This RFCS project aims to strengthen the European stainless steel industry by developing a technology-based assistance system for production scheduling and planning. “Technology-based” means that input-output models are derived to predict transfer/final quality in terms of flatness, cross-bow, thickness tolerance, elongation and other quality parameters, based on process data gained from topometrical measurement systems that will be temporarily installed at entry and exit of each plant.
The prediction models identified and suitable quality indices to be defined are then included in the optimisation task of production planning for the production chain: cold rolling (four plants) ---> annealing (three plants) ---> temper rolling (three plants).
This approach is completely new, i.e. it is not included in traditional production systems available at the market.
Based on collected process data, the algorithms for calculating the quality indicators will be developed. Different effect (weight) of each quality parameters with respect as to a safe cold processing as for respecting the final product destination will be provided. Furthermore, by considering, as mentioned above, the different effect of each single quality parameter, an overall indicator will be derived.
The following advantages can be envisaged:
• “Standardisation” of different quality parameters (i.e. camber and crown);
• Immediate detection of the coil quality status;
• Input for multi-object optimisation in the cold rolling area.
Several strategies for Multi-Objective Optimisation (MOO) will be investigated. Taking into account also the constraints on the computational time required to reach a feasible solution, simple approaches will be tested for first consisting in a combination of the different objectives into a single objective function.