本文選題：煉鋼廠 + 開澆決策��； 參考：《重慶大學(xué)》2016年博士論文

【摘要】：煉鋼流程是鋼鐵生產(chǎn)的關(guān)鍵環(huán)節(jié),生產(chǎn)作業(yè)計(jì)劃是其生產(chǎn)運(yùn)行控制的依據(jù)。合理的生產(chǎn)作業(yè)計(jì)劃可以降低物耗與能耗成本、增加收益、穩(wěn)定質(zhì)量,并直接提升企業(yè)的核心競爭力。為此,通過對(duì)煉鋼流程生產(chǎn)作業(yè)計(jì)劃相關(guān)的連鑄機(jī)開澆決策、生產(chǎn)運(yùn)行穩(wěn)定順行高效等基礎(chǔ)性問題的深入研究,以提升和發(fā)揮煉鋼廠制造執(zhí)行系統(tǒng)MES(Manufacturing Execution System,MES)的生產(chǎn)計(jì)劃調(diào)度功能,最終實(shí)現(xiàn)以信息化為基礎(chǔ)的生產(chǎn)運(yùn)行優(yōu)化控制模式取代傳統(tǒng)的人工經(jīng)驗(yàn)控制模式,成為鋼鐵企業(yè)廣泛關(guān)注并亟待解決的重要課題。由于煉鋼流程是一個(gè)由多階段大型高溫生產(chǎn)單元所構(gòu)成的、離散與連續(xù)工序相混雜的系統(tǒng),具有多目標(biāo)、多約束、動(dòng)態(tài)變化等復(fù)雜系統(tǒng)特征,而在不同的煉鋼廠由于工藝流程方面的各自不同特點(diǎn),致使建立統(tǒng)一且有效的生產(chǎn)作業(yè)計(jì)劃模型難度較大,而經(jīng)簡化抽象建立的通用性模型或算法與煉鋼廠的現(xiàn)實(shí)生產(chǎn)需求之間通常存在較大的差異。對(duì)煉鋼流程生產(chǎn)作業(yè)計(jì)劃編制相關(guān)問題的研究綜述和煉鋼廠制造執(zhí)行系統(tǒng)應(yīng)用情況的調(diào)研可見,已有研究對(duì)于連鑄機(jī)開澆、連連澆決策問題、多目標(biāo)要求下的連鑄機(jī)組澆開澆優(yōu)化問題,能動(dòng)態(tài)反映煉鋼廠現(xiàn)實(shí)生產(chǎn)穩(wěn)定性特征的仿真方法問題等方面認(rèn)識(shí)不足、手段有限;煉鋼廠MES系統(tǒng)的生產(chǎn)計(jì)劃調(diào)度功能與生產(chǎn)管理要求之間尚存在不適應(yīng)性,影響了煉鋼廠“有序、穩(wěn)定、高效”的生產(chǎn)目標(biāo)的實(shí)現(xiàn)。有鑒于此,以現(xiàn)實(shí)生產(chǎn)為背景,提出開展“煉鋼流程生產(chǎn)作業(yè)計(jì)劃編制相關(guān)基礎(chǔ)性問題研究”的博士論文課題,主要圍繞煉鋼廠的連鑄機(jī)開澆/連澆決策、連鑄機(jī)組澆開澆多目標(biāo)優(yōu)化、生產(chǎn)運(yùn)行的仿真優(yōu)化手段等問題開展建模、優(yōu)化求解算法與生產(chǎn)組織運(yùn)行優(yōu)化分析等研究工作,為煉鋼流程生產(chǎn)作業(yè)計(jì)劃已有研究成果的有效應(yīng)用提供更可靠的前提條件,并為煉鋼廠生產(chǎn)管控人員對(duì)生產(chǎn)穩(wěn)定控制的認(rèn)識(shí)提供一種新的仿真分析手段。論文的主要?jiǎng)?chuàng)新點(diǎn)及研究結(jié)論概述如下:1)建立了連鑄機(jī)開澆決策的混合整數(shù)規(guī)劃模型,并基于MATLAB軟件的YALMIP優(yōu)化工具進(jìn)行模型求解。為了通過生產(chǎn)物流穩(wěn)定順行來降低生產(chǎn)成本,針對(duì)煉鋼廠連鑄機(jī)開澆時(shí)是否連澆及開澆時(shí)間確定問題,以控制積壓液態(tài)金屬量成本最小和連續(xù)澆鑄的連續(xù)化程度收益最大為目標(biāo),在綜合考慮進(jìn)鐵量、安全生產(chǎn)線液態(tài)金屬量(簡稱安全在線金屬量)、金屬損耗、澆鑄鋼水量等涉及鐵鋼資源平衡的各因素之間的相互關(guān)系以及時(shí)間與生產(chǎn)線液態(tài)金屬量(簡稱在線金屬量)約束的條件下構(gòu)建了連鑄機(jī)開澆決策的混合整數(shù)規(guī)劃模型,并設(shè)計(jì)相應(yīng)的模型求解方法。2)構(gòu)建了連鑄機(jī)的開澆爐次與時(shí)間決策的多目標(biāo)優(yōu)化模型,并設(shè)計(jì)了改進(jìn)的非支配排序遺傳算法-INSGAII算法。針對(duì)現(xiàn)實(shí)連鑄機(jī)開澆決策中需同時(shí)確定爐次選擇、排序與開澆時(shí)間的多目標(biāo)優(yōu)化難題,以煉鋼廠生產(chǎn)批量計(jì)劃執(zhí)行情況的總懲罰、生產(chǎn)線積壓液態(tài)金屬量、優(yōu)質(zhì)鐵水非有效利用量最小為目標(biāo)函數(shù),構(gòu)建了連鑄機(jī)的開澆爐次與時(shí)間決策的多目標(biāo)優(yōu)化模型;基于非支配排序遺傳算法設(shè)計(jì)了改進(jìn)的NSGAII算法進(jìn)行模型求解,以預(yù)選池內(nèi)選擇的爐次序號(hào)為基因的編碼方式來減小模型解的無效搜索空間,采取調(diào)整傳統(tǒng)精英解集計(jì)算順序、限定計(jì)算擁擠距離個(gè)體數(shù)目的改進(jìn)措施來減輕計(jì)算負(fù)荷,利用對(duì)pareto解進(jìn)行模糊選優(yōu)的方法來確定最終優(yōu)化解。3)建立了煉鋼廠生產(chǎn)線金屬量控制的系統(tǒng)動(dòng)力學(xué)仿真模型,并以某煉鋼廠的實(shí)際生產(chǎn)數(shù)據(jù)對(duì)模型進(jìn)行仿真分析,來動(dòng)態(tài)反映煉鋼廠現(xiàn)實(shí)生產(chǎn)的穩(wěn)定性特征�；谙到y(tǒng)工程的思想方法,在綜合考慮生產(chǎn)作業(yè)計(jì)劃需求流量、生產(chǎn)線金屬量的目標(biāo)庫存、實(shí)際庫存、庫存偏差等信息,以及相關(guān)物質(zhì)流影響的基礎(chǔ)上,建立了以煉鋼廠生產(chǎn)線金屬量為控制水平變量的系統(tǒng)動(dòng)力學(xué)仿真模型。4)以某煉鋼廠生產(chǎn)條件為對(duì)象的優(yōu)化模型應(yīng)用測試表明:(1)煉鋼廠連鑄機(jī)的開澆時(shí)間決策優(yōu)化模型可以實(shí)現(xiàn)連鑄機(jī)澆次開澆時(shí)間的科學(xué)計(jì)算,有助于穩(wěn)定各班次之間的生產(chǎn)條件,降低生產(chǎn)線上的積壓液態(tài)金屬量,編制出合理的煉鋼廠生產(chǎn)作業(yè)計(jì)劃;(2)連鑄機(jī)開澆爐次與時(shí)間多目標(biāo)決策模型有利于連鑄機(jī)上各爐次澆鑄周期的穩(wěn)定控制,并有利于在煉鋼流程切實(shí)推行計(jì)劃管理,改進(jìn)的非支配排序遺傳算法的效率優(yōu)于傳統(tǒng)的非支配排序遺傳算法與強(qiáng)度pareto進(jìn)化算法。系統(tǒng)動(dòng)力學(xué)模型仿真研究表明:(1)有生產(chǎn)作業(yè)計(jì)劃指導(dǎo)下煉鋼-連鑄區(qū)域的運(yùn)行狀態(tài)明顯較無作業(yè)計(jì)劃指導(dǎo)的鐵水預(yù)處理區(qū)更穩(wěn)定,因此,煉鋼廠應(yīng)實(shí)行按全流程的生產(chǎn)作業(yè)計(jì)劃的運(yùn)行管控;(2)當(dāng)加快生產(chǎn)節(jié)奏提高連鑄機(jī)的拉速時(shí),在線金屬量應(yīng)同步提高,否則會(huì)因?yàn)樵诰�金屬量的降低,而導(dǎo)致生產(chǎn)不穩(wěn)定。5)為進(jìn)一步檢驗(yàn)所建立的多目標(biāo)開澆決策與系統(tǒng)動(dòng)力學(xué)模型的實(shí)用性,進(jìn)行了聯(lián)合仿真實(shí)驗(yàn)研究。通過將前者決策結(jié)果作為后者輸入?yún)?shù),并以實(shí)例煉鋼廠的生產(chǎn)數(shù)據(jù)為依據(jù),分別對(duì)連鑄機(jī)的開澆時(shí)間、品種鋼比例、連鑄機(jī)數(shù)量等指標(biāo)與煉鋼廠各區(qū)域在線金屬量的影響關(guān)系進(jìn)行聯(lián)合仿真實(shí)驗(yàn),仿真研究表明:在進(jìn)鐵流量一定的情況下,(1)推遲開澆時(shí)間或者增大品種鋼比例會(huì)使在線金屬量增高;(2)增加連鑄機(jī)數(shù)量會(huì)降低煉鋼廠各區(qū)域在線金屬量;(3)在進(jìn)鐵流量存在差異情況下,平穩(wěn)進(jìn)鐵較隨機(jī)進(jìn)鐵更有利于在線金屬量的穩(wěn)定控制。綜上所述,本文所建立的煉鋼廠連鑄機(jī)開澆時(shí)間決策優(yōu)化模型、連鑄機(jī)開澆爐次與時(shí)間決策的多目標(biāo)優(yōu)化模型,為科學(xué)確定煉鋼廠連鑄機(jī)的開澆時(shí)間與開澆爐次提供了新的技術(shù)手段,為煉鋼流程生產(chǎn)作業(yè)計(jì)劃編制的假設(shè)條件問題提供了科學(xué)決策方法;所建煉鋼廠生產(chǎn)線金屬量控制的系統(tǒng)動(dòng)力學(xué)模型,以及針對(duì)多目標(biāo)開澆決策與系統(tǒng)動(dòng)力學(xué)模型的聯(lián)合仿真實(shí)驗(yàn)研究,為深入認(rèn)識(shí)煉鋼廠生產(chǎn)線上液態(tài)金屬量動(dòng)態(tài)變化特性提供了新方法,給煉鋼廠生產(chǎn)運(yùn)行的穩(wěn)定控制提供了有效的仿真分析手段。
[Abstract]:The steelmaking process is the key link in the steel production, and the production operation plan is the basis for its production control. The reasonable production operation plan can reduce the cost of material consumption and energy consumption, increase the income, stabilize the quality, and directly improve the core competitiveness of the enterprise. In order to improve and exert the production planning and scheduling function of MES (Manufacturing Execution System, MES) in the steelmaking plant, the production operation and operation system based on the information based production and operation optimization control mode is finally realized and the traditional artificial experience control model is replaced by the iron and steel enterprise. Because the steelmaking process is a multi stage large and high temperature production unit, the discrete and continuous processes are mixed with complex systems, which have complex system characteristics such as multi-objective, multi constraint, dynamic change and so on. In different steelworks, the different characteristics of process flow are caused by different characteristics of different steelmaking plants. It is difficult to establish a unified and effective production operation plan model, and there is a large difference between the general model or algorithm established by simplified abstract and the actual production demand of the steelmaking plant. A survey on the related problems of the production planning of steelmaking process and the investigation of the application of the manufacturing execution system of the steelmaking plant In view, there have been research on the problem of continuous casting casting, continuous casting decision, optimization of casting cast in cast in continuous casting machine under multi objective requirements, lack of knowledge and limited means, which can dynamically reflect the practical production stability characteristics of the steelworks, and the means are limited, and the production planning and scheduling function of the steelworks MES system is still between the requirements of production management. The existence of inadaptability affects the realization of the production target of "orderly, stable and efficient" in the steelmaking plant. In order to provide more reliable precondition for the effective application of the existing research results of the steelmaking process production operation plan, it provides more reliable prerequisite for the effective application of the existing research results of the steelmaking process production operation plan, and the stable control for the production management and control personnel of the steelmaking plant. The main innovation and research conclusions of this paper are summarized as follows: 1) 1) the mixed integer programming model of casting casting decision making is established, and the model is solved based on the YALMIP optimization tool of the software. In order to reduce the production cost through the stable production of the production logistics, the continuous caster of the steelmaking plant is made. In order to control the minimum cost of the backlog of liquid metal and the maximum yield of continuous casting, the aim is to control the minimum cost of the backlog of liquid metal and the continuity of continuous casting. The factors involved in the balance of iron and steel resources are considered in the comprehensive consideration of the amount of iron intake, the liquid metal content of the safety production line, the metal loss and the casting amount of the steel. The relationship between the time and the time and the liquid metal content of the production line (on the line metal amount) constrains the mixed integer programming model of the casting casting decision of the continuous casting machine, and the corresponding model solving method.2 is designed. The multi-objective optimization model of the time decision of the open casting furnace and the time of the continuous casting machine is constructed, and the improved non support is designed. According to the -INSGAII algorithm of sorting genetic algorithm, it is necessary to determine the multi-objective optimization problem in the decision making of the real caster casting at the same time, the total penalty of the sequencing and opening time, the total penalty of the production batch plan in the steelmaking plant, the backlog of liquid metal and the minimum non effective utilization of the high quality iron water as the objective function, and the continuous casting is constructed. The multi-objective optimization model of the timing and time decision of the open casting furnace is made. Based on the non dominated sorting genetic algorithm, an improved NSGAII algorithm is designed to solve the model. The invalid search space of the model solution is reduced by the encoding method of the selected furnace sequence number in the pre selected pool to reduce the invalid search space of the model solution, and the calculation order of the traditional elite solution set is adjusted and the calculation support is limited. In order to reduce the number of individual numbers, to reduce the calculation load and to determine the final optimal solution by using the method of fuzzy optimization of Pareto solution to determine the final optimal solution.3), the system dynamics simulation model of the metal quantity control of the production line of the steelmaking plant is established, and the simulation analysis of the mould type is carried out with the actual production data of a steelmaking plant to dynamically reflect the steelmaking plant. Based on the thought method of the system engineering, based on the comprehensive consideration of the production plan demand flow, the target inventory of the metal quantity of the production line, the actual inventory, the inventory deviation, and the influence of the related material flow, the system dynamics imitation is established with the control level variable of the production line metal quantity of the steelmaking plant. True model.4) the application test of optimization model based on the production conditions of a steelmaking plant shows that: (1) the optimization model of casting time decision optimization model of continuous casting machine in steelmaking plant can realize the scientific calculation of casting time in cast in place of continuous casting machine, help to stabilize production conditions between various flights, reduce the backlog of liquid metal in production line, and make reasonable The production operation plan of the steelmaking plant; (2) the multi-objective decision model of the continuous casting furnace time and time is beneficial to the stability control of the casting cycle of the continuous casting machine, and is beneficial to the practical implementation of the plan management in the steelmaking process. The efficiency of the improved non dominated sorting genetic algorithm is better than the traditional non dominated sorting genetic algorithm and the strength Pareto. The simulation study of system dynamics model shows that: (1) the operation state of the steelmaking and continuous casting area under the guidance of production operation plan is more stable than that without the guidance of the operation plan. Therefore, the steelmaking plant should carry out the operation control of the production operation plan according to the whole process; (2) when the production rhythm is accelerated, the continuous casting machine should be increased. At the speed, the on-line metal amount should be increased synchronously, otherwise the production is unstable due to the reduction of the metal content on the line, which leads to the production of unstable.5). A joint simulation experiment is carried out to study the practicability of the multi target opening decision and the system dynamics model established by the further inspection. Based on the production data of the factory, the simulation experiments are carried out on the relationship between the opening time of the continuous casting machine, the proportion of variety steel, the quantity of continuous casting machine and the influence of the on-line metal quantity in the various regions of the steelmaking plant. The simulation study shows that (1) the delayed opening time or the increase of the ratio of the variety of steel to the on-line metal under the certain flow rate of the iron intake (2) (2) increasing the number of continuous casting machines will reduce the online metal content in each area of the steelmaking plant; (3) in the case of difference in the flow of iron, the steady intake of iron is more favorable to the stable control of the online metal quantity than the random incoming iron. The multi-objective optimization model of decision making provides a new technical means for determining the opening time of the continuous casting machine and the opening of the open cast furnace for the scientific determination of the steelmaking plant. It provides a scientific decision method for the hypothetical conditions of the production planning of the steelmaking process, the system dynamics model of the metal quantity control of the production line of the steelmaking plant, as well as the multi-objective opening. The joint simulation experiment of casting decision and system dynamics model provides a new method to understand the dynamic change characteristics of liquid metal content on the production line of steelmaking plant, and provides an effective simulation analysis method for the stable control of the production operation of the steelmaking plant.
【學(xué)位授予單位】：重慶大學(xué)
【學(xué)位級(jí)別】：博士
【學(xué)位授予年份】：2016
【分類號(hào)】：TF758
，

本文編號(hào)：1954769