Please use this identifier to cite or link to this item:
http://hdl.handle.net/11375/23376
Title: | Dynamic Optimization, State Estimation and Control of Electric Arc Furnace Operation |
Authors: | Shyamal, Smriti |
Advisor: | L.E. Swartz, Christopher |
Department: | Chemical Engineering |
Publication Date: | Nov-2018 |
Abstract: | Electric arc furnaces (EAFs) are commonly used in the steel industry for production of steel by melting down the scrap metal and altering its chemistry. The highly energy intensive steelmaking operation is a complex batch process, and involves limited automation. The main aim of this research is to develop a mathematical model and computational framework for on-line optimization-based decision support and control of EAF operation. A dynamic model of the EAF process is implemented and used within an optimization framework to determine the optimal input trajectories. Industrial implementation of the model-based optimization systems is envisaged to generate significant savings through reduced consumption of electric power, natural gas, oxygen, carbon and fluxes such as limestone and dolomite. In the first part of the study, we present the advances made through this research toward implementing an economic optimization of EAF operations. The three key building blocks for a real-time implementation are discussed: a dynamic model, dynamic simulation and dynamic optimization. We also present the simulation and optimization deployments for the industrial members of the McMaster Steel Research Centre (SRC) through which users can interact with the model. In the second part of the study, state estimation is investigated through implementing multi-rate Moving Horizon Estimation (MHE) for real-time model calibration. A parameter estimation based multi-rate MHE framework is developed to handle measurements with different sampling rates. The MHE application is further extended for the full batch time with the use of a simultaneous solution approach. A novel MHE initialization scheme is also proposed to reduce the numerical computation time. The estimator showed strong performance in tracking the internal states in the presence of plant-model mismatch and measurement noise, improving from poor initial guesses of the states. The multi-rate MHE is then coupled with an economics-based dynamic optimizer to form an online decision support tool. %(DST). The DST is able to reconstruct the states and provide optimal decisions to operators in real-time despite the use of a nonlinear large-scale EAF model. In the third part of the study, an energy management approach is introduced that effectively curtails the energy cost in real-time through the implementation of economically optimal operating decisions. An economics-oriented shrinking horizon nonlinear model predictive control (NMPC) algorithm that exploits time-varying electricity prices is coupled with a multi-rate MHE to form an integrated decision-making framework. A novel initialization scheme is also developed for obtaining fast on-line solutions of the economic NMPC and multi-rate MHE dynamic optimization problems. The energy usage optimization results indicated a significant reduction in the operating cost and peak electricity demand compared to the case where the electricity price profile is not updated. In the fourth part of the study, a real-time dynamic optimization-based advisory system that employs a first-principles EAF model is introduced to support the operator decision making in real-time. Economically optimal process operation is achieved by employing the first-principles dynamic EAF model in the optimization formulation. A dynamic optimization calculation can be triggered by the operator at any point in the batch, an action that can be repeated multiple times during the batch. The advisory system incorporates a multi-rate MHE that continually computes estimates of the process states utilizing current and past inputs and measurements. End-point constraints and potential extension of the batch duration are handled through a multi-tiered optimization algorithm. Our case studies demonstrate a major economic improvement when the dynamic optimization-based advisory system is used. |
URI: | http://hdl.handle.net/11375/23376 |
Appears in Collections: | Open Access Dissertations and Theses |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
Shyamal_Smriti_201806_PhD.pdf | 6.29 MB | Adobe PDF | View/Open |
Items in MacSphere are protected by copyright, with all rights reserved, unless otherwise indicated.