摘要
A neural network friction model with very good prediction accuracy was developed on the basis of industrial data. Simulative calculation indicated that the accuracy of rolling force and motor power calculation can be improved using the calculated friction coefficient. It was found that reduction ratio and deformation resistance of strip has more effects than other parameters, and the effects of most parameters are affected by rolling speed.
A neural network friction model with very good prediction accuracy was developed on the basis of industrial data. Simulative calculation indicated that the accuracy of rolling force and motor power calculation can be improved using the calculated friction coefficient. It was found that reduction ratio and deformation resistance of strip has more effects than other parameters, and the effects of most parameters are affected by rolling speed.
