铁钢界面多模式运输的机车调度计划智能优化  

Intelligent optimization of locomotive scheduling for multi-mode transportation on ironmaking and steelmaking interface

作　　者：何姝静 郑忠[2] 朱道飞[1] 张开 王飞 沙远洋 HE Shujing;ZHENG Zhong;ZHU Daofei;ZHANG Kai;WANG Fei;SHA Yuanyang(Faculty of Metallurgical and Energy Engineering,Kunming University of Science and Technology,Kunming 650093,Yunnan,China;College of Materials Science and Engineering,Chongqing University,Chongqing 400044,China;Faculty of Mechanical and Electrical Engineering,Kunming University of Science and Technology,Kunming 650500,Yunnan,China;Manufacturing Department,Shougang Jingtang United Iron and Steel Co.,Ltd.,Tangshan 063200,Hebei,China)

机构地区：[1]昆明理工大学冶金与能源工程学院,云南昆明650093 [2]重庆大学材料科学与工程学院,重庆400044 [3]昆明理工大学机电工程学院,云南昆明650500 [4]首钢京唐钢铁联合有限责任公司制造部,河北唐山063200

出　　处：《钢铁》2024年第4期188-199,共12页Iron and Steel

基　　金：国家自然科学基金重点资助项目(51734004,52334008)。

摘　　要：目前铁钢界面“一包到底”模式的机车调度研究中,多包型铁水包运送任务存在供需匹配、运输路径选择、机车组包运输等统筹优化能力不足,限制了机车调度计划的优化。为减少机车使用数量,降低机车空载率,从而降低机车运输成本,在保障高炉出铁空包需求、各炼钢厂铁水准时供应的前提下,兼顾运输路径特点、铁包包型、机车运输能力、运包模式、组合运包时间、机车安全运行时间窗口等约束,建立以最小化多模式下机车运输成本为优化目标的机车调度计划优化数学模型;设计了一种有机结合遗传算法与Dijkstra算法的混合遗传算法求解此模型,实现了多包型多运包模式下机车调度计划和运输路径的快速优化。基于某企业12 h实际生产数据的测试结果表明,机车调度计划数学模型和求解算法能在7 min内制订出合理的机车调度计划。与传统蚁群算法求解相比,混合遗传算法收敛更快,机车使用台数由17台减至16台,且机车运输任务次数减少2.94%,机车空载率下降4.59%;与生产实际相比,机车使用台数由18台减至16台,且机车运输任务次数减少10.81%,机车空载率下降35.43%。该方法可有效降低机车运输成本,降低不同机车调度员的个人经验对机车调度方案编制合理性的影响,减轻调度员优化机车调度方案的难度和工作量。In the current research of locomotive scheduling on the one-ladle technology on ironmaking and steelmaking interface,the multi-ladle transportation task has insufficient overall optimization ability,such as supply and demand matching,transportation path selection,and the transportation mode of the ladles,which limits the optimization of locomotive scheduling.In order to reduce the number of locomotives used,the empty-loaded rate of locomotives and the cost of locomotive transportation,on the premise of ensuring the demand of empty ladles for blast furnace tapping and the demand of heavy ladles in various steelmaking plants,a locomotive scheduling's mathematical model which takes the minimized multi-mode locomotive transportation cost as an optimization object is established by taking the transportation path characteristics,the types of ladles,the locomotive transportation capacity,the transportation mode of the ladles,the combined transportation time of the ladles and the safe time window of the locomotive into consideration,and a hybrid genetic algorithm combining genetic algorithm and Dijkstra algorithm is designed to solve the model,which realizes the rapid optimization of locomotive scheduling and transportation route under multi-ladle mode.The simulation experiment results based on the actual production data of an enterprise within 12 h show that the mathematical model and solution algorithm of locomotive dispatching plan can work out a reasonable locomotive scheduling plan within 7 min.Compared with the traditional ant colony optimization algorithm,the hybrid genetic algorithm has faster convergence speed,the number of locomotives is reduced from 17 to 16,the number of locomotive transportation tasks is reduced by 2.94%,and the empty-loaded rate is reduced by 4.59%.Compared with the actual production,the number of locomotives used is reduced from 18 to 16,the number of locomotive transportation tasks is reduced by 10.81%,and the empty-loaded rate is reduced by 35.43%.This method can effectively reduce the c

关 键 词：一包到底 机车调度计划 多运包模式 混合遗传算法 铁钢界面 

分 类 号：TF086[冶金工程—冶金物理化学]