A dynamic, first-principles process model for a steelmaking electric arc furnace has been developed. The model is an integrated part of an application designed for optimization during operation of the furnace. Special care has been taken to ensure that the non-linear model is robust and accurate enough for real-time optimization. The model is formulated in terms of state variables and ordinary differential equations and is adapted to process data using recursive parameter estimation. Compared to other models available in the literature, a focus of this model is to integrate auxiliary process data in order to best predict energy efficiency and heat transfer limitations in the furnace. Model predictions are in reasonable agreement with steel temperature and weight measurements. Simulations indicate that industrial deployment of Model Predictive Control applications derived from this process model can result in electrical energy consumption savings of 1–2%.

No abstract

No abstract

No abstract

This paper studies the problem of multi-plant manganese alloy production. The problem consists of finding the optimal furnace feed of ores, fluxes, coke, and slag that yields output products which meet customer specifications, and to optimally decide the volume, composition, and allocation of the slag. To solve the problem, a nonlinear pooling problem formulation is presented upon which the bilinear terms are reformulated using the Multiparametric Disaggregation Technique (MDT). This enables global optimisation by means of commercial software for mixed integer linear programs. We demonstrate the model and solution approach through case studies from a Norwegian manganese alloy producer. The computational study shows that the model and proposed optimisation approach can solve problem sizes of up to ten furnaces to a small optimality gap, that global optimization approach with MDT scales well with larger, real problem instances, and that the model outperforms the current operational practice.