input.