基于线性规划和神经网络的优化烧结配料系统开发  

A SINTER FEED PROPORTION OPTIMUM SYSTEM BASED ON LINEAR PROGRAMMING AND NEURAL NETWORK

作　　者：王炜[1] 陈畏林[2] 贾斌[2] 徐智慧[2] 杨海林[2] 

机构地区：[1]湖北省钢铁冶金重点实验室(武汉科技大学) [2]武汉钢铁集团公司

出　　处：《烧结球团》2006年第1期27-30,共4页Sintering and Pelletizing

摘　　要：合理经济地确定烧结配料问题,具有十分重要的意义。本文从技术要求和经济效益两方面综合考虑烧结配矿的最佳经济性问题。根据线性规划法的思路,以生产实践为基础,建立目标函数,确定约束条件,对烧结所用铁矿粉进行优化配料。同时,利用神经网络模型预测烧结矿的性能指标,依据其结果调整线性规划的约束条件,从而得到满足烧结矿质量要求的最佳配矿比。in the sintering process, it is essential that sinter is established economically and reasonably. In this paper, the best economical condition for the charge proportion for sintering is comprehensively considered from technologic demand and economic efficiency. According to LP and based on practice and trial, aimed function and confined condition is defined, computation programs are made to get optimum charge proportion. A back propagation neural network was also carried out to predict the tinter quality and adjust the confined condition of the LP model. An optimum charge proportion can be find out and the quality of the tinter also be assured by the model.

关 键 词：烧结 配料 神经网络 线性规划 

分 类 号：TF046.4[冶金工程—冶金物理化学] P631.325[天文地球—地质矿产勘探]