基于支持向量机的轧制过程灵敏度分析

利用支持向量机(SVM)建立轧制力模型,并由该模型分别对各输入变量进行偏微分,以计算轧制过程出口厚度灵敏度系数,从而解决解析方法难以求解的轧制过程模型“代数环”问题．对轧机第一机架试验仿真,结果表明,基于SVM方法获得的灵敏度具有很高的精度,且由灵敏度确定的控制量可以获得很好的控制效果,同时也表明第一机架辊缝对出口厚度的影响比张力要大,以辊缝控制方式调节出口厚度比张力控制方式具有更高精度．

Sensitivity Analysis for Rolling Process Based on Support Vector Machine(SVM)

This paper presented a method for the calculations of the sensitivity factors for rolling process which is obtained by differentiating the roll force model based on SVM. It can eliminate the "algebraic loop" of the analytical model for rolling process. The simulations in the first stand of five stands tandem rolling mill indicate that the calculation for sensitivity by the proposed method can obtain a good accuracy, and the appropriate adjustment on the control variables determined directly by the sensitivity has a high compensation performance. Moreover, the roll gap has the larger effect on the exit thickness than both front tension and back tension, and it is more efficient to select the roll gap for control variable of the thickness control system in the first stand.