基于万有引力算法的铝热连轧规程优化设计

Aluminum Strip Hot Rolling Schedule Optimization Design Based on the Gravitational Search Algorithm

提出了以预防打滑和平衡4个机架功率为目的的多目标函数，建立了一种基于支持向量回归机的轧制力模型来取代传统数学模型，有效地提高了轧制力预报的精度。同时采用万有引力算法对轧制规程进行优化，该算法不仅能够防止陷入局部极值点而且避免了盲目搜索，大大提高了进化速度。该规程应用在河南某厂“1＋4”铝热连轧改造现场的精轧部分，轧制规程效果理想。

A reasonable rolling schedule is a strong guarantee for the quality of aluminum strip and the low power consumption of mill. this procedure set put forward multi-objective function to prevent skidding and balanced four mill power, establish a rolling force model based on Support Vector Regression to replace the traditional mathematical model and improve the accuracy of the prediction of the rolling force effectively. At the same time, it uses Gravitational Search Algorithm to optimize the rolling schedule. The algorithm can not only be able to prevent falling into local extremum point, but also avoid a blind search. So it improved the speed of evolntion greatly. This procedure used in the finishing part of“ 1 ＋ 4” aluminum hot strip in Henan province, the rolling procedure effect is ideal and it has a guiding significance for the secondary setting calculation.