Physics-Based Modeling of Electric Operation, Heat Transfer, and Scrap Melting in an AC Electric Arc Furnace
- PDF / 1,435,474 Bytes
- 15 Pages / 593.972 x 792 pts Page_size
- 53 Downloads / 232 Views
NTRODUCTION
ELECTRIC arc furnaces (EAF) are used to melt steel out of steel scrap or direct reduced iron. In 2013, electric arc furnaces contributed with approximately 30 pct to steel production worldwide (in Reference 1, p. 16). The main power inputs are electric energy from the grid and energy from fossil fuels. Due to its high total energy demand of approximately ð650 . . . 850Þ kWh=t,[2] engineers and researchers have been trying to optimize the EAF process since the beginning of its industrial usage. While approaching a minimum threshold in energy input in the 1990s (a value of approx. 400 kWh=t of total energy is the physical limit),[3] the interest in developing predictive models of the process grew. Such models could be used to predict the energy demand, test different modes of operation or examine the effect of optimization measures. In 1992, Ko¨hle published a statistical model of the EAF process, which determines the energy demand by a regression analysis.[4] This analysis is based on a data collection among the operators of fourteen furnaces in Germany. The main determining variables identified in this work are operational parameters like weight of charge, tapping temperature, duration of heat, and so on. In further papers, the model was enhanced and adapted to new data by Ko¨hle and co-workers.[5–7] Ko¨hle’s analysis was highly recognized in literature and further researcher applied the method to their data or FLORIAN OPITZ, Scientific Assistant, and PETER TREFFINGER, Professor, are with the Offenburg University of Applied Sciences, Badstr. 24, 77652 Offenburg, Germany. Contact e-mail: florian.opitz@ hs-offenburg.de Manuscript submitted October 5, 2015 METALLURGICAL AND MATERIALS TRANSACTIONS B
compared the findings to theoretical thermodynamic considerations.[8–10] Because of their very own nature static statistical models can not be used for prediction of a single heat, but rather to compare the overall performance of a furnace to average values. Due to this limitation of statistical models, dynamic physical models have been developed. All these models describe the processes in the vessel by means of balancing equations for mass and energy between a limited number of control volumes. Cameron et al.[11] introduced a model with four phases (gas, metal, slag, and organic solid) and six interfaces. Energy and mass transfer between the phases are calculated at the interfaces. Electric energy input and heat losses are need to be given as input data from measurements, whereas no fitting parameters are needed. The model is validated by comparison of the simulated off-gas composition to plant data. Bekker et al.[12] focused on the generation of carbon monoxide in order to implement a real-time dynamic control in the off-gas system. Their model comprises five phases (solid scrap, solid slag, liquid slag, molten metal, and gas phase), for which heat and mass transfer are calculated by 14 ordinary differential equations. Eight fitting parameters are used. The off-gas mass flow rate and the electric arc power are given as input v
Data Loading...
