您的位置:首页  > 论文页面

基于主成分分析(PCA)与GA-BP结合的转炉炼钢终点锰含量预测研究

发表时间:2019-04-30  浏览量:3491  下载量:747
全部作者: 张壮,曹玲玲,林文辉,孙建坤,冯小明,刘青
作者单位: 北京科技大学钢铁冶金新技术国家重点实验室;新余钢铁集团有限公司
摘 要: 为提高转炉炼钢终点锰含量预测模型的预测精度,研究分析了影响转炉炼钢终点锰含量的因素,提出了将主成分分析(principal component analysis,PCA)与遗传算法(genetic algorithm,GA)和BP(back propagation)神经网络结合的数据驱动建模方法。使用PCA法对多个影响终点锰含量的因素进行降维处理,将处理后所得的主成分变量输入GA-BP模型进行训练而得到转炉炼钢终点锰含量预测模型。通过将PCA-GA-BP模型预测结果与GA-BP神经网络模型预测结果相比较,结果显示,基于PCA的GA-BP模型的预测精度较高,泛化性能好,预测误差在±0.025%范围内的命中率达到86%,均方误差为2.74×10^-8,且模型的训练速度有了显著提升。
关 键 词: 钢铁冶金;转炉;终点锰含量;BP神经网络;遗传算法;主成分分析
Title: Prediction study of end-point manganese content for BOF steelmaking process based on principal component analysis (PCA) and GA-BP
Author: ZHANG Zhuang, CAO Lingling, LIN Wenhui, SUN Jiankun, FENG Xiaoming, LIU Qing
Organization: State Key Laboratory of Advanced Metallurgy, University of Science and Technology Beijing; Xinyu Iron & Steel Co. Ltd.
Abstract: In order to improve the prediction accuracy of the prediction model of end-point manganese content for BOF steelmaking process, a data-driven modeling method based on the combination of principal component analysis (PCA) and genetic algorithm-black propagation (GA-BP) neural network was presented by analyzing its influential factors of end-point manganese content during BOF steelmaking process. By using PCA, the amount of variables which affected the end-point manganese content would be reduced. Then the principal components were employed to train the GA-BP neural network in order to obtain the prediction results of end-point manganese content for BOF steelmaking process. Comparing the prediction results of GA-BP neural network and PCA-GA-BP models, the results showed that GA-BP neural network prediction modeling method based on PCA method had the better prediction accuracy and the better generalization performance. The hit ratio of the model was 86% when the predictive errors of the model were within ±0.025%, and the mean square error was 2.74×10^-8. And this modeling method has high training speed.
Key words: ferrous metallurgy; basic oxygen furnace; end-point manganese content; BP neural network; genetic algorithm; principal component analysis (PCA)
发表期数: 2019年4月第2期
引用格式: 张壮,曹玲玲,林文辉,等. 基于主成分分析(PCA)与GA-BP结合的转炉炼钢终点锰含量预测研究[J]. 中国科技论文在线精品论文,2019,12(2):328-335.
 
6 评论数 0
暂无评论
友情链接