(170e) Real-Time Energy Management for Electric Arc Furnace Operation

Electric arc furnaces (EAFs) are extensively employed by steel production companies to produce steel from scrap metal. The energy intensive batch operation involves melting of the scrap using electrical and chemical energy, and modifying its chemistry. Electrical energy is added via electrodes, while combustion reactions provide chemical energy. The melting continues and a flat bath of molten metal is achieved. Oxygen lancing leads to the formation of metal oxides due to oxidation, which floats on the top as slag. The slag chemistry is altered by flux additions (carbon, lime and dolomite), carbon and oxygen lancing. The amount and timing of input additions are determined mainly by the operators. Since EAFs are operated under very harsh conditions with limited measurements available during the heat, process automation is a challenging task. The composition of offgas is the only set of measurements generally available at a regular frequency. With electricity contributing approximately 60% of the energy requirement [1], external variations in electricity price needs to be considered for optimal energy usage. An energy-efficient control strategy is envisaged to generate significant savings due to an optimal use of electrical power, natural gas, oxygen, carbon and the fluxes in response to electricity price changes.

Control strategies have been developed in the past for the EAF electrode system [2], with linear model predictive control (MPC) for the offgas system [3]. The use of a first-principles model for nonlinear MPC assuming full state feedback was reported in [4]. Economic model predictive control, on the other hand, has become popular due to its straightforward handling of an economic target. The traditional set-point tracking and input move suppression objectives are either replaced by, or used in conjunction with, an economic objective [5, 6]. A time-varying stage cost in EMPC [7] is suitable for incorporating a changing electricity price profile. EMPC for continuous processes has been extended to include demand response (DR) [8]. Literature in which DR is considered for batch operation of EAFs is sparse. Real-time control applications require state estimation to work in tandem with MPC. Multi-rate moving horizon estimation (MHE) has been used for EAFs to reconstruct the states using both the fast and slow measurements [9]. A decision support tool for EAF operators based on a combined application of shrinking horizon dynamic optimization and multi-rate MHE was proposed in [10]. The study is extended here to employ a coupled EMPC-MHE implementation for EAFs to utilize time varying electricity cost for optimal control.

In this work, we aim to develop an EAF control strategy that effectively reduces the energy costs in real-time while exploiting the electricity price profile. A shrinking horizon EMPC is coupled with a multi-rate MHE to develop an integrated decision-making framework. The EMPC objective function comprises time-varying cost coefficients to handle a volatile wholesale electricity market. The control structure employs a modified first-principles differential-algebraic equation (DAE) model proposed in [11, 12]. The model divides the furnace into four zones (solid scrap, molten metal, slag-metal, gas) with mass transfer limitations across the zone interfaces. Chemical equilibrium is considered for the slag-metal and the gas zone through an embedded Gibbs free energy minimization. The complex DAE model is discretized and used within a simultaneous solution framework for solving the dynamic optimization problems. Online solve times of EMPC-MHE problems are critical for a real-time implementation of the control strategy. We introduce a novel initialization scheme to obtain fast solutions of EMPC-MHE optimization problems.

To evaluate the proposed strategy, we conduct case studies to investigate the performance of EMPC under various Day Ahead Market conditions, in which the electricity price changes every hour. The case studies are chosen based on a detailed analysis of actual and forecasted 2016-2017 prices for the Ontario (Canada) wholesale market. The EMPC is shown to balance the electrical and chemical energy consumption in an optimal way to generate cost savings. Moreover, the peak electricity demand during the batch operation is reduced when a time-varying electricity price profile is used. The multi-rate MHE demonstrates a strong tracking ability to support the EMPC. Finally, the computational results show a significant reduction in EMPC-MHE solve times when the proposed initialization scheme is used. In summary, an EMPC-MHE based strategy is presented for curtailing the energy costs in real-time. Current work includes formulations for emission reduction through constraints or a carbon credit-based approach. As a future work, we intend to study stochastic formulations for differentiating forecasts from actual prices. Additionally, the strategy will be tested and compared for 15 minute and 5 minute electricity markets. References

[1]	R. J. Fruehan, The Making, Shaping, and Treating of Steel: Ironmaking volume, vol. 2, AISE Steel Foundation, 1999.
[2]	B. Boulet, G. Lalli and M. Ajersch, "Modeling and control of an electric arc furnace," in American Control Conference (ACC), 2003.
[3]	J. G. Bekker, I. K. Craig and P. C. Pistorius, "Model predictive control of an electric arc furnace off-gas process," Control Engineering Practice, vol. 8, no. 4, pp. 445-455, 2000.
[4]	R. D. M. MacRosty and C. L. E. Swartz, "Nonlinear predictive control of an electric arc furnace," IFAC Proceedings Volumes, vol. 40, no. 11, pp. 285-290, 2007.
[5]	R. Amrit, J. B. Rawlings and D. Angeli, "Economic optimization using model predictive control with a terminal cost," Annual Reviews in Control, vol. 35, no. 2, pp. 178-186, 2011.
[6]	M. Ellis, H. Durand and P. D. Christofides, "A tutorial review of economic model predictive control methods," Journal of Process Control, vol. 24, no. 8, pp. 1156-1178, 2014.
[7]	D. Angeli, A. Casavola and F. Tedesco, "Theoretical advances on Economic Model Predictive Control with time-varying costs," Annual Reviews in Control, vol. 41, pp. 218-224, 2016.
[8]	J. Feng, A. Brown, D. O'Brien and D. J. Chmielewski, "Smart grid coordination of a chemical processing plant," Chemical Engineering Science, vol. 136, pp. 168-176, 2015.
[9]	S. Shyamal and C. L. E. Swartz, "A Multi-rate Moving Horizon Estimation Framework for Electric Arc Furnace Operation," IFAC-PapersOnLine, vol. 49, no. 7, pp. 1175-1180, 2016.
[10]	S. Shyamal and C. L. E. Swartz, "Optimization-based Online Decision Support Tool for Electric Arc Furnace Operation," in IFAC World Congress, Toulouse, France, 2017 (accepted).
[11]	R. D. MacRosty and C. L. E. Swartz, "Dynamic modeling of an industrial electric arc furnace," Ind.Eng.Chem.Res., vol. 44, pp. 8067-8083, 2005.
[12]	Y. Ghobara, "Modeling, Optimization and Estimation in Electric Arc Furnace," Master's thesis, no. McMaster University, 2013.