焦炭质量的DE-BP神经网络预测模型研究  

Prediction Model of Coke Quality Based on DE-BP Neural Network

作　　者：陶文华[1] 袁正波 Tao Wenhua;Yuan Zhengbo(School of Information and Conlxol Engineering, Liaoning Shihua University, Fushun 113001, China)

机构地区：[1]辽宁石油化工大学信息与控制工程学院,辽宁抚顺113001

出　　处：《系统仿真学报》2018年第5期1650-1656,共7页Journal of System Simulation

基　　金：国家自然科学基金面上基金(61473140);国家自然科学基金青年基金(61203021)

摘　　要：焦炭的质量对高炉冶炼的生产有着重要的影响,为解决焦炭质量预测线性方法计算量大、检验难问题,在分析焦炭质量影响因素基础上,采用主元分析的方法确定配合煤参数的输入向量,以焦炭质量指标中灰分Ad、硫分Std、抗碎强度M40、耐磨强度M10作为输出向量,建立了焦炭质量的DE优化BP神经网络预测模型。预测效果表明该模型焦炭质量各项指标的预测值与真实值的相对误差均在4%以下,克服了BP神经网络预测精度低、易陷入局部极小值的缺点,可以满足生产工艺的要求,对焦炭生产具有一定的使用价值。The quality of coke has a great effect on the furnace process. In order to solve the problem of large amount of calculation and inspection of coke quality prediction linear method, the coke quality prediction model based on DE-BP neural network is established on the basis of analyzing the factors affecting coke quality, which uses principal component analysis method to determine the parameters of the input vector with the coal and coke ash quality indicators Aa, sulfur Sta, crushing strength M40, abrasion resistance Mlo as the ouqTut vector prediction. The result of simulation shows that the relative errors of real and estimated values of the indicators are below 4%, overcoming the low prediction precision of BP neural network and the shortcoming of easily trapped in local minimum. It can meet the requirements of production process and has a certain value for coke production.

关 键 词：BP神经网络 焦炭质量 差分算法 预测模型 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]