Coke Production Process Based On Multi-objective Optimization Model To Optimize The Methods Of The Research And Application

The coking plant process is a very complex industrial process. The features of strong multi-variable coupling, large time delay, multi-constrained are the major challenges, what'more, as lack of the relationship design among the processes, the comprehensive production targets can not be achieved. Thus, the most critical problem for the the coking plant process optimization is how to adjust process operation parameters in the production process.On the basis of the intensive analysis of coking process technology, an optimization strategy is proposed for the process optimization problem. That is the local optimization targets are added between comprehensive production targets and the process operation parameters. First, the correlation model between the comprehensive production targets and the local optimization targets are established, so as to map the comprehensive production targets to the local optimization targets. Second, the local optimization targets are set as values in the process control systems and decision variables in the optimal scheduling system. Only need to adjust process operation parameters in each process control systems to reach the local optimization targets. According to this optimization strategy, a process optimization idea is proposed, which is with the maximum coke yield and minimum energy consumption as the optimization objective, with coke quality and technological requirements as the constraint conditions and with the target flue temperature, collector pressure and coking time as the decision variables.Firstly, according to the mechanism analysis and artificial experience, qualitative analysised the influencing factors of the comprehensive production targets, then quantitative calculation the gray colleration degree which is the major influencing factors of comprehensive production targets by the method of gray colleration analysis, thus the input variable of multi-objective optimization models are determined. Secondly, building the Netural Netwrok multi-objective optimization models for abnormal situation and building the multiple regression multi-objective optimization models for normal situation. Finally, the MOGA algorithm and the LWS+GRG algorithm are applied to solve this multi-objective optimization problem, the optimization results, that is local optimization targets, are set as values in the process control systems and decision variables in the optimal scheduling system. So the seamless connection is finished beteen optimization system and the subsystem.The coking plant process optimization software is implemented with VC++. Application results show that the multi-objective optimization models under normal condition and abnormal condition have high precision which are builded in this paper, MOGA algorithm and LWS+ GRG algorithm have global convergence and high optimization effective rate. The labour intensity is reduced after optimization software running, and the optimized result can provied better operater guiding for action coking plant, and better meet the enterprise needs.