基于自适应遗传算法的冷连轧轧制力模型自学习

为提高轧制力模型的设定精度,以包含冷连轧带钢轧制过程变形区金属塑性变形和入口、出口弹性变形的Bland-Ford-Hill模型作为冷连轧轧制力模型,提出利用改进自适应遗传算法对平均变形抗力和摩擦系数进行寻优搜索,得出满足实际轧制力精度的平均变形抗力和摩擦系数,进而通过指数平滑法计算出平均变形抗力和摩擦系数的自学习系数,实验结果表明,该模型自学习后轧制力的设定精度可以满足在线控制的要求。

Adaptive Learning of Rolling Force Model Based on Adaptive Genetic Algorithm in Tandem Cold Rolling

To improve the precision of rolling force model in tandem cold rolling,the Bland-Ford-Hill model was used for rolling force model in which the plastic deformation in bite zone and elastic deformation at entry and exit of bite zone were considered.Aiming at the actual precision of the rolling force,an improved adaptive genetic algorithm was proposed in order to search the deformation resistance and the friction coefficient.Further, the adaptive learning coefficient of the deformation resistance and the friction coefficient could be obtained by way of exponential smoothing average.The experiment results showed that the accuracy of the actual value of rolling force could meet the requirement of on-line process control.