基于改进PSO的中厚板轧制规程能量优化设计

Energy optimization of medium plate rolling schedule based on improved particle swarm optimization algorithm

基于中厚板轧制规程优化是多个连续决策变量的复杂非线性优化问题,以最小化轧制能耗为目标,建立轧制规程的能量优化设计模型;基于粒子群优化算法(particle swarm optimization,PSO)进行轧制规程优化设计,通过规划最佳道次板厚分配实现轧制能耗目标最优;针对中厚板轧制生产特点及其对算法全局搜索能力和收敛速度的要求,提出一种改进PSO,通过构造新的惯性权重和加速系数非线性调整策略,平衡算法不同阶段的粒子全局搜索和精确搜索能力,引入局部平均值的同时构造非线性局部加速系数,以减小对收敛速度的影响。研究结果表明:改进PSO应用于轧制规程优化,能耗明显降低,全局搜索能力强且收敛速度快,综合性能比遗传算法、标准PSO及几种常用改进算法的优。

Considering that optimization of medium plate rolling schedule is a complex nonlinear optimization problem, optimization model for minimizing the energy consumption during hot rolling was built with regard to multiple continuous variables, and the best allocation of reduction ratio among rolling passes was planned by particle swarm optimization （PSO） algorithm to achieve the optimal target. For the characteristics of medium plate rolling and the algorithm requirements for global search ability and convergence speed, an improved PSO was proposed, in which a new strategy for adjusting the inertia weight coefficient and the acceleration coefficient nonlinearly was constructed to balance the global and accurate search ability at different stages, while introducing local mean of particle velocity, a nonlinear local acceleration coefficient was also constructed to reduce the side effects on convergence speed. The results show that the energy consumption of the optimized rolling schedule reduces significantly, and considering that global search ability and convergence speed, the overall performance of the improved PSO is better than that of GA, standard PSO and several other improved PSO algorithms.