【作者】 蔡雁；

【导师】 吴敏；

【作者基本信息】 中南大学 ， 控制科学与工程， 2013， 博士

【摘要】 烧结综合料场是存放钢铁企业原料的场地,综合料场占地面积较大,存储的原材料是企业正常生产的前提。综合料场生产工序复杂,生产成本大,建立高效的烧结综合料场作业管理与优化系统对于钢铁企业节约生产成本具有重要意义。目前,烧结综合料场的作业过程并未得到充分优化,在料场储位配置方面,存在存储混乱,料场利用率及原料稳定性不高的问题；在库存量管理方面,存在原料库存量采购不合理,导致库存过多或过少的问题；在料场管理方面,存在堆取料机作业易失误,料场信息不准确等问题。针对上述问题,本文围绕烧结综合料场作业管理与优化方法展开研究,主要的研究工作与创新点如下：(1)基于层次分析法的料场储位模糊多准则优化方法由于钢铁企业原料来源广泛、品种繁多、数量庞大、外在影响因素比较严重等特点,所以原料在综合料场中的各个存储方案的评价是一个定性和定量相结合的多目标优化问题。一般的优化算法不能从根本上解决料场利用率和原料成分稳定性的问题。本文通过分析料场的原料存储情况,以料场利用率和原料场成分稳定性为目标建立储位优化模型。首先根据储位配置的主要影响因素建立七个优化准则,然后应用层次分析法确定各个准则的权重值,解决准则难以量化的问题；最后采用三角模糊数的方法表示权重值和期望值,提高优化结果的合理性。现场运行结果表明,采用该方法建立的系统可以达到预期优化目标,大大降低企业管理的成本。(2)基于多模型集成的原料库存量预测方法原料由于价格、产地及品味的不同在重要程度上存在差异,但是,在形成中和粉的过程中,不同重要程度的原料所占比例是大致不变的,所以原料总量的库存量变化存在一定的规律性,不同重要程度的原料变化也是具有一定规律性的。本文通过深入分析原料的库存量变化特点及影响因素,提出一种基于灰色系统模型与时间序列模型的预测方法来预测原料库存量。首先利用灰色系统可以反映数据序列整体发展趋势的优点,通过对库存量历史数据的滤波、统计、累加或累减生成,建立库存量灰色预测模型,然后利用时间序列模型可以反映数据序列细节波动的优点,对数据序列进行时序分析,建立时间序列自回归积分移动平均模型；最后,建立基于信息熵的原料库存量集成预测模型,该集成模型可以充分集合不同类型模型的优点,准确预测原料库存量,为后续的原料库存量优化提供可靠的依据。(3)基于GA-PAO算法的烧结料场原料库存量优化料场库存量预测的目的是为了防止断料现象的发生,保证生产的连续性,而库存量优化的则是为了最大程度地优化采购库存成本,二者共同构成了料场的原料库存量优化管理。原料采购库存成本的约束是钢铁企业流动资金的制约瓶颈。本文针对钢铁企业烧结料场原料采购与消耗的特点,以企业原料库存费用最小为目标建立烧结综合料场原料库存量优化模型,构建了详细的模型目标函数及约束条件；提出基于标准PSO算法的适应度函数,将GA算法用于PSO算法的改进,设计详细的算法参数,并对GA-PSO算法与标准的PSO算法进行寻优结果对比。同时,应用某钢铁企业烧结生产线的综合料场实际生产数据进行仿真,结果表明,该库存量优化模型结合GA-PAO算法实现了原料库存成本的优化,为钢铁企业采购计划的制定提供决策支持。(4)烧结综合料场作业管理与优化系统的设计及其工业应用结合工业现场实际,根据某钢铁企业360m2烧结生产线对系统的软件、硬件结构进行分析,建立烧结综合料场作业管理与优化系统,并阐述了系统的应用软件模块、优化控制算法的实现流程以及数据通信技术。通过对系统的实际运行结果进行分析,发现该系统可以实现企业兼顾采购成本与保证原料供应稳定性的综合优化,同时实现料场储位的优化配置,提高料场利用率及原料稳定性,取得明显的经济效益。更多还原

【Abstract】 In iron and steel enterprises, sintering comprehensive stockyard is used to store raw materials, and normally cover large area, in order to guarantee routine large-scale production. Because the production process is complicated, and the production cost is large, it is of great significance to establish an efficient sintering comprehensive stockyard operation management and optimization system for production cost savings. Currently, the operating procedure of sintering comprehensive stockyard has not been fully optimized. In the aspect of storage configuration, there exist problems such as confusion storage mode, low utilization rate of stockyard and low stability of raw materials. In the aspect of inventory management, the procurement of raw materials is unreasonable, which results in too much or too little inventory. In the aspect of stockyard management, there exist mistakes of stacker reclaimers’job, inaccurate information of stockyard and other problems. In response to these issues, this paper conducts research on job management and optimization of sintering comprehensive stockyard, the main research work and innovations are as follows:(1) Fuzzy multiple criteria optimization method based on analytic hierarchy process for storage location choiceThe supply of raw materials in iron and steel enterprises always shows its complexity because of large variety of sources and composition, large purchase quantity, and serious impacts coming from production. So it is a combination of qualitative and quantitative multi-objective optimization problem for the evaluation of raw materials storage solutions. Conventional optimization algorithms can hardly improve the stockyard utilization and the materials ingredients stability fundamentally. Based on the analysis on the situation of raw materials storage, this paper establishes a storage locations optimization model, which aims at stockyard utilization and materials ingredients stability. Firstly, seven optimization criterions are established based on the main factors affecting storage locations configuration. Secondly, the weight values of various criteria are determined by using analytic hierarchy process method. Finally, the method with triangular fuzzy numbers is used to represent the weight values and the desired values, in order to improve the reasonability of optimization results. The running results show that, the system established can achieve the desired optimization goal, and reduce the cost of enterprise management greatly.(2) Prediction method based on multi-model integration for iron mine powders inventoriesDue to the difference of origin and grade, raw materials have different price, show importance in different degrees in batching process. However, in the formation process of powder, the proportions of different kinds of raw materials are nearly unchanged. Hence there is certain regularity about the changes of the total raw materials inventory, as well as the changes of the special kinds of raw materials with different importance. This paper proposes a prediction method based on gray system model and time series model to predict raw materials inventory. Firstly, as gray system can reflect the overall trend of the data sequence, gray prediction model for inventory is established by filtering, statistics, cumulation, or regression based on historical inventory data. Secondly, because of the advantage in reflecting fluctuations of the data sequence, according to the timing analysis of the data sequence, the time series model is applied. Finally, the integrated prediction model is established based on information entropy for raw materials inventory. This integrated model can fully reflect the advantages of different types of models, predict the raw material inventory accurately, and provide a reliable basis for the following optimization of raw materials inventory.(3) Raw material inventory optimization algorithm for sinter material plant based on GA-PSOThe goal of stockyard inventory prediction is to prevent the occurrence of the raw materials shortage, so as to ensure the continuity of production. At the same time, inventory optimization aims to reduce the cost of purchasing inventory. The two aspects constitute the raw materials inventory optimization management for stockyard. The constraints of raw materials inventory and procurement cost are the bottleneck of liquidity in iron and steel enterprises. According to the characteristics of procurement and consumption of raw materials in sintering comprehensive stockyard, this paper establishes an optimization model for raw materials inventory, which targets on minimum inventory costs, and builds the detailed model of the objective function and the constraints. Firstly, a standard PSO algorithm is proposed based on fitness function. Secondly, GA algorithm is applied to improved PSO algorithm and the detailed algorithm parameters are designed. Lastly, the optimization results of GA-PSO algorithm and standard PSO algorithm are compared. Further more, the actual production data of sintering comprehensive stockyard in a steel enterprise are used for simulation, and the results show that this inventory optimization model combined with GA-PSO algorithm has achieved the optimization of raw materials inventory costs, which provides decision support for the procurement plan in iron and steel enterprises.(4) The design and industrial applications of the job management and optimization system for sintering comprehensive stockyardBased on the analysis of system software and hardware structure of a360m2sintering production line, and the actual situation of the industrial field, the job management and optimization system for sintering comprehensive stockyard is established. What’s more, the implementation process of application software modules, the optimization control algorithm and the data communication technology of this system are introduced. After the analysis of the actual operation of the system results, it is found that this system can realize the comprehensive optimization of procurement costs and stability of raw materials supply, and achieve the goal of optimization configuration of the stockyard storage locations, improve the stockyard utilization and raw materials stability. That makes significant economic benefits.更多还原