Optimization Methods For Steel-making And Continuous Casting Production Scheduling With Time-of-use Electricity Price

The steel-making and continuous casting production process is an important link for the production of iron and steel enterprises,with the characteristics of multi-stage,multi-machine as well as discrete and continuous production combined.The production scheduling optimization thereof can effectively link up the various procedures of the iron and steel enterprises,accelerate the production rhythm,improve the production efficiency,and reduce the production cost,so it is always the hot issue concerned by the business circle,and the research field.The type of energy consumed in the steel-making and continuous casting production process is varied,during which,there will be huge power consumption,and the power cost is closely related to time-of-use(TOU)electricity price.Theoptimization of power consumption and power cost under TOU electricity price can reduce the total energy consumption and energy cost in the production process,and improve the economic benefits of the enterprises,possessing important theoretical research significance and practical application value.In this dissertation,the complicated issue of optimizing the power consumption and power cost under TOU electricity price for the steel-making and continuous casting production scheduling is researched,which includes the cases of identical technological route,complicated technological route,uncertain processing time,and multiple objectives,etc.A scheduling optimization model for minimizing the power consumption and power cost under TOU electricity price in the steel-making and continuous casting production process with identical technological route is established.For the sharp increase of 0-1 variables after considering TOU electricity price in the model,the complicated target calculation,the slow solving speed and other issues,a kind of hybrid heuristic cross entropy(HHCE)algorithm based on local search is proposed.The HHCE algorithm includes matrix coding strategy and backward decoding method based on stages,and combines the strategy of hybrid generation of samples based on FIFO heuristic rule with the strategy of local search with rows swap and columns swap.It can obtain a high-qualified schedule within a relatively short time and has good stability and convergence.Simulation results indicate that,comparing with the optimization of residence time of charges for reducing power consumption indirectly in the steel-making and continuous casting production process,this optimization model can exert a better effect,especially for optimizing the power cost under TOU electricity price.A scheduling optimization model for minimizing the power consumption and power cost under TOU electricity price in the steel-making and continuous casting production process with complicated technological route is established.Comparing with the case of identical technological route,the coding strategy and decoding method of cross entropy algorithm incurred by complicated route constraint are more difficult.The consideration of TOU electricity price enlarges the decision variable scale in the model by three times at least,and makes the target calculation more complicated and the model solving more difficult.Therefore,a kind of hybrid adaptive cross entropy(HACE)algorithm based on dynamic parameters is proposed.The HACE algorithm adopts backward decoding method based on operations,and combines the strategy of hybrid generation of samples based on global selection and random permutation heuristic rules,the strategy of local search with matrix partition and rows swap and columns swap,with the strategy of dynamic parameters adjustment.It has high solving quality,fast solving speed,and strong adaptive ability.Simulation results indicate that,this optimization model can effectively describe the more complicated large-scale steel-making and continuous casting production process.And with respect to the optimization of power consumption and power cost under TOU electricity price,it can exert a better effect than merely optimizing the residence time of charges.For the uncertain LF refining time,and the uncertain basic processing time of charges and other uncertainties in the steel-making and continuous casting production process,a scheduling optimization model for minimizing the power consumption and power cost under TOU electricity price is established.Since integer variables and their constraint conditions increase largely,and the LF refining time shall be adjusted,this model has more decision variables,bigger scale,and it is more difficult to solve.Thus,a cascade cross entropy(CCE)algorithm mixed with the discrete and continuous cross entropy algorithms is proposed.In the CCE algorithm,the uncertain processing time and the machine assignment for charges are respectively solved,which can simplify the coding and decoding process of the algorithm,reduce the quantity of infeasible solutions,avoid several issues occurred in the genetic algorithm,including too long chromosome,complicated crossover and mutation operations,and difficult decoding,and then reduce the solving time.Moreover,a hybrid adjustment method based on critical charge is proposed to adjust the LF refining time of some charges,so as to compensate for their temperature loss,and simultaneously reduce the power consumption and power cost increased under TOU electricity price.Simulation results indicate that,comparing with the solving results of random instances and special instances with certain processing time,this model can optimize the combination of uncertain processing time,which is reasonable and effective in reducing the power consumption and power cost.Finally,to solve the problem of tough decision on the contradictions and mutual influence among power consumption,power cost,make span and other objectives after considering TOU electricity price in the steel-making and continuous casting production process,a multi-objective scheduling optimization model is established.For the issue of varied types and complicated calculation of objectives in the model,the issue of difficult individual ranking and evaluation,bad variety and relatively concentrated distribution of solving result and other issues,a hybrid multi-objective cross entropy(HMOCE)algorithm with Pareto optimal is proposed.Combining strategies of the hybrid and multiple generation of samples,the individual evaluation based on fast and non-dominated ranking,the diversity holding based on crowding distance and elite sample cluster and other strategies,the HMOCE algorithm can obtain a good solving effect.Especially for the introduction of clustering algorithm,it can effectively avoid the non-dominated solutions gathering to the middle part of the Pareto front,so improves the diversity and distribution universality of non-dominated solutions.Simulation results indicate that,this model and its solution can not only provide a compromise schedule that balance various objectives for the scheduling decision makers,but also provide a preferred schedule inclined to a certain objective without worsening other objectives on the whole.