基于多目标差分进化算法的高炉煤气系统调度

摘要

Considering that the scheduling of blast furnace gas(BFG)system is crucial for energy saving in iron and steel industry,this study proposes a novel scheduling method for BFG system based on dynamic Bayesian network(DBN)and improved multi-objective differential evolution(IMODE) algorithm. On account of the dynamic characteristic of the BFG system and the output uncertainty of time prediction model,this study models the BFG system with the causality based DBN method. Simultaneously,the optimization target is that the gas cabinet reaches the desired value fast with a large margin for adjustment. When optimizing the scheduling schemes,the crowding distance of the particles is involved into the searching mechanism of IMODE algorithm to improve the searching precision. Furthermore,in view of the fact that the gas tank cannot run securely by adjusting a single user and the differences of adjustment ability of different users,a multi-user scheduling scheme method is proposed. In order to verify the effectiveness of the proposed method,experiments are carried out with the BFG system production data of a domestic steel enterprise. The results show that the proposed method is more effective than others for the scheduling of the BFG system.%针对钢铁工业中高炉煤气(blast furnace gas,BFG)系统的调度问题,提出了一种基于动态贝叶斯(dynamic Bayesian network,DBN)和改进的多目标差分进化(improved multi-objective differential evolution,IMODE)算法的 BFG 系统调度方法.考虑到 BFG 系统的动态特性和时间预测模型的输出不确定性,采用基于因果关系的DBN算法对BFG系统的煤气柜建立模型,并以煤气柜快速达到期望值且具有较大的调节余量为优化目标.在优化调度时,将粒子拥挤距离引入到多目标差分进化算法的搜索机制中,从而提高模型的搜索精度.此外,针对单个用户调整不能使煤气柜安全运行的情况,同时考虑到不同消耗用户调整能力的差异,提出了多用户调整方案.为了验证所提算法的有效性,采用国内某钢铁企业BFG系统生产数据进行实验,结果表明该方法相比其他的方法在BFG系统调度调整中具有更好的效果.