Hybrid Optimal Scheduling Method Of Steelmaking-Continuous Casting And Its Application

The steelmaking-continuous casting process in modern large steel plant is constituted by several casts, heats, converters, refining furnaces, casters and multiple refining modes. Initially, the molten steel is poured into the ladle through the converter (which is called a heat), then it is transported to the refining furnace to be refined, and finally it is sent to the tundish. The molten steel that is constituted of several heats is cast to be the slab (which is called a cast) in the caster.On the basis of the cast plans, the steelmaking-continuous casting production scheduling aims at determining the heats in the cast plan production machines including converter, refining furnace, and their starting and end process time corresponding converter, refining furnace and caster. The target that the cast plan is carried out punctually, the heats in the cast is cast on caster continuously, the heats production time cannot be conflicted mutually in the same machine, and in front of the caster, waiting time of the heats do not exceed given threshold, subjecting to the constraints of fixed production process, fixed production time in each equipment and transporting time between different machines, eventually obtaining the production timetable (is called the schedule). However, the schedule will actually implemented by the ladle. The molten steel will extend the processing time if there is no ladle to accept the molten steel from converter, when the molten steel have completed smelting in converter. Moreover, if the ladle temperature is too low, it is necessary to extend the baking time on baker, thus increases energy consumption. Therefore, the ladle matching problem should be considered during the steelmaking-continuous casting production scheduling.The scheduling process of steelmaking-continuous casting with several converters, refining furnaces, casters and multiple refining modes not only requires that every cast plan should be cast on time, the heats in the cast plan should be cast continuously on the same caster, and in front of the caster, waiting time of the heats do not exceed the given threshold; but also demands that the heats production time cannot be conflicted mutually in the same converter or refining furnace. The latter constraints cannot be described by accurate mathematical models and thus cannot be solved using the traditional optimization method effectively.Because the existing scheduling models are not for multiple refining modes the way, have not consider the heats do not exceed the given range (threshold), and ignore the ladle effect on schedule, the existing research methods can not be applied to the steel plant with multiple converters, multiple refining furnace and casters, multiple refining modes. As a result, the steel plant had to use manual scheduling and matching ladle. However, it (manual scheduling) brings the problem of long redundant waiting time between different machines, which makes the molten steel temperature drop to such a low degree that every cast plan cast on time can not be guaranteed. Moreover, the randomicity of the manual matching easily lead to energy waste.Supported by "Key Technical Research and Demonstration of Application on MES in Irion and Steel Enterprises" of the National863High Technology Program project (2004AA412010), this PhD project aims at tackling the above mentioned practical challenging issue. The work takes a large steel plant as the system to be considered, whose production unit consists of three converters, seven refining furnaces, three casters and three refining modes. As the research background, a hybrid optimal scheduling approach and its application are studied in this dissertation, where the major contributions are listed as follows:1.Via the analysis of the state of production a large steel plant with three converters, seven refining furnaces, three casters and three refining modes, the casting plan, schedule and manual scheduling, this dissertation develops the continuous casting scheduling models for steelmaking, which consists of a performance index equation and a constrain equation. The purpose of the performance index equation aims at minimizing the waiting time, keeping the punctuality of the casting plan and the continuity of the casting process which is subjected to the constraints of fixed production process, fixed production time in each equipment, fixed transporting time between different machines, in front of the caster, waiting time of the heats do not exceed given threshold and the un-confliction in the schedule so as to eventually obtain the production timetable. Moreover, the reason that the existing optimization algorithm cannot be applied to the proposed issue is also analyzed. The constraints and objectives of the selected ladle are established based on factors (including ladle material properties, temperature, nozzle etc) from manual selecting ladle knowledge. Moreover, the difficulties in accurately modeling of matching ladle are analyzed, and the problems of manual scheduling and ladle matching are pointed out.2. Aiming at tackling the problems that manual made schedule low efficiency, the heat wait time too long between different machines is caused easily, ladle is selected by man likely to causing energy wastage. A hybrid optimization scheduling strategy is proposed by integrating dynamic program, linear program, orthogonal design and man-machine interface. The proposed strategy is constituted by optimization schedule of main equipment and ladle matching. On the basis of the cast plans, optimization schedule of main equipment aims at determining the every heat in the cast plan production machines including converter, refining furnace and their corresponding starting process time as well as starting casting time in casters. Finally goals that the cast plan is carried out punctually without break, the waiting time is minimized is achieved. Ladles are selected as hot as possible ladle according to the ladle matching, under the constraints including ladle material properties, ladle nozzle and converter tapping time. This reduces empty ladle baking time so as to achieve low consumption energy.3. Aiming at tackling the problems that solving main equipment scheduling model, a main equipment optimal scheduling method is proposed for steelmaking-continuous casting process. It consists of equipment assignment algorithm based on dynamic program (DP) technique and conflict elimination algorithm based on linear program (LP) technique. The LP based conflict elimination algorithm was composed of a setting penalty coefficients for conflict elimination model by an orthogonal design approach and to determine the final program schedule based on human-computer interaction. Consider that the ladle matching model is developed difficultly, by transforming the knowledge of the ladle experts into rules, the well-known rule-based reasoning (RBR) technique is adopted to select a high temperature ladle for every charge and to ensure the normal execution of scheduling plan.4. Simulations are carried out for the proposed optimization schedule of main equipment approach through actual data consisting of3casts plan (7heats,3refining modes) Simulation results show that the average waiting time in the cast plan is3minutes, which is shorter than the average waiting time in the artificial schedule by16minutes. Time needed to produce the scheduling is3seconds which is obviously shorter than that needed in the artificial scheduling by7minutes. The maximum waiting time for all the heat is10minutes that is shorter than its counterpart in the artificial scheduling by26minutes. An experiment on the efficiency comparison between the proposed algorithm whose penalty coefficients are obtained by the orthogonal design and the existing algorithm based on the artificial trial and error method is carried out using the above data. The results indicate that the overall waiting time for the algorithm based on the former is21minutes, shorter than28minutes for the latter algorithm. There is no break between different heat in the same cast plan and the cast plan is carried out punctually, while there is a delay (about4minutes) between the actual casting time and expected one.5. Several industrial experiments for the proposed approach are taken in a large steel plant, where its continuous casting process consists of three converters, seven refining furnaces, three casters and three refining modes. The schedule is created by the proposed method for10groups of industrial data, every group including3cast plans,20-23heats,3refining modes. The results are demonstrated as follows:The average time required to generate the schedule is6seconds which is much shorter than1minutes needed by the steelmaking plant. For the various groups of industrial test cases, the maximum waiting time of the heats in front of the caster do not exceed10minutes. Maximum of deviation to the targeted casting time is16that lower14minutes than the requirement of the steelmaking plant. The approach guarantees the heats of each cast plan continuous casting on the same caster. Moreover, the requirement for high temperature ladle can also be met due to the ladle matching function.