Accepted Manuscript Title: Optimal Operation Strategy and Gas Utilization in a Future Integrated Steel Plant Author: Hamid Ghanbari Frank Pettersson Henrik Sax´en PII: DOI: Reference:
S0263-8762(15)00244-0 http://dx.doi.org/doi:10.1016/j.cherd.2015.06.038 CHERD 1949
To appear in: Received date: Revised date: Accepted date:
14-8-2014 23-5-2015 30-6-2015
Please cite this article as: Ghanbari, H., Pettersson, F., Sax´en, H.,Optimal Operation Strategy and Gas Utilization in a Future Integrated Steel Plant, Chemical Engineering Research and Design (2015), http://dx.doi.org/10.1016/j.cherd.2015.06.038 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Optimal Operation Strategy and Gas Utilization in a Future Integrated Steel Plant Hamid Ghanbari1,*, Frank Pettersson1,2 and Henrik Saxén1 1
2
ip t
Thermal and Flow Engineering Laboratory, Department of Chemical Engineering, Åbo Akademi University, Biskopsgatan 8, FI-20500 Åbo, Finland Process Design and System Engineering, Department of Chemical Engineering, Åbo Akademi University, Biskopsgatan 8, FI-20500 Åbo, Finland
cr
*
[email protected], Tel: +358-2-215-4438, Fax: +358-2-215-4792
us
Highlights:
Integration of steelmaking with a polygeneration plant can reduce carbon dioxide emission from the system.
There is a trade-off between economic profit and environmental impact that can be tackled by flexible operation.
Combination of blast furnace top gas recycling and blast oxygen enrichment may reduce steelmaking emissions.
Analysis of the system under periodic external power demands provides robust states of operation.
M
d
Abstract
an
Ac ce p
te
In this work future perspectives of primary steelmaking are numerically studied with the aim to find ways to increase the sustainability of this industrial sector. The key options studied are emerging blast furnace operation technologies combined with carbon capturing and utilization units and integration with a polygeneration system producing district heat, electricity and methanol. A mathematical model is developed using the suggested superstructure to optimize the use of residual gases minimizing the internal energy demand under specified operating costs, simultaneously considering investment costs for new process units. The results of the study, which illustrate both the optimal operation of the blast furnace and the required unit processes for utilization of the residual gasses in the plant, provide guidelines on how this industrial sector can be developed in the future to considerably reduce harmful emissions and to make maximum use of raw materials. A large number of scenarios were studied and the net present value, steelmaking costs, specific emissions and methanol production in the optimal states were analyzed. The results reveal the optimal technology for gas treatment under periodic optimization considering a varying seasonal external energy demand. It demonstrates that the net present value of the system for a time horizon can be increased and that the CO2 emissions from the system can be reduced by up to 30 % by an optimal design and flexible operation of the system.
Keywords: Blast furnace, Optimization, Carbon capturing and utilization, Polygeneration system
1. Introduction Steelmaking is an important industry in many countries, but it is also a very energy intensive sector with a specific energy consumption of about 20 GJ/t of steel (World Steel Association, 2013). In most steel plants residual gases and heat are (partly) recovered to provide electricity and heat for internal and external demands.
1 Page 1 of 23
Steelmaking is still responsible for a big share of the industrial carbon dioxide emissions, which according to EU policy should be reduced by at least 80 % from the emission levels of 1990 (European Climate Foundation ECF Roadmap 2050, 2010).This policy commits industry to adapt to energy efficient, low carbon operation. Much research has been undertaken to increase the energy efficiency and to suppress carbon dioxide emissions from steel manufacturing. Carbon dioxide emission reduction methods in conventional steelmaking may be categorized into 1) use of alternative reducing agents, such as carbon neutral or low-carbon carriers, and 2) reduction of CO2 emissions that are inevitable in the process by capturing CO2 followed by sequestration.
us
cr
ip t
The Ultra-Low Carbon Dioxide Steelmaking (ULCOS) project is a major research and development effort carried out within Europe since 2004. A selected scenario, top gas recycling and oxygen enrichment in the blast furnace, has been studied theoretically and implemented in the experimental blast furnace in Luleå, Sweden (Danloy et al., 2009), where a capturing process was chosen to remove carbon dioxide for sequestration and treatment of residual top gases before they were injected back into the blast furnace as reducing agents (Birat, 2009). Application of such technologies in industrial scale requires investment in a CO2 capturing plant. In practice, a future implementation would depend on legal issues (e.g., whether the CO2 can be pumped in underground storages) and the costs of carbon dioxide emissions, transportation and sequestration. A more sustainable option may be a stronger integration of steelmaking with other industry and society, for optimal use of residual gases, by-products and energy.
te
d
M
an
Researchers have used different process integration approaches to analyze the available potential for CO2 reduction in the steel industry. Optimization techniques have been used to minimize carbon dioxide emissions (Wang et al., 2008) and energy consumption (Grip et al., 2013) from current conventional steelmaking. By developing different technologies through the ULCOS program, further analysis has been done on energy performance of steel industry including CO2 capturing units (Kuramochi et al., 2012). Some of the prospective scenarios on energy efficiency and CO2 emission in the EU steelmaking have been introduced. Analyses on cost effectiveness of the implementation of some of the best available technologies have been also reported (Moya and Pardo, 2013). The sue of biomass-derived reductants has attracted attention in research as this is a possible way reduce CO2 emission from blast furnace steelmaking (Suopajärvi et al., 2014). Large volume of off-gases generated during steelmaking is today mainly used as fuels in the process units within the plant, but depending on cost of fuels and CO2 emission penalties, producing chemicals from the gases could also be an option (Lundgren et al., 2013).
Ac ce p
The economics of future scenarios for CO2 emission reduction in the iron and steelmaking sectors is the key driving force toward feasibility of alternative primary steelmaking routes (Fischedick et al., 2014). An investigation on cost of energy saving and CO2 reduction in China shows that some of the technologies may not be cost effective in the current situation based on the fuel price and estimated CO2 cost (Li and Zhu, 2014). This would be an even more important challenge for steel industry within EU, where the role of restrictions and different scenarios of operation of an integrated system, including novel technologies, should be investigated to find the best path of operation. A process integration approach has been proposed to integrate steelmaking with a polygeneration system, which can produce electricity, district heat and methanol as by-products (Ghanbari et al., 2013). The motivation for this integration was that polygeneration systems are more energy efficient than standalone processes, as they have higher flexibility to switch between different forms of energy products depending on regional and seasonal demands (Liu, 2009). A benefit of liquid methanol is that it can be stored and used to substitute traditional energy carriers or as a feed material for chemical industries. Additionally, the steel plant could avoid paying for emission or sequestration of CO2 by converting (part of) the residual carbon to methanol, which extends the life cycle of carbon. This becomes a more interesting option when considering the effect of different auxiliary fuels in the steel plant, e.g., residual oil, natural gas, coke oven gas and pulverized coal or carbon neutral carriers such as biomass in the form of dried wood chips, torrefied biomass or charcoal (Ghanbari et al., 2012; Lundgren et al., 2013). In a recent work (Ghanbari et al., 2013), the present authors extended the analysis to include an optimal design of the gas treatment system, including CO2 separation, in a polygeneration plant integrated with a steel plant. However, the analysis undertaken was restricted to a few scenarios (conventional blast furnace), and a systematic analysis of different options to operate the future steel plant was not presented, and the flexibility offered by the system under (changing) external demand of energy was not studied.
2 Page 2 of 23
ip t
In order to address the limitations of the study presented in (Ghanbari et al., 2013), the mathematical model was modified to simulate novel operation concepts of the steel plant including different states for preheating and oxygen enrichment of the blast and residual top gases. The model was applied to solve an optimal design of the gas treatment part of the system, simultaneously optimizing the operation of the steel plant with the objective set for the entire system. In section 2 of the present paper, the system is described by introducing each unit process and the main assumptions made in the modeling. In the third section the optimization model is presented, treating the main unit operations. Section 4 presents some results regarding the optimal state of blast furnace operation and furthermore investigates the effect of seasonal variation in the external energy demand on the optimal state of the system. The fifth and final section presents some conclusions.
cr
2. Problem Statement and Processes Model Description
Ac ce p
te
d
M
an
us
The process industry has faced both economic crises and more stringent environmental restrictions specifically concerning carbon dioxide emissions. This has led to a focus of research in process engineering on sustainable development through better use of resources and increased energy and raw material efficiency by optimal design and operation. In this study, the operation of an integrated steel plant with facilities for CO2 capturing and utilization has been numerically investigated. Figure 1 illustrates the main processes in primary steelmaking, including units for agglomeration of the enriched ores (sinter plant) and the dry distillation of coal to coke (coke plant), which are the main burden materials for the ironmaking unit, the blast furnace. The liquid iron, called hot metal, from the blast furnace goes to the basic oxygen furnace, where carbon and impurities are blown off and the iron is refined to steel. The gases arising in the steel plant are usually partly recovered and used for preheating the combustion air (blast) for the blast furnaces in large regenerative heat exchangers (hot stoves), in the power plant for production of electricity and heat, in slab reheating furnaces and in the rolling mill, etc. In very few plants in the world the residual gases are used for other purposes, e.g., upgrading to liquid fuels. Because of severely limited storage capacity, the gases must be used almost immediately, which also means that several percent in practice must be flared in spite of the fact that auxiliary fuels are uses simultaneously.
Figure 1 Primary blast furnace steelmaking (Freitag and Richerson, 1998)
Different ways to utilize residual gases from the blast furnace, coke oven and basic oxygen furnace have been considered in a superstructure for process synthesis and analysis, which provides a framework for describing the distribution of the gases through the system under different constraints. Figure 2 shows the superstructure of a future steel plant (left part) integrated with a polygeneration system (right part). The steel plant includes a coke plant, sinter plant, hot stoves, blast furnace, basic oxygen furnace, combined heat and power plant, and air separation unit. The gas treatment units, methane gasification and methanol unit include optional technology for (pressure or temperature) swing adsorption, membrane adsorption, chemical absorption, steam methane
3 Page 3 of 23
M
an
us
cr
ip t
reforming, partial oxidation, carbon dioxide reforming, gas and liquid methanol reactors, etc., in addition to purification units, heat exchanges and compressors. Merged boxes denote alternative technologies (e.g., pressure swing adsorption, PSA, or membrane separation, MEM for coke oven gas treatment after initial compression), among which the optimizer has to make a decision.
Ac ce p
te
d
Figure 2 Superstructure for future Integrated Steel Plant (ISP). Full lines depict solid and liquid phase material flows and dashed lines illustrate the residual gas network. The conventional steel plant (left part) includes a Coke Plant (CP), Sinter Plant (SP), Hot Stoves (ST), Blast Furnace (BF), Basic Oxygen Furnace (BOF), Combined Heat and Power Plant (CHP) and Air Separation Unit (ASU). Potential technologies for carbon capturing and sequestration and methanol production are Pressure Swing Adsorption (PSA), Membrane adsorption (MEM), Steam Methane Reforming (SMR), Partial Oxidation Reactor (POR), Carbon Dioxide Reforming (CDR), Water Separation (WSP), Liquid phase Methanol reactor (LPMEOH), Gas phase Methanol rector (GPMEOH), Gas Separation unit (GSP), Dimethyl Ether purification (DME), Methanol purification (MEOH), Temperature Swing Adsorption (TSA), Chemical absorption unit (COPURE), Water Gas Shift reactor (WGS), CO2 Chemical Absorption (CCA), and Vacuum Pressure Swing Adsorption (VPSA).
2.1. Steelmaking Technology
The suggested system compromises primary steel making from sintermaking and cokemaking through the blast furnace up to liquid steel (denoted by subscript ls). Hence the process units for casting and rolling are not included in this study. The blast furnace, which is the heart of the steel plant, acts as large shaft-like counter current heat exchanger and chemical reactor, where the agglomerated iron-bearing burden is charged with coke in alternate layers. The combustion of coke is maintained by the supply of preheated air (blast), providing carbon monoxide to reduce iron oxides to iron and delivering energy to heat and melt the iron and gangue. The blast furnace top gas has a low calorific value, but it is traditionally recovered and used in the hot stoves (for blast preheating) and in the power plant for electricity (and possibly heat) production. A first principles model (Helle et al., 2011) has been developed for simulation of the blast furnace and has been verified on data from a Finnish blast furnace. The model has two main control volumes, the elaboration and preparation zones, which describe the state of the two parts of the furnace separated by a thermal and chemical reserve zone, where equilibrium is approached. In the elaboration zone, the model estimates the material and energy flows and compositions into and out of the control volume on the basis of input blast parameters and specific injection rates of auxiliary reductants. The results are used to express the mass and energy balances in the upper part of blast furnace, the preparation zone. Table 1 reports the blast furnace input and some output variables and their practical constraints imposed in this work. The inputs were selected among a set of potential
4 Page 4 of 23
ones to yield a unique solution of the equations at hand. For a more detailed description of the model, the reader is referred to the appendix of (Helle et al., 2011).
cr
ip t
By running the BF model for a large set of possible input variable values, a set of feasible data with more than half a million observations was generated, considering states ranging from conventional blast furnace operation to operation with blast oxygen enrichment and top gas recycling. Analysis of the results showed a very nonlinear behavior of the process, as was also observed by Helle and coworkers (Helle, 2014), who linearized the model only in the high-oxygen blast range. Therefore, a piecewise linear regression approach was made to build a surrogate model of the blast furnace. The three different states of input gas preheating considered in this work are depicted in Figure 3. In state 1 the blast and recycled top gas are pressurized and preheated in two sets of compressors and hot stoves. Thus, this state requires an additional set of compressors and hot stoves (650 m3) compared to traditional ironmaking, leading to higher investment costs. In the other two states only the available compressors and hot stoves are used for preheating, of either blast (state 2) or recycled top gases (state 3), maximally to 1200 C.
us
Table 1 Upper and lower bounds of blast furnace operation in this work
Variable BF bosh gas volume BF solid residence time BF slag basicity BF sinter feed flow Coke production flow
Range
an
Rang
Ac ce p
te
d
M
Variable BF blast oxygen content BF specific oil rate BF blast temperature BF top gas recycling BF top gas recycling temperature BF top gas temperature
Figure 3 Different states of blast preheating and oxygen enrichment. State 1: preheating recycled top gas (TGR) and blast (BL) in two sets of hot stoves, State 2: preheating BL in one set of hot stoves, and State 3: preheating TGR in one set of hot stoves.
Table 2 Regions of operation used to generate surrogate model. The states refer to those in outlined in Figure 3.
Oxygen enrichment of the blast (%) 21-32 55-65 84-99
Top gas recycling rate (km3n/h) 0 80-100 180-200 States 1, 2 and 3 States 1, 2 States 1, 3 States 1, 3 States 1, 2, and 3 States 1, 3
5 Page 5 of 23
Table 2 shows the three steps of oxygen enrichment from conventional blast furnace operation
,
ip t
and full oxygen blast operation considered, and the feasibility of strong enrichment these concepts according to practical constraints. The motivation for selecting the ranges was that the first one corresponds to conventional blast furnace operation, while the third one to (practically) full oxygen blast also studied earlier (Helle, 2014). The second range was selected as an intermediate state between the other two. Hence, a stepwise surrogate model based on oxygen enrichment and top gas recycling degree was developed. It should be noted that in the cases with no top gas recycling but state 3 (cf. Figure. 3), the compressors are used to pressurize the conventional blast and oxygen injection is realized through lances. The regression model is expressed as
cr
(1)
in the mixed-integer model using a convex hull reformulation
M
associated with a binary variables
an
us
and are regression coefficients, denotes the specific oil rate, pellet rate, blast oxygen content, where blast temperature and top gas recycling rate, which are the five inputs of the blast furnace model. The sixteen , of the surrogate model are the specific coke rate, volume flow rate of top gases , outputs, top gas temperature, sinter flow rate, blast volume flow rate, (raceway) flame temperature, burden residence time, bosh gas volume, lime stone flow rate, quartzite flow rate, slag flow rate and oxygen enrichment in hot represents the fourteen different segments applied for regression, which are stoves and lances. Superscript
(2)
(3)
Ac ce p
te
d
where LB and UB are the lower and upper bounds for oxygen enrichment and recycling rate in each time period . In addition to the blast furnace model, empirical equations with physical restrictions based on a data from a reference steel plant were used for the steelmaking unit processes (Ghanbari et al., 2012; Ghanbari et al., 2015; Helle et al., 2011). The cokemaking unit is considered to produce coke, coke oven gas, tar and residual fuel oil. The main part of the coke goes to the blast furnace, while a smaller amount of coke breeze goes to the sinter plant. For practical reasons the capacity of the coke production was considered to be fixed at an upper limit (55 t/h) Therefore, any deficit/excess of coke (after the requirement of the sinter plant and blast furnace) was taken to be bought/sold. The conversion of coal to metallurgical coke in the coke plant is simply modeled by linear relations between the feed rate of coal and the flow rates of produced coke and coke oven gas. The sinter plant is considered to have an upper limit of the production rate (here 200 t/h). Coke flow rate, iron ore flow rate and limestone flow rate are the inputs and sinter and recovered heat the outputs of the model of this unit. Pellets from kg pellet per ton of hot an external producer are available, if required, in the system in the range of metal (thm). The BOF converter, where scrap is added (here taken to be 25 % of the hot metal flow) is assumed to produce liquid crude steel with 0.1 w % of carbon and 50 % of the converter gases (with a carbon monoxide to carbon dioxide composition ratio of 9:1) is assume to be recovered. Oxygen is taken to be produced by cryogenic air separation, as this is the most robust way to produce the large volumes of oxygen required in a steel plant, but it is an energy intensive process: The specific power requirement for low pressure gaseous oxygen production was assumed to be 0.3 kWh/m3n (Castle, 2002).
2.2. Carbon Capturing and Utilization Technology Short-cut models and empirical equations are used to estimate material and energy flows for each operation unit (Ghanbari et al., 2013). The carbon capturing and utilization plant includes three main units: gasification, gas separation and polygeneration system. To implement top gas recycling in the blast furnace, off-gas treatment is necessary. Off-gases from the blast furnace and converter are sent to the carbon capturing unit. Amin-based
6 Page 6 of 23
cr
ip t
chemical absorption (CCA) (Hooey et al., 2013) and vacuum pressure swing adsorption (VPSA) (Kuramochi et al., 2012) are considered for carbon dioxide removal. The removed carbon dioxide is taken to be pressurized to 110 atm (which was set as the required pressure for transport CO2) and sent via a pipeline for sequestration. The residual gases are divided between the blast furnace and the polygeneration system. Further treatment such as carbon monoxide separation by temperature swing adsorption (TSA) (Ruthven et al., 1994) and chemical absorption (COPure) (Costello, 2011) combined with a water gas shift (WGS) reactor is considered to provide a carbon monoxide to hydrogen ratio required for methanol production. In case of conventional blast furnace operation a nitrogen separation unit is needed, where membrane separation (MEM) (Henis J.M.S. and Tripodi, 1980) and pressure swing adsorption (PSA) are considered as alternatives. Thus, some units such as the nitrogen separation unit can be omitted in case of oxygen (“nitrogen-free”) operation of the blast furnace. In the optimization problem formulation, a set of binary variables are introduced to select unit operation in the superstructure expressed as
us
(4)
Ac ce p
te
d
M
an
is the separation unit and are binary variables indicating selection of the technology. Another gas where treatment is that of coke oven gas (COG), which consists of mainly methane and hydrogen. COG can be used as a fuel or treated to produce chemicals such as methanol. In this study the possibility of both usages was implemented by introducing a gasification plant. After primarily cleaning, coke oven gas can be sent to the power plant or to the gasification unit. In the gasification unit, methane is removed from the gas by Pressure Swing Adsorption (PSA) and Membrane (MEM) technologies and sent to the methane gasification reactor to produce more syngas. Partial oxidation (POR), steam reforming (SMR) and carbon reforming (CDR) reactors were taken to be available technologies (Van Dijk et al., 1983; Wang et al., 1996; Zhu et al., 2001). Among the technologies steam reforming is pressurized, which implies a need of compressor(s) and considering gas liquefaction, water must be removed in a water separation unit before sending the gases to the polygeneration system. Therefore, water is removed by depressurizing the syngas products. In the polygeneration system, the available syngas is distributed between the power plant and the methanol production according to the demand for electricity and steam within and off the site, including electricity for the compressors and heat for unit processes transferred by a heat exchanger network in the system. Liquid-phase and gas-phase methanol reactor technologies with purification units are considered for methanol production (Van Dijk et al., 1983; Vaswani, 2000). In the gas-phase methanol reactor, 2 % percent of dimethyl ether is produced in the stoichiometric process. Equations for units with reactors can be expressed by
(5)
where are the available technologies in the gasification and methanol production units. The main material and energy balances considered for each unit process are (6) (7) (8) (9) where
and T are mass flow rates and temperature of the material (
for each unit (
in period , while
and
c
express the mass and energy balance terms and h is the cost function. The energy used in the compressor can
7 Page 7 of 23
be calculated by determining the work required for compressing from inlet to outlet pressure. The reference compressor is assumed to operate isentropically, and the true operation is estimated with adiabatic, motor drive and mechanical efficiencies of , and , respectively. The outlet temperature, specific enthalpy and compressor work are calculated by
(10)
ip t
(11)
and
are the outlet and inlet absolute temperatures,
is the average of the specific heat ratio of the components,
and
are the outlet and inlet pressures,
is the absolute temperature,
us
where
cr
(12)
is the standard
an
enthalpy in of component k and are parameters obtained from NIST Chemistry Web book (The National Institute of Standards and Technology, 2013). A network of compressors and heat exchangers provide the pressure and temperature level for each unit operation. An efficiency coefficient of 0.7 was applied to the heat exchanger network. Table 3 shows some operational conditions for the unit processes in the superstructure. More detail about the mass and energy balance equations for each unit can be found in (Ghanbari et al., 2015). Table 3 Operational condition for reactors and separation technology
,
te
d
M
Operational condition CH4/H2O = 3.681, P = 20 bar, T = 1153-1300 K, CH4/CO2 = 1, P = 1 bar, T = 1143-1313 K, CH4/O2 = 2, P = 1 bar, T = 1073-1473 K, H2/CO ≥ 2, P = 50 bar, T = 523 K, H2/CO ≥ 2, P = 50 bar, T = 533 K, Operational condition P = 3.4 bar, T = 318 K, , R=1.5 P = 11.2 bar, T = 318 K, , R=20 P = 11.2 bar, T = 318 K, P = 20 bar, T = 383 K, P = 1 bar, T = 273-573 K, Low P, T,
Ac ce p
Unit SMR CDR POR LPMEOH GPMEOH Separation Column MEOH DME GSP WSP TSA COPure VPSA CCA
Low P, T,
2.3 Carbon Dioxide Emissions
The CO2 emissions from the system are calculated on the basis of a carbon balance equation, including all fossil carbon-bearing inputs (coal, oil, external coke, limestone) and excluding the outflows of carbon with liquid steel, external coke, methanol, and stripped CO2. The emissions associated with the production of external raw materials (e.g., pellets) were not considered, as the units were outside the balance boundaries of the system.
3. Superstructure Optimization The proposed superstructure was applied to optimize the integrated plant using mixed integer nonlinear programming (Ghanbari et al., 2013). In the superstructure optimization, if a certain technology is chosen for a step and the operational conditions are not satisfied, the necessity of using, e.g., compressors and heat exchangers has to be considered. This has been formulated with Boolean variables (B) and as logical
8 Page 8 of 23
propositions. Equations (13) - (19) show how the different technologies were implemented, with upper and lower bounds and non-negativity constraints for continuous variables. The equations are implemented by applying binary variables and big-M formulation.
ip t
(13) (14) (15) (16) (17) (18) (19)
an
us
cr
If a specific unit process is chosen, then mass and energy balances are enforced and the corresponding cost terms are imposed in the objective function. In the case of bilinear terms a reformulation strategy was considered by convex and concave envelopes of the bilinear terms over the given bounds. The optimal design and operation is estimated by maximizing net present value (NPV), which determines equipment sizing and technology selection in the superstructure, and material and energy flow rates and distribution in the system. The problem is formulated as (20)
M
(21) (22)
d
where is the total capital investment given by fixed and working capital, is the bare module cost of equipment for development of conventional steelmaking, which is expressed by a linear approximation with fixed cost charge of the power law modular method (Biegler et al., 1997) with cost update factor for 2010, base capacity and cost, and
in each period ,
and
are the
is an exponential factor, NP is the net profit of the integrated system in period , are the life and depreciation time of the project and
Ac ce p
the tax rate,
te
is the maximum of the related parameter to sizing of an equipment (capacity)
is
is the annual discount rate. Fixed
capital investment was assumed to be the sum of manufacturing and non-manufacturing cost and is estimated as 1.4 times the calculated bare module cost with 25 % contingency. Working capital cost and direct expenses were considered to be 19.4 % and 4.0 % of the fixed capital investments, respectively (Biegler et al., 1997). Table 4 presents the economic terms and parameters used to estimate the net present value of the system. Note that the emission allowance for 1 ton of CO2 was assumed to cost 56 $.
Table 4 Key economic parameters in the objective function
Parameter
Value
Parameter
Value
Parameter
Value
4. Result and Discussion 9 Page 9 of 23
us
cr
ip t
The integrated system was modeled as a mixed integer nonlinear problem in the General Algebraic Modeling System (GAMS) using DICOPT as solver (GAMS Development Corporation, 2014). Due to bilinear terms and polynomial terms in the energy balances, the model is nonlinear and non-convex; therefore, global optimum is not guaranteed. Some of the bilinear terms were reformulated using convex hull to related linear forms. Linear under-estimators of the cost function (Biegler et al., 1997) are generated with an existence and sizing value which is represented by binary and continuous variables, respectively. The model of Section 4.1 has 60 binary variables, 4,596 continuous variables and 10,453 constraints, while for the model of Section 4.2, 171 binary and 18,141 continuous variables and 41,723 constraints are needed to represent different seasonal demands. Solution times for different scenarios are in order of minutes (< 15 min), so the use of a global solver with the same gap ratio was not considered due to its much longer computational time in comparison with that of DICOPT. The models were applied to investigate a steelmaking facility with a production rate of 170 tls/h, considering price for feed materials and by-products as reported in Table 4, and some physical/operational constrains (Table 1). It should be noticed that all feasible scenarios have been reported in the results for the sake of comparison and for investigation of the effect of the preheating states and different scenarios on the operation of the system.
4.1 Optimal Blast Furnace Operation Scenarios
Ac ce p
te
d
M
an
The base system was investigated for two different cases: Case 1 refers to operation without any external obligation: The integrated plant can completely flexibly distribute by-products according to the price to reach the was imposed on the system. maximum net present value. In Case 2 an external demand for electricity Figure 4 show the net present value and estimated steelmaking costs for both cases applying different technologies and states of operation, here called scenarios, and referred to as ISP (integrated steel plant). The numbers in the scenarios on the horizontal axis refers to the blast furnace state, top gas recycling rate and blast oxygen enrichment, as defined in the caption of the figure. For instance, scenario ISP-213 means BF state 2 with intermediate top gas recycling rate and cold oxygen injection.
Figure 4 NPV and steel production cost of the integrated plant without (Case 1, left blue bars) and with (Case 2, red right bars) external energy demand considering different scenarios of blast furnace operation. The number codes on the abscissa express the BF operation scenario: The first number expresses the BF states (1-3, cf. Figure 3), the second the top gas recycling rate (0 = no, 1 = intermediate and 2 = high recycling) and the third the oxygen enrichment (1 = normal, 2 = high enrichment and 3 = cold oxygen injection). Conv. represents conventional steelmaking without integration with CCU and MEOH plants.
From the figure it can be seen that scenario ISP-323 (i.e., preheating only recycled top gas, high recycling rate, full oxygen blast) has the highest NPV for Case 1, while imposing the external electricity demand has a strongly reducing effect on the NPV value, also in comparison with the value for optimized conventional BF operation. Also ISP-313 (with intermediate top gas recycling rate) turns out to be promising in Case 1. It is also seen how
10 Page 10 of 23
cr
ip t
the combination of recycling and oxygen enrichment affects the final cost of steel production. Imposing the external power demand increases the steelmaking costs by about 20 $/tls. The carbon dioxide flow from each optimized scenario has been presented in Figure 5. This figure shows the sum of the specific emission and sequestration from the system in the two cases. It is seen that in all integrated scenarios the specific emissions from the system are reduced compared to conventional blast furnace operation, but the specific carbon dioxide outflow may increase, particularly in Case 2. In Case 1 a stronger reduction is achieved due to methanol production. For ISP-123 the emissions are comparable with ISP-323, but it has a clearly lower NPV (cf. Figure 4). This difference is mainly due to the required investment in a new set of hot stoves and compressors in the former system. For the system without top gas recycling and low oxygen enrichment, where the single hot stove set is used to preheat the blast (ISP-101 and ISP-201), the effect of preheating states can be studied. For Case 1, scenarios ISP-101 and ISP-201 yield the same optimal structure and show similar behavior in terms of economics and environmental impact. External power demand (Case 2) makes the optimal operation different with and without sequestration. Figure 4 and Figure 5 show that the maximum profit is reached without any CO 2 sequestration (specific emission ), while by sequestration of about in ISP-201 the . It is interesting to note that adding a carbon capturing unit increases
specific emission decreases to
Ac ce p
te
d
M
an
us
the specific carbon dioxide generation, which is caused by an additional oil use in the CHP (cf. Figure 7). ISP301, in turn, shows the lowest NPV and highest steelmaking cost mainly due to the state of the gas preheating, which is also reflected in the fuels needed in the system.
Figure 5 Specific carbon dioxide flow from the integrated plant without (Case 1, left bars) and with (Case 2, right bars) fixed external power demand, considering different scenarios of blast furnace operation. The scenario number code on the abscissa is defined in the caption of Figure 4.
11 Page 11 of 23
ip t cr us
an
Figure 6 Electricity, district heat and methanol production of an integrated plant without fixed external demand of electricity (Case 1) and methanol production for the case with fixed external energy demand (Case 2), considering different scenarios of blast furnace operation. The scenario number codes are defined in the caption of Figure 4. The constraint considered as lower bound (external energy demand) is active for all scenarios in Case 2.
Ac ce p
te
d
M
Figure 6 shows the production of electricity, district heat and methanol for Cases 1 and 2. Electricity and district heat production are on their lower bounds in all scenarios in Case 2 (and therefore not shown), but the production of methanol may vary by up to 35 t/h from case to case (e.g., in ISP-101). The conventional scenario is seen to have a high production of electricity and heat in Case 1, while the values are clearly lower for the other scenarios, except ISP-211 and ISP-111 (where oxygen enrichment is low). Overall, the production of methanol is lower in Case 2 compared to Case 1. Thus, imposing an external electricity demand gives the system less (economic) potential to produce methanol because the need of (more expensive) auxiliary fuels increases. The promising scenario ISP-323 has a moderate methanol production of about 10 t/h. Figure 7 shows the flow rate of oil injected into the blast furnace to partially replace coke (upper parts of bars), and extra oil needed in the polygeneration system to supply energy needed in the system (lower parts of bars) for the optimized solutions of different scenarios and cases. An upper bound of 50 t/h for the oil supply to the polygeneration system was imposed, but this constraint becomes active only in ISP-122, Case 2.
12 Page 12 of 23
ip t cr us
an
Figure 7 Oil flow rate of an integrated plant without (Case 1, left bars) and with (Case 2, right bars) fixed external power demand, considering different scenarios of blast furnace operation. The scenario number codes are defined in the caption of Figure 4. The constraint considered as lower bound (external power demand) applies to all scenarios in Case 2.
M
The results show that on average the need for oil would be about twice that used in conventional steelmaking at the reference plant, although there are exceptions, such as ISP-123, Case 1. In this scenario there is no byproduct production, which leads to higher net present value and lower specific emission than in conventional steelmaking.
Ac ce p
te
d
From the figures, we can compare different operational scenarios to get an overall view of how a steel plant adopting blast furnace top gas recycling stepwise could be integrated with a CO2 capturing and sequestration unit and a methanol plant, and how this novel site would perform compared to conventional steelmaking. It is interesting to analyze the flow of carbon in the different optimized systems. Figure 8 shows a comparison between the oxygen blast furnace system (ISP-323, inner ring) and conventional steelmaking (outer ring). In conventional steelmaking about 50 % of the input carbon leaves the system as emissions. This carbon is mainly supplied in two feed material flows, i.e., coal to the coke plant and oil as auxiliary reductant in the BF. The CO2 emissions from the system decreases to half for the ISP-323 plant: the difference leaves the system as sequestrated CO2 and as methanol. 4.2 Seasonal Operation of Integrated System To further investigate the effect of an external energy demand on the optimal operation of the integrated system, a time period of four seasons with 0, 15, 30 and 40 MW external electricity demand was assumed. Thus, these values were activated as lower bounds during the four periods and a multi-period task was optimized.
Figure 9 shows the optimized integrated steel plant in ISP-323, i.e., high top gas recycling rate with preheating and cold oxygen supply, which is the oxygen blast furnace concept, indicating the main unit processes and the connection of the streams.
13 Page 13 of 23
ip t cr us
Ac ce p
te
d
M
an
Figure 8 Comparison of carbon flows in conventional steel plant (Conv, outer ring) and novel integrated plant (ISP-323, inner ring).
Figure 9 Suggested integrated plant (ISP-323) and the main streams.
14 Page 14 of 23
Table 5 Main streams (in kmol/h) and BF variables in ISP-323. For stream numbers, see Figure 9.
CH4 222 222
50.7 997
11.1
11.1 13.2
438
an
Blast volume [km3n/h] Flame temperature [ºC] Recycled top gas temp. [ºC] Recycled top gas volume [km3n/h] Bosh gas volume [km3n/h] Top gas temperature [ºC] Burden residence time [h] Slag rate [kg/thm] COG volume [km3n/h] BOFG volume [km3n/h] Carbon dioxide send out [t/h]
M
1484 105 1076 302 33.2 261 42.1 458 81.6 154 21.3 37.1 160 9 30
MEOH
us
997 989 7.3 50.7 7.3 50.7
H2 O
DME
ip t
N2 50.7 50.7
cr
O2 1.1 1.1
8.7
23.9 1800 1200 180 183 193 8.8 260 17.6 6.2 116
te
d
CO CO2 H2 41.4 13.9 456 1 41.4 13.9 89.5 2 366 3 252 13.9 511 4 3910 2787 731 11 2647 12 3910 139 731 13 3618 135 703 14 292 3.4 27.4 15 439 13.9 877 21 105 3.4 27.4 22 21.9 0.7 4.3 23 31 32 33 34 Oxygen volume [km3n/h] Specific coke rate [kg/thm] Specific oil rate [kg/thm] Specific pellet rate [kg/thm] Coal flow rate [t/h] Ore flow rate [t/h] Limestone rate [t/h] Scrap rate [t/h] Sinter flow rate [t/h] External (sold) Coke flow rate [t/h] Average extra oil needed [t/h]
Ac ce p
Coke oven gas (stream-1) goes to the gasification plant. First, mainly hydrogen is separated from it in a PSA unit which operates at 10 atm, and is sent to the polygeneration system (stream-3). The rest of the gas, which has high methane content, is sent to the gasification reactor. The partial oxidation process was selected here, which operates at low pressure and at its optimal temperature (1143 K). A mixture of blast furnace and basic oxygen furnace gas (stream-11), taken to be mostly CO (46%), CO2 (34%), H2 (9%) and N2 (11%) after scrubbing, goes to a VPSA unit. Pressurized CO2 at 110 atm sends out of the system for sequestration. The residual gases rich in carbon monoxide are assumed to be divided between the polygeneration system and the blast furnace, before which they are preheated (stream-14). The solution has omitted the TSA unit based on the composition of the gases in this state (~88% CO). The residual gases from capturing and gasification are pressurized and sent to the polygeneration system to produce methanol, electricity and district heat according to the local demand, considering the economics. Table 5 shows the optimal values of the key variables in this specific system. For the sake of comparison, the optimization results of conventional steelmaking (Conv.), integrated conventional steelmaking (ISP-201) and integrated steel plant with novel blast furnace operation (ISP-323) will be presented. In these results, the design of the gas treatment units in each solution naturally stays the same for all periods, but the flows in the system may vary between the periods. In some cases, the results of the optimization show similar states of operation of the blast furnace for all periods. Figure 10 shows net present value, specific emissions, methanol production, fuel required in the polygeneration system, amount of CO2 sent out of the system, BF coke rate and cost of steel production. The net present value decreases by the first step of integration (from conventional operation to ISP-201), but on applying blast furnace top gas recycling and oxygen blast the net present value of the system increases. Considering the result of the
15 Page 15 of 23
case studies in the previous subsection, it is seen that seasonal operation may increase the total net present value of the system. The specific emissions decrease at each step of integration, from 1.7 to about , which is lower than for the single-period operation. There is a trade-off between fuel consumption and integration which indicates the importance of the types of fuel and fuel costs. The specific coke rate decreases by about 60 kg/thm by injecting top gases as reducing agents. From net profit analysis the cost of liquid steel production was estimated for each case, showing a decrease of more than 25 $/tls for the integrated system.
te
d
M
an
us
cr
ip t
Figure 11 shows the specific emissions for three cases in each of the periods. The specific emissions from the system are seen to increase along with the external electricity demand. Conventional steelmaking operates with up to Period 3. specific emissions close to
Ac ce p
Figure 10 Comparison of net present value (NPV), specific CO2 emissions, methanol production, external fuel (oil), carbon dioxide flow for sequestration, coke rate in the blast furnace and estimated cost of steel production for conventional steelmaking, ISP-201 and ISP-323. The numbers on the bars show the value of each term for the system under periodic external energy demand.
16 Page 16 of 23
ip t cr us
Figure 11 Specific emissions in each period for conventional steelmaking, ISP-201 and ISP-323.
an
The cumulative emissions (Figure 12) illustrate the lower emissions of the top gas recycling concepts, despite the increase experienced along with the growing external demand of energy.
M
Figure 13 shows the estimated steelmaking cost for the optimized cases in each period. The maximum difference in the steel cost occurs in Period 1 (without given external demand) which indicates the flexibility of the integrated system to reach lower costs by flexible production.
Ac ce p
te
d
The largest difference in the cost of steel production is between the conventional system and the system with top gas recycling and cold oxygen injection in the BF (ISP-323). The flexibility of the integration is also seen for the period with maximum external power demand (Period 4), where the steelmaking cost of ISP-323 are still lower than in conventional steelmaking without external constraints on the power production.
Figure 12 Cumulative specific emissions in all periods for conventional steelmaking, ISP-201 and ISP-323.
17 Page 17 of 23
ip t cr us
an
Figure 13 Comparison between steel costs for conventional steelmaking, ISP-201 with sequestration and ISP-323 for different periods.
M
Finally, in order to illustrate the sensitivity of the optimized results of a selected system (ISP-323), Figure 14 shows the changes in the net present value when perturbations of ±50 % in the price of the feed material and byproducts were introduced. Quite naturally, the price of ore has the highest influence (up to 25 %) on the NPV, followed by emissions, pellets and coal.
Ac ce p
te
d
Electricity is seen to have the lowest effect on the net present value due to the small power production in the initial state. As environmental restriction increases it gets more challenging to deliver low-price electricity from the steelmaking sector, but this could still be seen as a feasible future initiative to suppressing emissions from the steelmaking system, simultaneously increasing the life cycle of carbon.
Figure 14 Sensitivity analysis of ± 50 % change in price of key feed materials and byproducts on net present value for integrated steelmaking (ISP-323)
18 Page 18 of 23
5. Conclusions
ip t
This paper has presented a numerical study of the potential to reduce emissions by an integration of steelmaking with a polygeneration plant, producing methanol as a byproduct, and facilities for CO2 capture by an overall model. The model developed considers a steel plant that can be run in a conventional way or under novel blast furnace operation modes, with cold oxygen injection and top gas recycling, in order to throw light on the effect of steel plant operation on the optimal design of the polygeneration plant. The results of the study provide information about how a future steelmaking plant could be sustainably operated. Such operation and design optimization for sustainable steelmaking has not been presented in the literature before.
an
us
cr
A large number of scenarios of design and operation have been investigated in the analysis, and the complex entity has been optimized with respect to economic objectives. For the system without other external demands than on the steel produced, the most economical state (in terms of the net present value used as the objective) is the one with novel operation of the blast furnace (applying high recycling of top gas after preheating) and this also yielded low CO2 emissions from the system and a quite high production rate of methanol. Imposing an external electricity demand considerably reduced the NPV and increased the cost of steelmaking to become close to those of conventional (but optimized) operation. In general, it was found that applying top gas recycling and cold oxygen injection in the blast furnace could decrease the emissions from the system and the integration of the steel plant with a polygeneration system increased the economic profitability. The results also demonstrated that there is a trade-off between economic profit and emissions from the system. Another general observation that was made was that the flexibility of the integrated system was strongest in cases where no or little external constraints were imposed on the operation: Here, the system had the freedom to use the full innovation potential of the optimizer in the design to operate the integrated plant in the most economical way.
Ac ce p
te
d
M
Naturally, the specific results obtained depend on the assumed cost structure, but these parameters can be easily updated in future studies. In the future work on investigating different steel production concepts, a more general surrogate model of the blast furnace operation that could capture all operational states would be useful, as it would decrease number of binary variables in the optimization problem. Such a model is also estimated to better reflect the characteristics of the blast furnace operation. Further development of the model could also include an analysis of other auxiliary reducing agents with lower environmental impact, such as natural gas and biomass.
19 Page 19 of 23
References Biegler, L. T., Grossmann, I. E., Westerberg, A. W. (1st Eds), , 1997. Systematic Methods of Chemical Process Design. NJ, USA: Prentice Hall PTR.
ip t
Birat, J. P., 2009. Steel and CO2 -the ULCOS program, CCS and Mineral carbonation using Steelmaking Slag, 1st International Slag Valorisation Symposium , Leuven, Belgium. 6-7 Apr. Castle, W., 2002. Air separation and liquefaction: recent developments and prospects for the beginning of the new millennium, International Journal of Refrigeration-Revue Internationale Du Froid 25(1), 158-172.
cr
Costello, R., 2011. COPureSM Process,R.C. Costello & Assoc., Inc.
us
Danloy, G., Berthelemot, A., Grant, M., Borlee, J., Sert, D., van der Stel, J., Jak, H., Dimastromatteo, V., Hallin, M., Eklund, N., Edberg, N., Sundqvist, L., Skoeld, B. -., Lin, R., Feiterna, A., Korthas, B., Mueller, F., Feilmayr, C., Habermann, A., 2009. ULCOS - Pilot testing of the Low-CO2 Blast Furnace process at the experimental BF in Lulea, Revue De Metallurgie-Cahiers D Informations Techniques 106(1), 1.
an
European Climate Foundation ECF Roadmap 2050, 2010. A practical guide to a prosperous, low carbon Europe. Technical analysis. Initiative of the European Climate Foundation .
M
Fischedick, M., Marzinkowski, J., Winzer, P., Weigel, M., 2014. Techno-economic evaluation of innovative steel production technologies, Journal of Cleaner Production DOI: 10.1016/j.jclepro.2014.05.063.
d
Freitag, D. W., Richerson, D. W., 1998. Opportunities for Advanced Cermamics to Meet the Needs of the Industries of the Future, US Advanced Ceramics Association and Oak Ridge National Laboratory DOE/ORO 2076.
te
GAMS Development Corporation, 2014. GAMS (General Algebraic Modeling System) Software v23.7.3, http://www.gams.com .
Ac ce p
Ghanbari, H., Helle, M., Saxén, H., 2012. Process integration of steelmaking and methanol production for suppressing CO2 emissions—A study of different auxiliary fuels, Chemical Engineering and Processing: Process Intensification 61, 58-68. Ghanbari, H., Saxén, H., Grossmann, I. E., 2013. Optimal design and operation of a steel plant integrated with a polygeneration system, AIChE Journal 59(10), 3659-3670. Ghanbari, H., Pettersson, F., Saxén, H., 2015. Sustainable development of primary steelmaking under novel blast furnace operation and injection of different reducing agents, Chemical Engineering Science 129, 208-222. Grip, C., Larsson, M., Harvey, S., Nilsson, L., 2013. Process integration. Tests and application of different tools on an integrated steelmaking site, Applied Thermal Engineering 53(2), 366-372. Helle, H., 2014. Towards sustainable Iron-and steelmaking with economic optimization, Åbo Akademi University. Helle, H., Helle, M., Saxen, H., 2011. Nonlinear optimization of steel production using traditional and novel blast furnace operation strategies, Chemical Engineering Science 66(24), 6470-6481. Henis J.M.S., Tripodi, M. K., 1980. Multicomponent Membrane for Gas Separations, US Patent 4230463. Hooey, L., Tobiesen, A., Johns, J., Santos, S., 2013. Techno-economic Study of an Integrated Steelworks Equipped with Oxygen Blast Furnace and CO2 Capture, Energy Procedia 37, 7139-7151.
20 Page 20 of 23
Kuramochi, T., Ramírez, A., Turkenburg, W., Faaij, A., 2012. Comparative assessment of CO2 capture technologies for carbon-intensive industrial processes, Progress in Energy and Combustion Science 38(1), 87112. Li, Y., Zhu, L., 2014. Cost of energy saving and CO2 emissions reduction in China’s iron and steel sector, Applied Energy 130, 603-616.
ip t
Liu, P., 2009. Modelling and Optimisation of Polygeneration Energy Systems, Doctor of Philosophy, Imperial College London . Lundgren, J., Ekbom, T., Hulteberg, C., Larsson, M., Grip, C. -., Nilsson, L., Tunå, P., 2013. Methanol production from steel-work off-gases and biomass based synthesis gas, Applied Energy 112, 431-439.
cr
Moya, J. A., Pardo, N., 2013. The potential for improvements in energy efficiency and CO2 emissions in the EU27 iron and steel industry under different payback periods, Journal of Cleaner Production 52, 71-83.
us
Ruthven, D. M., Farooq, S., Knaebel, K. S., 1994. Pressure Swing Adsorption. USA: VCH wiley.
Suopajärvi, H., Pongrácz, E., Fabritius, T., 2014. Bioreducer use in Finnish blast furnace ironmaking–Analysis of CO2 emission reduction potential and mitigation cost, Applied Energy 124, 82-93.
an
The National Institute of Standards and Technology, NIST Chemistry Webbook, The National Institute of Standards and Technology 2013.
M
Van Dijk, C. P., Solbakken, A., Rovner, J. M., 1983. Methanol from coal and Natural Gas, US Patent 4407973. Vaswani, S., 2000. Development of Models for Calculating the Life Cycle Inventory of Methanol by Liquid Phase and conventional Production Process, Department of Civil Engineering, Raleigh, NC .
te
d
Wang, C., Larsson, M., Ryman, C., Grip, C. -., Wikström, J. -., Johnsson, A., Engdahl, J., 2008. A model on CO2 emission reduction in integrated steelmaking by optimization methods, International Journal of Energy Research 32(12), 1092-1106.
Ac ce p
Wang, S., Lu, G., Millar, G., 1996. Carbon dioxide reforming of methane to produce synthesis gas over metalsupported catalysts: State of the art RID C-5507-2008 RID A-2859-2008, Energy & Fuels 10(4), 896-904. World Steel Association, 2013. Sustainable Steel Policy and Indicators 2013, Worldsteel.org , Report NO. 9782-930069-75-3. Zhu, J., Zhang, D., King, K., 2001. Reforming of CH4 by partial oxidation: thermodynamic and kinetic analyses, Fuel 80(7), 899-905.
21 Page 21 of 23
Nomenclature
cr
te
d
M
an
us
coefficient for mass balance functions in process model Air separation unit Boolean variables Blast furnace Blast furnace gas Bare module cost Basic oxygen furnace Basic oxygen furnace gas cost factor ($) CO2 chemical absorption unit Carbon capturing and sequestration Carbon dioxide reforming unit Combined heat and power plant Coal Coke oven gas External coke bought or sold Chemical absorption unit Coke plant District Heat Dimethyl ether purification unit Dimethyl ether Depreciation time Depreciation time energy Electricity Emission CO2 Feedstocks mass flow rate (t/h) Fuel Gas phase methanol unit Gas separation unit function Hot stoves materials chemical process component lower bound Limestone Life time Life time Liquid phase methanol unit liquid steel mass Membrane adsorption unit
Ac ce p
A, b ASU B BF BFG BMC BOF BOFG c CCA CCS CDR CHP coal COG coke COPURE CP dh DME dme dp dp e el emis F f fuel GPMEOH GSP h HS i j k LB lime lp lp LPMEOH ls m MEM
ip t
Abbreviations:
22 Page 22 of 23
ip t cr us an
te
d
M
Methanol purification unit Methanol Methane gasification Methanol production net profit Net present value Iron ore products pressure (bar) Pellet Partial oxidation reforming unit Pressure swing adsorption unit Pyrolysis unit Quartz Reflux/recycle ratio reactor Recycled top gas Base capacity Scrap segment separator Sequestrated CO2 Steam methane reforming unit Sinter plant tone temperature (C) Capital investment Temperature swing adsorption upper bound volume flow rate (km3/h) CO2 capturing vpsa unit Water gas shift unit Water separation unit fraction converted per pass Inputs to BF model selection task Outputs from BF model Split ratio exponent in Guthrie’s formulation coefficient for fixed charge cost model annual discount rate time tax rate period
Ac ce p
MEOH meoh MG MP NP NPV ore p P pel POR PSA PYRU quartz R reac rtg S scrap seg sep seq. SMR SP t T TCI TSA UB V VPSA WGS WSP x X Y Z ζ
23 Page 23 of 23