Journal of Materials Research and Technology Journal of Materials Research and Technology
J Mater Res Technol 2017;6:258-70 DOI: 10.1016/j.jmrt.2017.05.007
Original Article
A theoretical study using the multiphase numerical simulation technique for effective use of H2 as blast furnaces fuel
Jose Adilson de Castroa,, , Cyro Takanob, Jun-ichiro Yagic
a Graduate Program on Metallurgical Engineering, Federal Fluminense University, Volta Redonda, RJ, Brazil
b Metallurgical and Materials Engineering Department, Polytechnic School, University of São Paulo, São Paulo, SP, Brazil
c Tohoku University, Sendai, Japan
Received 01 March 2017, Accepted 16 May 2017
Abstract

We present a numerical simulation procedure for analyzing hydrogen, oxygen and carbon dioxide gases injections mixed with pulverized coals within the tuyeres of blast furnaces. Effective use of H2 rich gas is highly attractive into the steelmaking blast furnace, considering the possibility of increasing the productivity and decreasing the specific emissions of carbon dioxide becoming the process less intensive in carbon utilization. However, the mixed gas and coal injection is a complex technology since significant changes on the inner temperature and gas flow patterns are expected, beyond to their effects on the chemical reactions and heat exchanges. Focusing on the evaluation of inner furnace status under such complex operation a comprehensive mathematical model has been developed using the multi interaction multiple phase theory. The BF, considered as a multiphase reactor, treats the lump solids (sinter, small coke, pellets, granular coke and iron ores), gas, liquids metal and slag and pulverized coal phases. The governing conservation equations are formulated for momentum, mass, chemical species and energy and simultaneously discretized using the numerical method of finite volumes. We verified the model with a reference operational condition using pulverized coal of 215kg per ton of hot metal (kgthm−1). Thus, combined injections of varying concentrations of gaseous fuels with H2, O2 and CO2 are simulated with 220kgthm−1 and 250kgthm−1 coals injection. Theoretical analysis showed that stable operations conditions could be achieved with productivity increase of 60%. Finally, we demonstrated that the net carbon utilization per ton of hot metal decreased 12%.

Keywords
Blast furnace, Mathematical model, Mix injection, CO2 emission
1Introduction

The hot metal elaboration unit operation during the primary metallurgy of the iron production is the most intensive in carbon, where the route blast furnace (BF) – basic oxygen furnace (BOF) is predominant due to economical and production scale reasons. Nevertheless, this production route is intensive in energy and demands specific and elaborated quality for the raw materials as coke and granular sinter, which allows higher efficiency of the reduction step with excellent gas utilization ratio. Conversely, the production of rich hydrogen gas from coal gasification process using fossil or renewable resources has become economically viable using newly developed technologies. Therefore, new efficient use of rich hydrogen gas in this industry is a promising issue to be addressed [1,2]. In the primary steel integrated production route, the BF operation unit is responsible for nearly 70% of the demanded energy. The fuel and reducing agent consumptions represent more than 60% of the total cost of pig iron production in this processing route. These materials are consumed in the blast furnace as granular coke or pulverized coal (PC) fed through the tuyeres of the reactor. Consequently, efforts have been addressed to decrease the reducing agent rate (RAR), or at least, to replace the coke consumption of the process by alternative materials injected through the blast furnace tuyere, which makes the process less carbon intensive and environment friendly [3–10]. The desired increase on the pulverized coal injection is a complex and challenging task for the blast furnace technology. The reasons are mainly due to the gas flow control and unburned coal particles accumulation in the lower part of the furnace, which can generate unstable solid and liquids descending motions [8–10]. To analyze the complexes phenomena that take place in the whole blast furnace comprehensive mathematical models based on multiphase multicomponent theory have been successfully applied with new features and continuous improvements [10–18]. Thus, we proposed to analyze, using a detailed mathematical model, a combined injection practice of H2, O2 and CO2 to take the advantages of the expected increasing of the raceway temperature by the H2 combustion. This excess energy is used to promote CO2 conversion through fast solution loss reaction with the pulverized coal in the raceway, avoiding unburned coal accumulation and allowing suitable slagging practice [19,20]. Accordingly, we propose the concept of simultaneous injection of H2, O2 and CO2 recycling into the tuyere of the furnace, as shown in Fig. 1. For this practice, external sources of hydrogen and oxygen rich gas are needed while CO2 is obtained by treating the blast furnace exit gas. With this practice, the BF off gas is enriched and its calorific value improved. Besides, it is expected that the kinetics of pulverized coal combustion in the raceway is enhanced and the temperature of the lower part of the furnace kept under control, since combustion of H2 will provide additional energy. In contrast, the enhanced water gas and solution loss reactions will consume this excess energy, depending on the amount of CO2 injected. The tiny balance of these reactions will keep suitable temperature distributions and materials flow in the blast furnace. Fig. 2 shows the expected mechanism and flow pattern, which is able to enhance the pulverized coal reactions within the raceway replacing coke reactions in this zone. Several blast furnaces around the world have been operated with injection rates higher than 180kgthm−1. Nevertheless, further increase on the injection rates are limited due to materials flows within the lower part of the furnace and cohesive zone. The unburned powder accumulation and high viscosity slag formed during ash melt deteriorate the permeability of the dropping zone leading to unstable conditions [13,20]. When injection of mix gases is carried out in the blast furnace, the temperature and gas flow inside the raceway are drastically changed. These effects on the raceway region define the whole furnace operation leading to mutual interactions of the phases and chemical species. Special concern is related with the cohesive zone shape and position, which play important role on the granular zone behavior regarding the fluid flow, heat transfer and reaction rates. The multiphase multicomponent model is able to take in to account these phenomena and quantify the yield of raw materials. In this paper we tested the model for an actual operation with high injection of PC (220kgthm−1) and simulated scenarios of higher injection (250kgthm−1) to demonstrate the new proposed operation taking the advantages of hydrogen injection in the raceway region. It is important to notice that, although hydrogen rich gas has been previously tested on the granular zone [7,14], we proposed a very new concept, which not only improve the reducing zone efficiency, but use the advantages of the thermal control of the lower part of the furnace with the addition of oxygen and hydrogen to enhance the exothermic reactions. This tiny control allows the smooth operation of the whole furnace.

Fig. 1.
(0.2MB).

Concept of mix injection of H2, O2 in raceway of the blast furnace and CO2 recycling technology.

Fig. 2.
(0.17MB).

Raceway reactions modifications for gaseous injection of H2, O2 and recycling CO2.

2Methods2.1Modeling principles

The mathematical model proposed in this study is assumed three-dimensional and analyses the in-furnace region of the packed bed. The domain is assumed from the slag surface in the hearth up to the burden surface in the throat. Five phases are modeled: gas, granular solids (coke, sinter, pellets, granular iron ore), hot metal, slag, and pulverized coal (PC). In the gas phase are included the newly considered chemical species and chemical reactions. Fig. 3 shows the continuous and dispersed phases and their mutual physical and chemical interactions assumed in this model. All phases are treated simultaneously due to mutual interactions and inter-phase mass, momentum and energy exchanges. Similar approach has been discussed in previous literature, which confirms the adequacy of this theory to be applied for simulating the BF process [3–5,13–22]. The mix injection of hydrogen, pulverized coal, oxygen and carbon dioxide recycling within the raceway, however, is proposed and demonstrated the technological feasibility for the first time in this study.

Fig. 3.
(0.18MB).

Phases and respective interactions assumed in the proposed model.

2.1.1General conservation equations treated in the model

A general formulation for the blast furnace process is proposed. Thus, the governing equations of all phases, that form a large set of strongly coupled non-linear differential equations, are solved simultaneously. In this model, conservation equations of motion, energy and chemical species are formulated and coupled with chemical reactions and physical properties dynamically calculated. For simplicity, all the conservation equations are represented in the compact form for a generalized coordinate system, as read in Eq. (1)[12–25].

In this equation, ϕ is the dependent variable, expressing the component velocities for the momentum equations of each phase, phase energy and the chemical species of the phase, i represents the phase being considered and k the index accounting for the variables used, which can assume components of the velocities for momentum equations, enthalpies for energy conservation and chemical species of each phase for compositions. α and ρ are phase volume fraction and density, respectively. U and t are phase velocity field and time, respectively. Γϕk is the effective transfer coefficient which represents effective dynamic viscosity in the momentum equations, effective thermal diffusivity for the energy equations and effective diffusion coefficient of the chemical species for the respective phase. The sources (Sϕk) are due to inter-phase interactions that can appear through chemical reactions, surface and bulk interactions and gravity [11–25]. The phases are composed of various chemical species and the general conservation equation is used to calculate the phase motion, the phase energy and the mass fraction of chemical species in each phase. The list of all five phases and their respective species treated in this model is presented in Table 1 together with their variables accounting for momentum and energy conservations.

Table 1.

Five phases and their chemical species considered in the model formulation.

Variables for continuous phases
GasMomentum  u1,g,u2,g,u3,g,Pg,εg
Energy  hg
Chemical species  N2, O2, CO, CO2, H2O, H2, SiO, SO, SO2, CH4, C2H6, C3H8, C4H10
SolidMomentum  u1,s,u2,s,u3,s,Ps,εs
Energy  hs
Chemical speciesLump coke  C, Volatiles, H2O, Al2O3, SiO2, MnO, MgO, CaO, FeS, S2, gangue 
Small coke  C, Volatiles, H2O, Al2O3, SiO2, MnO, MgO, CaO, FeS, S2, gangue 
Lump ore  Fe2O3, Fe3O4, FeO, Fe, H2O, Al2O3, SiO2, MnO, MgO, CaO, FeS, P2O5, gangue 
Sinter  Fe2O3, Fe3O4, FeO, Fe, H2O, Al2O3, SiO2, MnO, MgO, CaO, P2O5, K2O, Na2O, Ca2Fe3O5, Al2MgO4 
Fluxing agent  CaCO3, MgCO3, CaO, MgO, H2O, Al2O3, SiO2, MnO 
Pellet  Fe2O3, Fe3O4, FeO, Fe, H2O, Al2O3, SiO2, MnO, MgO, CaO, FeS, P2O5, K2O, Na2O, gangue 
Variables for discontinuous phases
Hot metalMomentum  u1,hm,u2,hm,u3,hm,εhm
Energy  hhm
Chemical species  Fe, C, Si, Mn, P, S
SlagMomentum  u1,slg,u2,slg,u3,slg,εslg
Energy  hslg
Chemical species  Fe3O4, FeO, Al2O3, SiO2, MnO, MgO, CaO, P2O5, CaS
PCMomentum  u1,pc,u2,pc,u3,pc,εpc
Energy  hpc
Chemical species  C, Volatiles, H2O, Al2O3, SiO2, MnO, MgO, CaO, FeS, S2

PC, pulverized coal.

Total of 140 partial differential equations numerically solved using the finite volume technique with multiphase approach.

2.1.2Boundary conditions and numerical approach used in this study

The boundary conditions assumed in this study use the observed information of the dynamics of the reactor and measured parameters on the actual operation practice. Therefore, the boundary conditions are applied on the boundary of the computational domain delimitated at the bottom by the slag surface, at the top by the burden surface profile and by lateral walls. At the top, the gas phase is assumed as fully developed flow and solid inflow is modeled based on the inflow rate given by local solid mass consumption due to chemical reactions and melting or gravity driven flows. At the tuyere injection, inlet of blast, additional oxygen, hydrogen or fuel gas and pulverized coal are given by their inflow rates. The blast flow rate and gaseous fuel injection are fixed while pulverized coal injection is iteratively calculated to reach the aimed pulverized coal injection rate, which are specified at the beginning of the iterative calculation. The blast temperature and pulverized coal temperature inlet are specified as fixed values throughout the calculation. At the side wall, mass fluxes across the wall are assumed null while heat loss is allowed by setting an overall cooling heat transfer coefficient calculated based on the cooling system data. For the gas velocity it is assumed null values perpendicular and tangential to the furnace wall. The solid tangential velocity on the wall surface assumes coulomb attrition law with a specified coefficient of 0.3 and the normal force is calculated using the local solid pressure, which is lumped into the source terms of the momentum equations of solid phase in the nearest volume of the wall surface and acts always as resistance term to the solid motion near the wall surface allowing sleep granular motion. The burden distribution is determined by the relative volume fractions of the inlet solids and their average diameter. The numerical method used to solve the transport equations is based on the finite volume method (FVM) formulated for a general non-orthogonal coordinate system [26–28]. The numerical mesh is constructed based on a body fitted coordinate system which allows accurate description of the blast furnace wall shape [26–28]. To solve the governing (momentum) equations of continuous phases the SIMPLE (Semi-Implicit Method for Pressure-Linked Equations) algorithm is applied on a staggered grid for covariant projections of the velocities and the numerical coefficients of the discretized equations are determined by using the power low scheme [26–28]. The convergence is assumed for momentum and energy equations of the five phases with the maximum relative error less than 10−3 while for all chemical species overall error less than 1% were accepted for finishing the calculations.

2.1.3Treatment of the source terms due to mutual interactions (Sϕk)

The treatment of the source terms is the key step of the accuracy of the multiphase model concept. The source terms in the conservation equations take into account chemical reactions, phase transformations, momentum exchange, external force, etc. [14–22]. The continuity and species conservation equations have mass sources due to chemical reactions and phase transformations. Enthalpy sources arise from inter-phase heat transfer, heat of reaction and sensible heat accompanied with the mass transfer due to chemical reactions and phase transformations [14–17]. The formulations for the phase interactions and chemical reactions have been published in previous reports [3–5,11–25]. The chemical reactions considered in this model are presented in Table 2. The momentum and energy sources are considered by taking into account the solid layer structure and cohesive zone formed depending on the soft-melting properties of individual burden [13,14].

Table 2.

Main chemical reactions considered in this multiphase model.

C(i)+O2(g)CO2(g)  C(i)+12O2(g)CO(g)   
C(i)+O2(g)2CO2(g)  C(i)+H2O(g)CO(g)+H2(g)   
S2(i)+2 O2(g)2 SO2(g)  S2(i)+O2(g)2 SO(g)   
CO2(g)+C(i)2CO(g)  SO2(g)+0.5S2(i)2SO(g)   
H2O(g)+C(i)H2(g)+CO(g)  SiO2(i)+3C(i)SiC(i)+2CO(g)   
SiO2(i)+C(i)SiO(g)+CO(g)  SiC(i)+CO(g)SiO(g)+2C(i)   
Volatiles (i)+α1O2α2CO2(g)+α3H2O(g)+α4N2(g) 
Volatiles (i)+α5CO2(g)α6CO(g)+α7H2(g)+α8N2(g) 
Note: the stoichiometric coefficients depends on the elemental analysis of the carbonaceous materials 
(i=coke, small coke and pulverized coal) 
CaCO3CaO+CO2(g)  MgCO3MgO+CO2(g)   
Fe2O3 (i)+H2(g)Fe3O4(i)+H2O(g)  Fe2O3 (i)+CO(g)Fe3O4(i)+CO2(g)   
w4w−3Fe3O4(i)+H2(g)→34w−3FewO(i)+H2O(g)  w4w−3Fe3O4(i)+CO(g)→34w−3FewO(i)+CO2(g)   
FewO (i)+H2(g)wFe(i)+H2O(g)  FewO (i)+CO(g)wFe(i)+CO2(g)   
(i=lump ore, sinter and pellets)     
H2O(i)H2O(g)  2CaO(i)CaO(s lg)   
MnO(i)MnO(s lg)  SiO2(i)SiO2(s lg)   
A2O3(i)A2O3 (s lg)  MgO(i)MgO(s lg)   
FexOy(i)FexOy(s lg)  SiO2 (s lg)+C (i)SiO(g)+CO (g)   
SiO(g)+C(hm)Si(hm)+CO(gMnO(s lg)+C(hm)Mn(hm)+CO(g)   
SO(g)+C(hm)S(hm)+CO(g2MnO (s lg)+Si (hm)2Mn(hm)+SiO2 (s lg)   
SO(g)+Si(hm)S(hm)+SiO(gP2O5(i)P2O5(s lg)   
CaO(s lg)+2S (hm)CaS(s lg)+SO(gP2O5 (s lg)+5C(hm)2P(hm)+5CO(g)   
(i=coke, small coke, lump ore, sinter and pellets)
Gaseous reactions
CH4+12O2→CO+2H2  C2H6+O22CO+3H2  C3H8+32O2→3CO+4H2 
C4H10+2O24CO+5H2  CO+12O2→CO2  H2+12O2→H2O 
CH4+CO22CO+2H2  C2H6+2CO24CO+3H2  C3H8+3CO26CO+4H2 
CH4+2O2CO2+2H2O  C2H6+5O22CO2+3H2O  C3H8+5O23CO2+4H2O 
C4H10+4CO28CO+5H2  C4H10+132O2→4CO2+5H2O  2SO2+3H22H2O+H2S 
CO2 (g)+H2(g)CO(g)+H2O(g) (gas equilibrium)     

Castro et al.

2.2Model validation using industrial data of large BF

The proposed model has been previously validated by using measured data obtained in an industrial blast furnace which has working volume of 3970m3 and instrumentations based on temperature and gas composition probes at the burden surface level [4,18,24]. The burden distribution is assumed in the model for the charging materials with radial distribution of mean solid diameter and their volume fractions. Fig. 4 shows steady state results for tridimensional raceway sector of influence and the burden materials distribution charged with their radial mean diameter distribution and charging pattern for the base case of 220kgthm−1 of pulverized coal and the respective inner temperature distribution and gas flow pattern. The charging distribution is iteratively calculated together with the temperature and gas flow pattern. The iterative procedure is driven by the measured temperature and gas composition profiles obtained by probing at the burden surface with small corrections for the radial distributions of the average burden particle diameters and materials volume fractions. As observed, central charging of granular coke was promoted and the temperature and gas flow reflect this charging condition and vice versa. In order to validate the model, averaged input data of 24h operation was used and the probe measurements for radial distribution temperature and gas composition were monitored in a total of 24 runs corresponding 24 interval of 1h. The averaged temperature and composition of CO and CO2 predicted by the model were used to correct the burden distributions patterns. The final accepted calculated results were compared with the probe data and showed excellent concordance, as can be seen in Fig. 5(a) and (b), respectively. Besides, the general trend of the monitored data reflects the central gas flow operation practiced in the charge system with granular coke charged mostly at the central region of the furnace.

Fig. 4.
(0.46MB).

Tridimensional raceway sector evidencing gas phase trajectories and burden radial distribution for the simulation of base case (PC: 220kgthm−1).

Fig. 5.
(0.22MB).

Off gas probe measurements for temperature and gas composition compared with model calculations for PC: 220kgthm−1.

3Results and discussions3.1Model verification for actual BF practice

Table 3 (column 1 and 2) summarizes the comparison for model predictions of global parameters and averaged industrial data of the furnace used in this study as the reference case. As can be observed, good agreement was obtained for all compared parameters. The model was applied for two sets of pulverized coal injections mixed with fuel gas. The cases 1–4 are the actual injection practice of pulverized coal with the newly suggested practice of additional injections of H2, O2 and CO2, while cases 5–10 are 250kgthm−1 of PC mixed with H2, O2 and CO2 injection. Table 3 shows the operational parameters predictions for the 220kgthm−1 injection cases, while Table 4 shows the parameters for 250kgthm−1 injection cases. As can be observed, for all cases of H2, O2 and CO2 injection, the granular coke or small coke charged at the burden decreased due to the replacement of granular coke by hydrogen fuel. Tables 3 and 4 also presented the computed values of carbon consumption on the lump coke and pulverized materials for solution loss reactions and water gas reactions, the main reactions that can take place in the PC replacing granular coke. As observed, the solution loss and water gas reactions were enhanced with the mixed injections of H2, O2, CO2 and PC in the tuyeres. It is very important to mention that for all cases of gas injection the net specific carbon emission on the blast furnace off gas decreased, which confirmed the effectiveness of these operation techniques to reduce greenhouse emissions, the main focus of this study. The reduction of the carbon emissions is attributed to the combined effect of coke reduction, CO2 recycling and enhancement of the blast furnace efficiency. It is also worthy to mention that the productivity of the blast furnace largely increased due to the additional oxygen injection and the replacement of the granular coke into the burden, which relatively increase the iron-bearing materials volume fractions. Figs. 6 and 7 show the burden distributions used in these calculations. As can be observed, the gas injection practice for the scenarios proposed did not required significant changes in the charging practice, which is a great advantage. Only the ratio of iron bearing materials and coke was changed to attain the solid inflow rates imposed by the increase in the productivity and therefore only the relative positions of the volume fractions curves shift upward to account for the mass flow rates and increase of the ore to coke ratios on the burden.

Table 3.

Comparisons of operational parameters calculated and model predictions for mix injections H2, O2 and CO2 with PC.

  BasePCI injection: 220kg/thm (actual injection practice)
  Calculated  Measured  Case 1  Case 2  Case 3  Case 4 
Productivity (thmday−1m−3 of working volume)  2.40  2.40  2.60  3.20  3.32  3.50 
Oxygen enrichment (%) (input)  7.34  7.30  7.34  7.34  7.34  7.34 
Blast (Nm3thm−1) (input)  944  944  843.3  641.2  627.2  622.3 
H2 injection (Nm3thm−1–  –  93.6  224.7  146.5  148.3 
CO2 injection (Nm3thm−1–  –  –  –  26.6  28.4 
O2 injection (Nm3thm−1–  –  46.1  110.6  72.2  78.2 
Raceway maximum temperature (°C)  2091  –  2272  2472  2241  2413 
Si  0.30  0.33  0.38  0.50  0.27  0.35 
Slag rate (kgthm−1250.5  254.4  242.8  247.5  253.5  252.3 
CaO/SiO2  1.29  1.25  1.43  1.32  1.26  1.24 
CaO (%)  43.9  –  45.4  44.6  43.9  44.2 
MgO (%)  8.8  –  9.1  8.7  8.5  8.6 
Al2O3 (%)  13.2  –  13.7  13.2  12.5  12.2 
Off gas (Nm3thm−11569  1515.8  1578.6  1513.3  1398.5  1422.1 
CO2/(CO+CO20.42  0.46  0.41  0.41  0.45  0.46 
CO2 (Nm3thm−1306.4  –  291.4  274.5  301.4  295.3 
CO (Nm3thm−1422.4  –  423.5  394.7  359.4  356.4 
H2 (Nm3thm−160.2  –  96.5  135.5  84.4  82.3 
N2 (Nm3thm−1646.9  –  579.4  443.1  429.6  423.2 
Temperature (°C)  164.8  156.5  169.2  182.4  170.9  169.3 
Coke consumption (kgthm−1219.9  215.5  216.4  201.3  201.8  200.3 
Small coke (kgthm−165.3  65.0  65.9  56.3  25.5  24.3 
PC (kgthm−1219.4  220.0  219.5  220.1  219.8  219.6 
Coke solution loss (kgCthm−165.1  –  59.3  68.5  70.3  69.2 
Coke water gas reaction (kgCthm−140.8  –  47.5  43.8  32.2  33.3 
Coke direct reduction (kgCthm−143.6  –  43.5  52.3  50.8  48.3 
PC solution loss (kgCthm−160.5  –  54.1  36.1  40.6  43.2 
PC water gas reaction (kgCthm−168.6  –  69.6  79.3  75.3  76.4 
PC direct reduction (kgCthm−10.0  –  0.2  0.8  2.5  2.8 
PC combustion efficiency (%)  99.6    99.7  100  100  100 
Specific carbon emission (kgCthm−1385.9  –  383.1  358.7  339.7  336.2 

PC, pulverized coal; thm, ton of hot metal (pig iron).

Castro et al.

Table 4.

Calculated operational parameters for H2 and CO2 mix injection with high PC injection.

  Only H2 injectionCO2 and H2 injections
  Case 5  Case 6  Case 7  Case 8  Case 9  Case 10 
Productivity (thmday−1m−3 of working volume)  2.63  2.72  2.95  3.35  3.41  3.94 
Oxygen enrichment (%) (input)  9.96  9.96  9.96  9.96  9.96  9.96 
Blast (Nm3thm−1) (input)  859.5  833.3  748.3  594.8  564.3  445.7 
H2 injection (Nm3thm−18.8  82.2  217.2  214.5  246.9 
CO2 injection (Nm3thm−11.2  48.7  84.2 
O2 injection (Nm3thm−14.4  44.6  111.6  109.4  121.6 
Raceway maximum temperature (°C)  2106  2129  2088  2262  2514  2324 
Si  0.32  0.28  0.41  0.47  0.33  0.31 
Slag rate (kgthm−1238  246  229  242  247  256 
CaO/SiO2  1.47  1.32  1.46  1.57  1.38  1.35 
CaO (%)  45.7  44.3  45.5  46.9  45.1  44.9 
MgO (%)  9.4  9.0  9.3  9.3  8.9  8.7 
Al2O3 (%)  13.7  13.3  13.9  13.5  14.0  12.9 
Off gas (Nm3thm−11481  1449.8  1464  1446  1476  1439 
CO2/(CO+CO20.44  0.42  0.42  0.45  0.45  0.42 
CO2 (Nm3thm−1299.7  300.0  309.2  287.1  295.2  298.1 
CO (Nm3thm−1398.8  391.7  397.6  398.4  352.1  417.3 
H2 (Nm3thm−163.9  66.8  61.8  93.3  134.5  135.4 
N2 (Nm3thm−1546.1  559.2  571.7  498.9  397.9  300.0 
Coke consumption (kgthm−1193.5  193.1  188.5  184.1  185.2  184.7 
Small coke (kgthm−153.1  52.9  52.2  26.2  26.7  26.7 
PC – pulverized coal (kgthm−1249.4  249.1  249.4  249.7  249.7  249.1 
Coke solution loss (kgCthm−157.9  55.4  54.4  51.5  56.5  57.5 
Coke water gas reaction (kgCthm−136.5  34.8  39.9  39.6  36.7  30.7 
Coke direct reduction (kgCthm−145.1  43.6  44.9  45.1  44.3  45.8 
PC solution loss (kgCthm−147.2  47.2  40.2  22.9  45.2  62.6 
PC water gas reaction (kgCthm−165.1  75.2  66.2  71.8  66.8  86.3 
PC direct reduction (kgCthm−10.1  0.3  0.6  0.5  8.1  1.4 
PC combustion efficiency (%)  100  100  100  100  100  100 
Net carbon emission (kgCthm−1378.5  369.1  367.2  346.7  346.4  338.1 

PC, pulverized coal; thm, ton of hot metal (pig iron).

Castro et al.

Fig. 6.
(0.34MB).

Steady burden distributions predictions for mix injection of H2, O2 and CO2 with PC: 220kgthm−1.

Fig. 7.
(0.33MB).

Calculated burden distributions for mix injection of H2, O2 and CO2 with PC: 250kgthm−1.

3.2Comparison of the in-furnace variables for mix injection in the raceway

In this section, the model is applied to investigate new operational conditions of H2, O2 and CO2 mixed with higher injection rates and compared with base case. Two sets of scenarios were design: (i) actual operational practice of 220kgthm−1 of pulverized coal injection with additional injection of H2, O2 and CO2 in the raceway and (ii) increasing the pulverized coal injection to 250kgthm−1 and additional injection of H2, O2 and CO2 in the raceway. The results discussed in this section were obtained by iterative calculations until stable phase flows and temperature pattern are reached and therefore represents predicted feasible operational conditions for the blast furnace operation. Fig. 8 shows the temperature and gas flow pattern for the first set of scenarios. As can be observed cases 1 and 2 (only H2 and O2 injections) considerably increased the temperature of raceway region and slowly increase the dropping zone temperature region but the changes in the shaft region temperature were negligible. Cases 3 and 4 (combined injection of H2, O2 and CO2) allows the CO2 recycling by using the lower part of the blast furnace excess energy and enhance the productivity with lower blast rate (see Tables 3 and 4). As can be observed in all cases the temperature patterns inside of the blast furnace are compatible with stable operation and therefore could be practiced in this reactor with the advantage of considerably decreasing the consumption of granular reducing agent and specific carbon emissions (see Tables 3 and 4). Fig. 9 shows the temperature patterns and gas flow distribution in a vertical plane passing through the raceway for the second set of scenarios with pulverized coal 250kgthm−1. As can be observed, cases 6–7 (only H2 and O2 injection), similarly to previous calculation of 220kgthm−1, increased the temperature due to the heat input promoted by H2 combustion. Cases 9–10 restored the temperature and gas flows by using the excess energy of H2 combustion to reform the CO2 injected. Figs. 10 and 11 shows the solid descending flow paths, residence time of the lump materials and the calculated position of the cohesive zone for the two sets of scenarios considered in this study. In this model the cohesive zone definition is not predetermined, instead, the softening-melting burden parameters are used, mainly the initial softening and melting temperature, interval of melting temperature, maximum pressure drop and the degree of shrinkage obtained in laboratory burden characterization that are used in all sub-models of gas flow resistance, heat transfer in the mushy zone and chemical reactions. Therefore, the actual position of the cohesive zones, thickness and shape are the result of coupled phenomena taking place in the whole blast furnace process. In all simulations the same set of burden materials parameters were used (Tsm – temperature of starting melting of 1228°C, shrinkage factor of 0.35 and interval of melting temperature for maximum pressure drop of 185°C). As can be observed, as the amount of injection gas was increased the equilibrium position and shape of the cohesive zones shift up and became thicker. Fig. 11 presents similar pattern and showed that the residence time of the granular materials in the shaft region decreased considerably due to increase in the productivity. For all scenarios proposed in this study, the model predicted the necessary parameters to be safety carried out in the blast furnace practice under mixed injection of H2, O2 and CO2 with PC. Therefore, all of the scenarios proposed could be used to enhance the blast furnace operation combining high productivity, simultaneous injection of pulverized coal, H2, O2 and CO2 with drastic decreasing of the blast rate, granular solid reducing agent and CO2 emissions. These features definitively would enhance the blast furnace performance not only from the perspective of productivity but also from the point of view of cleaner technology developments of ironmaking route.

Fig. 8.
(0.77MB).

Predicted gas flow pattern and temperature distribution for a raceway vertical plane for mix injection of H2, O2 and CO2 with PC: 220kgthm−1.

Fig. 9.
(0.73MB).

Calculated gas flow pattern and temperature distribution for a raceway vertical plane for mix injection and PC: 250kgthm−1.

Fig. 10.
(0.74MB).

Calculated cohesive zone location and solid flow descending pattern for a raceway sector for mix injection and PC: 220kgthm−1.

Fig. 11.
(0.75MB).

Calculated shape of cohesive zone and solid flow descending pattern for a raceway sector for mix injection and PC: 250kgthm−1.

4Conclusions

We presented improvements of the multiphase mathematical model for simulating the blast furnace with new features and applied to predict the new furnace condition under mix injections of PC and gaseous fuel enriched on H2, O2 and CO2 with the aim of reaching new possibilities for the process with lower environmental impact. The improved model uses multiphase interactions formulated for momentum, energy and chemical species. The model takes into account simultaneously in all phases the new chemical reactions occurred in the process due to new materials injections. The model predictions were compared favorably with a continuous monitoring BF data measured for a smooth operation period. In the sequence, new operation practice techniques based on higher mix injection of PC and rich hydrogen, oxygen and carbon dioxide fuels were simulated. The model predictions newly indicated that important decrease in the granular coke consumption in the BF process is possible by proposing new mix injections of 250kgthm−1 of PC and correspondent H2 and O2 injection to keep raceway temperature and dropping zone with stable liquids and gas flow conditions. Furthermore, it was found that CO2 injection is possible due to high temperature developed with H2 and O2 injection and for all simulated results the model indicated that lower carbon emissions can be achieved by the proposed operational conditions.

Conflicts of interest

The authors declare no conflicts of interest.

Acknowledgements

The first author would like to thanks CNPq (Universal and PQ) and Faperj (CNE-2014-2016) for partial support of this research.

References
[1]
R. Cyprés
Modern carbochemical processes for hydrogen production from coal
Int J Hydrogen Energy, 12 (1987), pp. 451-460
[2]
M. Ni,D.Y.C. Leung,M.K.H. Leung,K. Sumathy
An overview of hydrogen production from biomass
Fuel Process Technol, 87 (2006), pp. 461-472
[3]
J.A. Castro,H. Nogami,J. Yagi
Numerical investigation of simultaneous injection of pulverized coal and natural gas with oxygen enrichment to the blast furnace
ISIJ Int, 42 (2002), pp. 1203-1211
[4]
J.A. Castro,G.M. Araujo,I.O. Mota,Y. Sasaki,J. Yagi
Analysis of the combined injection of pulverized coal and charcoal into large blast furnaces
J Mater Res Technol, 2 (2013), pp. 308-314
[5]
E.P. Rocha,V.S. Guilherme,J.A. Castro,Y. Sasaki,J. Yagi
Analysis of synthetic natural gas injection into charcoal blast furnace
J Mater Res Technol, 2 (2013), pp. 255-262
[6]
J.G.M.S. Machado,E. Osorio,A.C.F. Vilela,A. Babich,D. Senk,H.W. Gudenau
Reactivity and conversion behaviour of Brazilian and imported coals, charcoal and blends in view of their injection into blast furnaces
Steel Res Int, 81 (2010), pp. 9-16
[7]
Y. Qie,Q. Liu,J. Li,C. Lan
Effect of hydrogen addition on reduction kinetics of iron oxides in gas-injection B
ISIJ Int, 57 (2017), pp. 404-412
[8]
J.A. Castro,H. Nogami,J. Yagi
Numerical analysis of multiple injection of pulverized coal, pre-reduced iron ore and flux with oxygen enrichment to the blast furnace
ISIJ Int, 41 (2001), pp. 18-24
[9]
K. Ichikawa,Y. Kashihara,N. Oyama,T. Hirosawa,J. Ishii,M. Sata
Evaluating effect of coke layer thickness on permeability by pressure drop estimation model
ISIJ Int, 57 (2017), pp. 254-261
[10]
J.A. Castro,A.W.S. Baltazar
Estudo numérico da reciclagem de CO2 na zona de combustão do alto forno
TMM, 6 (2009), pp. 13-18
[11]
P.S. Assis,W.B. Martins,C.B. Vieira
Avanços na injeção de carvão pulverizado para a sua aplicação em altos-fornos
REM, 56 (2003), pp. 281-285
[12]
J.A. Castro,A.J. Silva,H. Nogami,J. Yagi
Simulação computacional da injeção de carvão pulverizado nas ventaneiras de mini altos-fornos
Tecnol Metal Mater Miner, 1 (2004), pp. 59-62
[13]
Y. Shen,B. Guo,S. Chew,P. Austin,A. Yu
Modeling of internal state and performance of an ironmaking blast furnace: slot vs sector geometries
Metall Mater Trans B, 47B (2016), pp. 1052-1062
[14]
Z. Dong,Q. Xue,H. Zuo,X. She,J. Li,J. Wang
Gas-solid flow and shaft injected gas penetration in an oxygen blast furnace analyzed using a three-dimensional DEM-CFD coupling mathematical model
ISIJ Int, 56 (2016), pp. 1588-1597
[15]
J. Yagi
Mathematical modeling of the flow of four fluids in a packed bed
ISIJ Int, 33 (1993), pp. 619-639
[16]
P.R. Austin,H. Nogami,J. Yagi
A mathematical model for blast furnace reactions analysis based on the four fluid model
ISIJ Int, 37 (1997), pp. 748-755
[17]
J.A. Castro,H. Nogami,J. Yagi
Transient mathematical model of blast furnace based on multi-fluid concept, with application to high PCI operation
ISIJ Int, 40 (2000), pp. 637-646
[18]
J.A. Castro,H. Nogami,J. Yagi
Three-dimensional multiphase mathematical modeling of the blast furnace based on the multifluid model
ISIJ Int, 42 (2002), pp. 44-52
[19]
X.F. Dong,A.B. Yu,S.J. Chew,P. Zulli
Modelling of blast furnace with layered cohesive zone
Metall Mater Trans B, 41B (2010), pp. 330-349
[20]
S. Ueda,S. Natsui,H. Nogami,J. Yagi,T. Ariyama
Recent progress and future perspective on mathematical modeling of blast furnace
ISIJ Int, 50 (2010), pp. 914-923
[21]
A.T. Adema,Y. Yang,R. Boom
Discrete element method-computational fluid dynamic simulation of the materials flow in an iron-making blast furnace
ISIJ Int, 50 (2010), pp. 954-961
[22]
Z. Zhou,H. Zhu,A. Yu,B. Wright,D. Pinson,P. Zulli
Gas-solid flow in an ironmaking blast furnace-II: discrete particle simulation
Powder Technol, 208 (2011), pp. 72-85
[23]
A.W.S. Baltazar,J.A. Castro,A.J. Silva
Modelagem e simulação computacional da injeção de carvão pulverizado no canal adutor da raceway de altos-fornos
Rev Estud Tecnol, 2 (2006), pp. 65-77
[24]
J.A. Castro,Y.A.J. Silva,Y. Sasaki,J. Yagi
A six-phases 3-D model to study simultaneous injection of high rates of pulverized coal and charcoal into the blast furnace with oxygen enrichment
ISIJ Int, 51 (2011), pp. 748-758
[25]
Y. kato
Carbon recycling for reduction of carbon dioxide emission from iron-making process
ISIJ Int, 50 (2010), pp. 181-185
[26]
K.C. Karki,S.V. Patankar
Calculation procedure for viscous incompressible flows in complex geometries
Numer Heat Transf B, 14 (1988), pp. 295-307
[27]
S.V. Patankar,D.B. Spalding
A calculation procedure for heat, mass and momentum transfer in three-dimensional parabolic flows
Int J Heat Mass Transf, 15 (1972), pp. 1787-1806
[28]
M.C. Melaen
Calculation of fluid flows with staggered and non-staggered curvilinear nonorthogonal grids-the theory
Numer Heat Transf B, 21 (1992), pp. 1-19
Corresponding author. (Jose Adilson de Castro joseadilsoncastro@id.uff.br)
J Mater Res Technol 2017;6:258-70 DOI: 10.1016/j.jmrt.2017.05.007