Research And Application Of Hybrid Particle Swarm Optimization Algorithm In Load Distribution For Tandem Hot Metal Strip Rolling

The traditional rolling schedule in the hot strip rolling process is generally based on the load distribution obtained by the empirical distribution method,which combines the existing data obtained from the same strip production of the same type of equipment in production with the production practice,and finally determines the rolling steel.Although the traditional calculation method is reasonable,it is not optimal.The rolling schedule must be continuously optimized in the production practice to achieve the goal.The load distribution model has the characteristics of nonlinear multi-variables,especially when considering multiple optimization objective,the traditional optimization methods such as dynamic programming method will cause difficulty in solving,leading to lower accuracy of the optimization results and difficulty in meeting production requirements.In the era of rapid development of artificial intelligence in the modern industry,continuous exploration should be carried out to establish a new load distribution model and use artificial intelligence optimization methods to obtain more reasonable rolling procedures.This paper studies the advantages and disadvantages of different artificial intelligence algorithms,combines the advantages of both genetic algorithm and particle swarm optimization algorithm,and proposes a hybrid particle swarm optimization algorithm that is hybrid particle swarm optimization algorithm based on cross mutation(HPSO).The improved particle swarm optimization algorithm can effectively overcome the problems of the slow convergence speed and premature convergence of the PSO algorithm by introducing multiple group concepts and crossover mutation operations.This study establishes a load distribution model that takes into account load balance and shape optimization,and proposes an optimization idea that combines optimization of rolling forces and optimization of profile.Using the standard particle swarm optimization(BPSO),inertial weighted adaptive particle swarm optimization(IPSO)and hybrid particle swarm optimization(HPSO)algorithms respectively,the initial value of load distribution was optimized and the optimized rolling schedule was finally obtained.Experiments show that the optimization results of the BPSO optimization algorithm model and the IPSO algorithm model all have different degrees of algorithm defects,resulting in slow convergence optimization and less accurate convergence of the optimal solution.The obtained rolling parameters do not meet the plate shape well.The HPSO optimization algorithm can effectively get rid of the local optimal value by introducing selection,crossover and mutation operators,avoid premature convergence,and thus find the best global optimal value easily.The optimized rolling parameters can ensure the load balance and good plate shape,thus ensuring the rationality and stability of load distribution to a large extent.Therefore,there is a great deal of research value in the application of actual rolling schedules.