Dynamic Multi-objective Operation Optimization Of Continuous Annealing Production Process In Iron And Steel Enterprise

The task of operation optimization of production process is to determine optimal values of control variables in the process control layer to meet the production plan based on the production process constraints.It has an important impact on the product quality and the energy consumption.In the continuous annealing process of iron and steel enterprises,the key control parameters are usually set based on manual experience.Since the continuous annealing production has complex processing,many optimization objectives and dynamic environment changes,problems of high energy consumption and frequent fluctuations of strip quality appears frequently.To solve the above problems,in this thesis a dynamic multi-objective operation optimization model is established based on a large-scale iron and steel enterprise's continuous annealing production process,so as to optimize three conflicting objectives including production quality,unit capacity and energy consumption.According to the model's characteristics of dynamic,multivariable and multi-objective,a memory mechanism based hybrid multi-objective genetic algorithm is developed to solve the model.Detailed research contents are as follows.(1)For the operation optimization model of continuous annealing production,three optimization objectives of production quality,unit capacity and energy consumption are determined through practical investigation.Through the analysis of mechanism of the continuous annealing production process,the operational variables directly related to the optimization objectives and the process constraints are determined.Based on the analysis of dynamic factors in practical production,the multi-objective dynamic operation optimization model of the continuous annealing production process is established.(2)For the initial setting of operation variables at the beginning of production which is a static operation optimization problem,a hybrid multi-objective genetic algorithm is developed.Based on the multiple crossover operators,a new population generation method based on the personal best archives is proposed to improve the population quality by making effective use of previous search results.A parent solution selection method with guidance mechanism is proposed to speed up the convergence speed and keep the dispersion of population at the same time.Computational results based on benchmark problems and practical problems show the effectiveness of the proposed strategies and algorithm.(3)To handle the dynamic changes of production environment,a hybrid multi-objective genetic algorithm based on memory mechanism is proposed.Whenever the production environment changes,the population will be re-initialized based on the memory mechanism.The re-initialized population makes effective use of previous search results by selecting solutions from the previous external archives with Pareto ranking method,which can help to improve the convergence speed of the algorithm in the new environment.Computational results of simulated problem show that the proposed algorithm can get the new setting of operation variables in the new environment quickly and has good adaptability to environmental changes.