冷装工艺下钢坯库存作业与轧制计划的集成管理

发布时间：2018-01-19 01:05

本文关键词： 轧制计划编制库存作业优化单亲遗传算法启发式算法　出处：《北京科技大学》2015年博士论文　论文类型：学位论文

【摘要】：在钢铁企业中,尤其是冷装生产工艺下,钢坯库起着至关重要的作用,不仅接收连铸下线的入库钢坯,同时依照轧制计划向热轧工序提供原料(钢坯)。钢坯库内物流作业具有规模大、维度高、非线性等复杂特点,库内作业的优化对于热轧工序的生产具有重要的影响。本文对存放待轧制钢坯的钢坯库作业与轧制计划的集成管理展开研究,以特钢企业马莲垛式的钢坯库堆放方式和热轧计划编制与实施的实际情况为背景,深入分析库存的出入库作业与轧制计划编制之间的整合策略,针对以下4个关键问题设计优化模型及求解算法：考虑热轧生产需求的钢坯入库、出库作业优化；基于库存信息的热轧计划编制和调整。通过库存作业与轧制计划的集成管理,保证热轧生产的连续性和稳定性,同时降低库存作业成本。主要研究内容和创新点如下： (1)为保证后续热轧生产的连续和稳定生产,研究了考虑交货期的钢坯入库问题,对此问题建立了以最小化钢坯出库倒垛次数为目标的数学模型,并设计了基于动态调整策略的启发式算法。首先通过聚类算法将暂存在辊道上的钢坯进行聚类,形成钢坯入库批次；然后以批次为单位进行垛位指派；针对初步指派方案,构建了一种动态调整策略,动态调整每一个垛位最上层的钢坯批次,以降低整个入库计划的总倒垛次数。实验结果显示,算法能有效减少倒垛次数,并提高垛位的空间利用率。 (2)为有效减少轧制计划实施过程中的钢坯出库时间和降低出库成本,研究了一种考虑钢坯倒垛优化的轧制计划编制问题,即在粗轧制计划的基础上,为每个轧制位置匹配库存钢坯,并对轧制单元进行排序,形成可执行的轧制计划。通过深入分析问题特征,建立了以最小化轧制钢坯出库的总倒垛次数和轧制单元之间切换机架次数为目标的多目标整数规划模型,并构造了一种基于钢坯匹配的单亲遗传算法。实验结果验证了模型和算法的可行性和有效性。 (3)考虑到钢坯入库的实时动态性,为保证轧制计划的顺利执行,进一步研究了考虑实时库存的轧制计划调整问题,探讨了在实施轧制计划时钢坯库对轧制计划的影响因素,提出了轧制计划的三种动态调整策略：轧制钢坯替换、轧制单元内钢坯轧制顺序调整以及计划钢坯取消轧制调整。基于三种动态调整策略,建立了以最小化调整惩罚及钢坯库存倒垛次数的加权平均和为目标的数学模型,设计了一种三阶段启发式调整算法,为实施轧制计划提供了切实可行的方案。 (4)为提高轧制计划实施过程中的钢坯倒垛的有效性,打破了以往研究中障碍钢坯必须倒回的处理方式,综合考虑了轧制钢坯出库倒垛优化和为障碍钢坯重新指派位置的障碍钢坯落位优化,提出一种钢坯倒垛落位问题,并建立了两阶段数学模型：倒垛优化模型和落位优化模型。基于模型的两阶段特征,设计了两阶段求解算法：第一阶段,采用改进多回路启发式算法为轧制计划匹配最佳的钢坯库出库序列；第二阶段,基于垛位辐射邻域搜索算法为计划实施过程中的障碍钢坯指派最佳的落位位置。数据实验验证了模型和算法的有效性。
[Abstract]:In the iron and steel enterprise , especially the cold pack production technology , the steel billet warehouse plays a vital role , not only receives the incoming billet of continuous casting , but also provides the raw material ( billet ) to the hot rolling process according to the rolling plan . ( 1 ) In order to ensure the continuous and stable production of the subsequent hot rolling production , a mathematical model is established to minimize the number of stock issue stacks , and a heuristic algorithm based on the dynamic adjustment strategy is established . First , a dynamic adjustment strategy is constructed by clustering the steel billet on the temporary roller way to form a billet storage batch ; ( 2 ) In order to effectively reduce the stock issue time and reduce the issue cost in the rolling plan , a multi - objective integer programming model considering the optimized rolling schedule of the billet is studied , that is , based on the rough rolling plan , a multi - objective integer programming model with the objective of minimizing the total number of stacking times and the number of switching frames between the rolling units is established , and a single parent genetic algorithm based on the matching of the steel billet is constructed . ( 3 ) In order to ensure the smooth execution of rolling plan , the paper further studies the influence factors of rolling plan to ensure the smooth execution of rolling plan . Three kinds of dynamic adjustment strategies for rolling plan are discussed : rolling steel billet replacement , rolling sequence adjustment of steel billet in rolling unit and rolling adjustment of planned billet . Based on three dynamic adjustment strategies , a three - stage heuristic adjustment algorithm is designed to provide practical scheme for implementing rolling plan . ( 4 ) In order to improve the effectiveness of the steel billet stacking in the process of rolling plan , the paper breaks down the treatment way that the obstacle steel billet must be reversed in the past research , and proposes a two - stage mathematical model : the first stage , the optimized multi - loop heuristic algorithm is adopted to match the optimal steel billet stock issue sequence for the rolling plan ;

【学位授予单位】：北京科技大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：TF087;TG338

【参考文献】