In the current steel industry, it is very common to add micro-alloys into steels. In deforming the steels with high deformation and lower temperature, the mechanical properties of the steel are greatly increased while the costs of steels are only slightly increased.

In this process, however, it is almost necessary to add metallurgical models into the conventional Level 2 model, because the microstructure changes incurred during steel rolling, such as delayed incrystallization and metallurgical changes in the holding of steels in multi-pieces rolling.

Due to changes in chemical compositions and controlled rolling processes with lower temperature and higher reduction, model factors are changed in the process. So in a rolling mills, over 40,000 sets of model factors should be redesigned for all the steel grades. This set of models take into account of metallurgical phenomena such as delayed recrystallization and thus induced residual strain due to micro-alloying elements application. This procedure has also removed learning logics defects existing in almost every Level 2 package in the market.

It can also be provided in a set of simple, easy, inexpensive way of Level 2 System upgrade, in which the above-mentioned model factors can be integrated into the Level 2 System. This way of upgrade does not require major modification of the Level 2 source code, but can integrate large number (over 40,000 sets) of proven, high-quality models into the current Level 2 System, by using our Guided Two-Parameter Learning.

After the completion of the model upgrading with thus model accuracy increase, further higher-end improvements could be proceeded, such as plate shape improvement and rolled steel property prediction, so a Quality Management System can be developed.

The large number of models have been developed on the basis of the rolling process and product models in the last 30 years during the work in Germany and USA, etc.