Optimizing the melting process at AC-EAF with neural networks

摘要

Market pressures are forcing steelmakers to utilize more cost-effective production methods. Significant gains in process control and automation systems in the steel plant. In keeping with its ongoing innovation program, Stahlwerk Bous GmbH has invested in a neural network supported system for optimal control of energy input into its electric arc furnace at the Bous steel plant. Artificial intelligence in process automation and process optimization has brought a number of significant benefits to developers and operators alike. Dynamic electrode control of the electric arc furnace has great potential for increases in performance and efficiency of the process. Using the innovative hybrid furnace model Simelt NEC, developed by the Industrial Projects and Technical Services Group of Siemens AG, Erlangen, Germany, as a basis, it will be shown how the operation points of the electrode controller can be determined continuously with the aid of neural networks. The overall optimization target is to maximize the effectiveness of the melting process according to set optimization criteria. The optimization procedure can be used with any AC electric-arc furnace (EAF). The system is easy to maintain, and can be modified to suit specific operating requirements.