基于主客观证据融合的高炉悬料预测方法  

Prediction method of blast furnace hanging based on fusion of subjective and objective evidences

作　　者：曹卫华[1,2] 杜楠[1,2] 安剑奇[1,2] 吴敏[1,2] 

机构地区：[1]中南大学信息科学与工程学院,长沙410083 [2]先进控制与智能自动化湖南省工程实验室,长沙410083

出　　处：《北京科技大学学报》2014年第4期506-514,共9页Journal of University of Science and Technology Beijing

基　　金：湖南省自然科学省市(衡阳)联合基金资助项目(12JJ8012);国家自然科学基金资助项目(61203017);国家自然科学基金重点资助项目(61333002);国家高技术研究发展计划资助项目(2012AA040307)

摘　　要：针对高炉关键异常炉况悬料难以预测的问题,基于D-S证据理论,提出一种综合模糊专家推理和后验概率最小二乘支持向量机的悬料预测方法.首先,结合高炉生产过程和悬料现象,分析悬料形成的内在机理;其次,通过模糊专家推理提取基于专家规则的主观证据,再通过建立后验概率最小二乘支持向量机模型提取基于数据内在客观规律的客观证据;最后,基于D-S证据理论完成主客观证据融合,实现悬料预测.该方法充分利用专家经验和最小二乘支持向量机的自学习能力,能够提高预测精度.仿真结果表明本文提出的方法有效、准确.Aiming at the difficulty of predicting blast furnace hanging, a prediction method was proposed for the hanging based on the D-S evidence theory and in combination with fuzzy expert inference and a posterior probability least squares support vector machine. Firstly, the causes of hanging are obtained by mechanism analysis in consideration of blast furnace operations and hanging phenomena. Secondly, subjective evidences are extracted by fuzzy expert reasoning, while a posterior probability least squares support vector machine model is developed to extract objective evidences. Finally, in order to predict the hanging precisely, the subjective and objec-tive evidences are fused based on the D-S evidence theory, which makes full use of the expertise and the self-learning ability of the least squares support vector machine. Simulation results illustrate that the proposed method can make accurate prediction of the hanging.

关 键 词：高炉 预测 模糊逻辑 最小二乘逼近 支持向量机 信息融合 

分 类 号：TF543.1[冶金工程—钢铁冶金]