Application of Model Predictive Control to Finishing Auto Width Control Based on System Identification in Hot Strip Mills

In hot metal rolling mill system, achievement of strip width accuracy is important issue. However, temperature fluctuation caused by skid-marks in a furnace deteriorates the strip width quality. System identification is a statistical method to build mathematical models of dynamic systems from measured input-output data, and it is expected to be one of the most useful methods of modeling for big data. By using system identification, the dynamics of width fluctuation in bar was predicted on-line, and it is compensated for by model predictive control. As a result, the accuracy of width control is successfully improved.