|
|||||
|
|
| 石钢高炉铁水含硅量神经网络预报模型 | |
| 引用本文： | 郭建斌,郭建国. 石钢高炉铁水含硅量神经网络预报模型[J]. 冶金丛刊, 2006, 0(5): 4-8 |
| 作者姓名： | 郭建斌 郭建国 |
| 作者单位： | 1. 石家庄钢铁有限责任公司 2. 河北经贸大学计算机系 |
| 摘 要： | 本文按照现代控制理论，把高炉视为多输入——单输出系统，利用人工神经网络方法，结合高炉生产实际建立了石钢高炉铁水含硅量BP神经网络模型。通过引入动态步长和“惯性项系数”提高了网络收敛速度。采用“修正式”预报模式提高了铁水含硅量预报的命中率。结果表明：在允许误差为0．1％时，命中率达到了86．67％，可以为高炉操作提供指导。 |
| 关 键 词： | 高炉 铁水含硅量 预报 神经网络 |
| 文章编号： | 1671-3818(2006)05-0004-04 |
NEURAL NETWORK PREDICTION MODELS FOR CONTENT OF Si IN SHIGANG BLAST-FURNACE HOT IRON |
|
| Guo jianbin,Guo jianguo. NEURAL NETWORK PREDICTION MODELS FOR CONTENT OF Si IN SHIGANG BLAST-FURNACE HOT IRON[J]. Metallurgical Collections, 2006, 0(5): 4-8 | |
| Authors: | Guo jianbin Guo jianguo |
| Abstract: | According to modern control theory,blast furnace was regarded as a multiple input single output system.Combined with the production experience,the BP neural networks were used to predict the content of Si in blast-furnace hot iron.The rapidity of convergence was improved with introducing dynamic step size and inertial coefficient,and the prediction precision was improved with introducing modified prediction model.The results showed that the prediction value was 86.67%under the permissible error 0.1%. |
| Keywords: | blast furnace content of Si in hot iron prediction neural network |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ | |