Steel production via electric arc furnaces (EAFs) is a very energy-intensive process that accounts for almost 25\% of the total crude steel production worldwide. In modern steelmaking, finding an economically beneficial mode of operation that reduces the energy consumption of the plant and the environmental impact of the process is the priority. To address this challenge, this thesis presents a model-based optimization strategy that can reduce the energy demand of the EAF process. First, mathematical models of an electric arc and of an electric arc furnace were developed and used to answer two fundamental questions: a) How do the electrical set-points of the furnace affect the geometry of the arc and the heat exchange between the arc and the metal phases in the furnace?, and b) How do various operative set-points affect the melting dynamics of the process?. The developed models were validated using experimental and process data of an industrial ultra-high-power EAF. Second, a dynamic optimization framework (DO) with the goal to minimize the electrical losses of the process is proposed. Two important operational questions were addressed: a) What is the optimal operation strategy that reduces the energy demand of the process?, and b) How can dynamic optimization and scheduling be integrated to achieve an optimal operation of the steelmaking plant?. The DO problem is solved using a control vector parametrization strategy that computes an optimal input trajectory for a batch of steel that consists of several charges. The computed control policy was tested in an industrial EAF, and the energy consumption of the process was reduced by 4.5% for a family of steels.weiterlesen