Development And Application Of Scheduling Optimization Software For Steel-making And Continuous Casting

Multi-charges (one charge denotes that the molten steel in a converter) and multi-casts(one cast denotes that the continuous casted charges in the same caster) compose thesteel-making and continuous casting process by using multi-converters, mutli-refiningsmachines and mutli-casting machines with different refining modes. Molten iron transformsinto molten steel by the converters.. After that, the molten steel is poured into steel ladleswhich are the container for the molten steel. Steel ladles are transported to the refiningposition by trolleys and shop travelers for process of one or multi-refining modes (onerefining mode denotes that one charge is processed by one refining machines, multi-refiningmode denotes that one charge is processed by multi-refining machines). The refinedmolten steel is poured into the tundishes (tundish is the container for the molten steel), andthen transported to the continuous casting machines. Finally, the molten steel becomes slabsby the process of continuous cast.Providing that each charge’s process path (The total number of manufacturing procedureand the selected machine type in each manufacturing procedure for each charge) is given.The goal of the schedule of steel-making and continuous casting (SCC) is to strictly ensurethat the charges in the same cast should be continuously casted and the consecutive processedcharges in the same machine cannot conflict; The factors of energy saving and consumptionreduction are the difference value between the actual beginning time of the cast in the casterand the ideal beginning time of the cast in the caster and the waiting time between the twoconsecutive processed charges; The constraints are to satisfy the machine capacity and theprocessing time. Based on the requirements above, the schedule of SCC is to decide thebeginning time in each procedure for each charge and the machine type in each procedure,and form the table of SCC, that is the schedule of SCC.Now, the large steel-making enterprise is still using the manual scheduling method toinstruct the production. Because steel-making process is a high-speed and multi-routesprocess, it is difficult to instruct the production by using the manual scheduling method. Theoutput of the manual scheduling method only tell the scheduler the beginning time of the casting process and the ending time of the process in converters. For the process of refining,there is no the exactly instruction like the train table. This method will not let the schedulerhas the intuitively instruction for the production, and could easily leads the equipment is inthe idle status or the traffic for the transportation. Sometimes, the manual schedulingmethod also could lead to the breaking cast and the freezing of the molten steel. So, How tomake a good schedule in a computationally efficient manner and be able to assess the qualityof the schedule to improve the steel production and effectively assist the reschedule during theprocess of steel-making and continuous casting are the key factors for iron and steelproduction manufacturing productivity. In this paper, the development of the optimizationsoftware for the scheduling of steelmaking and continuous casting process adopts themathematical modeling technology to get the scheduling results. According to therequirements of the real production, we designed the initial module, equipment assignmentmodule, model generation module, conflict solution module and the data saving module.According to the different function of each module, this paper give the exactly instruction fo.Based on the mathematical model and the algorithm which is proposed in this paper, wedeveloped the steelmaking and continuous casting software. We took several industrialexperiments with the proposed approach in this paper based on a large steel plant, where itscontinuous casting process consists of three converters, seven refining furnaces, three castersand three refining modes. The schedule is created by the proposed method for10groups ofindustrial data, every group including3cast plans,20-23heats,3refining modes. The resultsare demonstrated as follows: The average time required to generate the schedule is1.2seconds which is much shorter than6seconds in real production. For the various groups ofindustrial test cases, the maximum waiting time of the heats in front of the caster do notexceed2minutes33seconds. Maximum of deviation to the targeted casting time is3minutes11seconds which is better compare with the results in real production. The test demonstratedthat the mathematical model and the algorithm has a wide application in the steel-makingindustry.