【作者】 於春月；

【导师】 王成恩；

【作者基本信息】 东北大学 ， 系统工程， 2009， 博士

【摘要】 钢铁工业是国民经济的支柱产业,在国民经济的发展中发挥着重要的作用。为了达到节能降耗的目的,自20世纪80年代起,国外钢铁企业开始广泛采用连铸坯热装热送技术,致力于开发炼钢—连铸—热轧—体化生产。一体化生产工艺复杂,连铸和热轧之间物流衔接紧密,生产计划与物流管理难度大,面向炼钢、连铸、热轧三大工序的一体化生产计划和调度问题,是钢铁企业一体化生产管理的核心。本文综述了钢铁企业一体化生产的特点以及生产计划和调度研究的现状,在此基础上,以国内某钢铁企业的中厚板一体化生产线为研究背景,围绕“提高热装比”这个核心目标,分别在合同计划、生产批量计划和作业计划(生产调度)三个层次上,重点研究了四个具体的钢铁一体化生产计划与调度优化问题:一体化生产合同计划、连铸与热轧生产计划协调优化、板坯热装调度和热轧生产调度问题。一体化生产合同计划优化问题。将炼钢—连铸—热轧一体化生产过程,化分为炼钢与热轧两大加工阶段,以板坯热装比最大、交货提前/拖期率和组炉余材率最小为优化目标,综合考虑炼钢与热轧的产能,以及钢种、板坯和成品规格等约束条件,建立了生产合同多目标优化模型,分别采用单目标优化和多目标优化方法进行求解。采用加权的方法,将一体化生产合同计划多目标优化问题转变成单目标问题,设计了双层染色体遗传算法进行求解,计算结果表明了模型的合理性与算法的有效性。运用改进的非支配排序遗传算法NSGA-Ⅱ(Non-dominated Sorting Genetic Algorithm-Ⅱ),求解钢铁一体化生产合同计划多目标优化问题。对NSGA-Ⅱ算法进行改进的方法是变异目标空间中的重合个体,以及在每一代增加若干个新个体。标准多目标数学测试问题、多目标背包问题和生产合同计划问题的计算结果表明,改进NSGA-Ⅱ算法的非劣解集,具有很高的相对覆盖度,证明了对算法的改进是有效的。中厚板热轧生产调度问题。以轧机利用率最大化为优化目标,分别研究了同规格板坯交叉轧制调度问题和异规格板坯交叉轧制调度问题。分析了交叉轧制的工艺特点,归纳出三类交叉轧制模式:均匀交叉轧制模式、对称交叉轧制模式和成组交叉轧制模式,并给出这三种交叉轧制模式的特点、可行条件、最佳开轧间隔时间及轧机利用率计算公式。考虑了辊道长度约束,对计算公式中的相关参数进行了调整,给出了一种同规格板坯交叉轧制生产调度的启发式方法。计算结果说明,该启发式方法优于现有启发式方法。异规格板坯交叉轧制调度问题,是一个有优先约束、等待时间和缓冲容量有限的单机调度问题。用AON(Activity-On-Node)网络对问题进行描述,提出并证明了面向单机调度问题的AON网络平衡定理,根据平衡定理,建立了非线性约束优化数学模型。设计了一个邻域搜索算法,将非线性约束问题转变为可用标准软件求解的线性约束问题。计算结果表明,这个邻域搜索算法可以找到比启发式方法更好的调度方案。板坯热装调度问题。连铸与热轧之间,板坯热装物流复杂,难以使用数学规划的方法,确定每块板坯的热装方式、出/入炉顺序和时间。本文采用基于仿真的方法,开发和利用物流仿真系统,确定每块板坯的热装方式、制定板坯的出/入库计划和每个加热炉的加热计划。连铸—热轧生产计划协调优化问题。以“铸轧序差M_O”最小为优化目标之一,综合考虑热轧计划约束和连铸计划约束,建立多目标生产计划协调优化模型,设计基于邻域的NSGA-Ⅱ算法进行求解。仿真结果表明,本文提出的连铸—热轧生产计划协调优化方法可行,调整后的热轧计划和浇铸计划,热装比有所提高。根据以上研究,本文开发了一体化生产计划与调度系统和板坯物流仿真系统,验证了本文提出的优化模型与算法。更多还原

【Abstract】 Iron & steel industry provides raw materials for other industries and plays an essential role in the industrial economy.Since 1980s,steelmaking-continuous casting-hot rolling integrated production and continuous casting slab hot charging technology have been studied and applied by overseas large iron and steel plant to save energy resources and reduce wastage.In steelmaking-continuous casting-hot rolling integrated production,production technology is complex and the logistics management between continuous casting and hot rolling is difficult.The integrated planning and scheduling problem between steelmaking, continuous casting and hot rolling is the key of the integrated steel production management.In this dissertation,the characters of the steel production are analyzed and the sate of art of integrated production planning and scheduling is also reviewed.Taking a national steel plate integrated steel production line as study background and aiming at increasing hot charging ratio,four special planning and scheduling problems about integrated steel production order planning,batch planning and scheduling are solved.The four problems are the production order planning problem of integrated steel production,the plate hot rolling scheduling problem,the slab hot charging scheduling problem and the continuous casting-hot rolling integrated optimization planning problem.For the production order planning problem of integrated steel production,a multi-objective optimization model is formulated,where steelmaking and hot rolling are regarded as the key steps.The objectives are to maximize the hot charging proportion of slabs and minimize the early/delayed delivery time and the slabs which are surplus to the requirements for hot/rolling.The main constraints including steelmaking and hot rolling capacities,steel grade and slab and product dimensions are all taken into consideration in the model.Single objective and multi-objective optimization method are used to solve the model. The multi-objective model is transformed into the single objective model by using a weighting method and is solved by a two-level chromosomes genetic algorithm.Computation results of different order planning problems show the rationality of the model and validity of the algorithm.An improved NSGA-Ⅱ(Non-dominated Sorting Genetic Algorithm) algorithm is designed to solve the multi-objective model.The NSGA-Ⅱalgorithm is improved by mutating the superposition individuals in the objective space and adding some new individuals on every generation so that the population diversity is improved significantly. Computation results of different standard multi-objective mathematic test problems, multi-objective knapsack problems and production order planning problems show the improved NSGA-Ⅱalgorithm is able to find solution set with higher coverage than NSGA-Ⅱ.For the plate hot rolling scheduling problem,with the objective to increase the mill utilization,two types of scheduling problems of different composite multi-plate rolling processes are researched.One is the same dimension composite multi-plate rolling process. The other is different dimension composite multi-plate rolling process.On the first simple problem,three modes of the process are classified,i.e.,equal-interval alternating, symmetric-interval alternating and grouped alternating.Thus the characteristics and feasible conditions for each of the three modes are described with the formulae of optimum time interval for starting rolling operation and mill utilization both given.Taking account of the restriction of mill table length,the parameters relevant to the formulae are adjusted properly to provide a heuristic method.The computation results showed that the proposed heuristic method is better than existing heuristic method with wider applications available.On the second complex problem,it is a typical single-machine scheduling problem subjecting to multiple constraints including precedence constraints and limited waiting time and finite buffer capacity.Describing the plate hot rolling scheduling problem with an AON (Activity-On-Node) network,an equilibrium principle is formulated for single-machine scheduling problems on the AON networks.Subsequently,a nonlinear constrained optimization model is built for plate hot rolling scheduling problems.A neighborhood search algorithm is proposed to convert the nonlinear model into a linear programming model that can be solved by standard software.Numerical computations indicate that the neighborhood search algorithm may find out better production schedules than the heuristic methods.For the slab hot charging scheduling problem,it is difficult to solve it based on mathematical programming methods.It is complex problem to compute every slab’s charging/outcharging order/time and hot charging mode.A logistics simulation system to make slab heating furnace hot charging planning and slab-yard charging/out charging planning is developed.For the continuous casting-hot rolling integrated optimization planning problem,a multi-objective optimization model is formulated.Minimizing the casting-rolling order diversity is the main objective.The constraints include continuous casting constraints and hot rolling constraints.NSGA-Ⅱalgorithm based on neighborhood is proposed to solve the problem.Simulation results show that the proposed integrated optimization method is feasible and hot charging ratio is increased. Based on the upwards works,integrated production planning and scheduling system and slab hot charging logistics simulation system are developed which validated the models and algorithms proposed in the dissertation.更多还原