Rolling force model is the core of the rolling process control system for wide strip tandemcold rolling mill, which is one of the important factors determines the strip thickness and flatnessquality. Therefore, it is very significant to study on rolling force prediction model and the way toimprove its precision for automating rolling process.Based on2080cold tandem mill of wisco PCM2as the research object, combined with theactual production to establish the mathematical model of rolling force for cold tandem rollingand and the scheme of improving the precision of rolling force prediction model was tested byindustrial experiment. Meanwhile, combining with the actual production i analysis the variousfactors affecting the rolling force, and to establish the mathematical model of rolling force;Thenthe deformation resistance model, friction coefficient model and their self adaptive approachwere built. And use the rolling force, forward slip, etc can be detected before rolling factors,through BP neural network to correct the reference point effect coefficient;At last a method toimprove the precision of rolling force prediction model is put forward and online test wasconducted. The result shows that the mean rolling force predicting deviation is improved from9.79%by the old model to6.7%by the new one,which Achieved good improvement effect andhas created significant economic benefits for the enterprise. |