Introduction

Blast furnace (BF) is the most significant ironmaking technology to produce hot metal from ferrous materials. However, the BF ironmaking process is responsible for about 70 pct of the energy consumption and CO2 emission in an integrated works.[1] Moreover, the minimum total fuel rate in the form of coal and coke tends to approach the limit for traditional blast furnaces (TBFs).[2] Therefore, TBFs are difficult to achieve significant carbon consumption and CO2 emission reduction. This situation raises an imminent need for new BF ironmaking technologies so that CO2 emissions will be reduced substantially. Such development will also help slow down the consumption of coke, which is indispensable in BF ironmaking but has depleting resources on the earth.

In recent years, various innovative low-carbon BF ironmaking technologies have been proposed. Some examples include iron–carbon agglomerates, biomass and waste plastic injection, reducing gas injection, and high and even full oxygen/hydrogen enrichment.[3] In the reducing gas injection technology, CO- and/or H2-enriched gases, e.g., coke oven gas (COG), reformed COG (RCOG),[4,5,6] natural gas,[7,8,9] top recycled gas,[10,11,12,13,14] and high-purity hydrogen[15,16,17,18] are injected into BFs from the hearth and/or the shaft tuyeres. Generally, shaft injection operations are considered useful in utilizing different reducing gas resources to adjust heat balance and improve the reduction potential in the furnace shaft.[19,20] Therefore, it is helpful to enhance BF efficiency and reduce CO2 emissions. However, this technology suffers some inherent deficiencies. Specifically, it has been recognized that the gas injected into the shaft has shallow penetration, which could compromise the benefits of shaft injection.[5,6] In addition, it has been argued in the research community that shaft gas injection could disturb the main flow stream inside a BF and the associated thermochemical behaviors, thereby reducing BF performance. Therefore, in-depth knowledge of the impact of shaft gas injection on BF performance is highly desired.

Numerous experimental efforts have been conducted previously to test the shaft gas injection technology mainly based on oxygen blast furnaces (OBFs) (see, e.g., References 21,22, through 23). Such studies aimed to test the feasibility of OBF, where shaft gas injection offers a measure of overcoming the excessively low temperature (i.e., thermal shortage) in the OBF shaft.[24] To date, only a few experimental studies have been reported to analyze shaft injection technology itself. For instance, Usui et al.[25] developed a lab-scale apparatus to quantify the influence of hydrogen injected from the BF shaft. They reported that the indirect reduction could be intensified under shaft injection conditions. Higuchi et al.[26] used a softening-melting tester and a counter-current reaction simulator to study how RCOG shaft injection influenced the reduction behavior of sintered ores. The results indicated that the injection rate of RCOG should be above 120 Nm3/tHM to minimize the reduction degree in the radial direction due to the poor penetration depth by the shaft injection. Additionally, the Japanese COURSE50 program tested the shaft injection of RCOG into a 9-m3 experimental BF (EBF).[5] The results revealed that the ratio of the gas penetration depth to the furnace radius at the shaft injection tuyere level was only 0.22 to 0.28 for this small EBF.

Although useful, the above experimental studies were mainly conducted under laboratory or experimental conditions. Shaft gas injection operations have rarely been tested based on industrial BFs. To date, this technology has not been generally industrialized. Therefore, developing a viable method to explore shaft gas injection operations under industrial conditions is necessary. In particular, this exploration at an industrial level could be extremely costly. Experimentally, BF inner states are difficult to measure to underpin new technology development. Therefore, numerical models have been widely employed to explore BFs.[27,28,29,30] As reviewed below, some numerical efforts have also been made for the shaft gas injection technology.

Natsui et al.[31,32] studied shaft gas injection penetration using a three-dimensional (3D) approach of computational fluid dynamics and discrete element method (CFD-DEM). They reported that the penetration was restricted to the peripheral region near the injection level, and the ascending gas was pushed to the center at a higher SIG flow rate. Using a 2D CFD-DEM model, Dong et al.[33] studied the influences of particle diameter, shaft tuyere size, and the ratio of gas injection flow rates through shaft and hearth tuyeres. Later, they extended the model from 2 to 3D to study the effect of penetration behavior in the radial direction according to gas pressure and velocity vectors.[34] These CFD-DEM studies have generated insights into shaft gas injection under cold conditions. However, real BFs are more complicated than this, involving intensive interactions in flow and thermochemical behaviors among gas, solid, and liquid phases. Recently, a CFD-DEM process model has been developed to describe these complex phenomena.[35,36,37] Because of extremely high computational requirements, such a model cannot simulate an industrial BF.

Compared with CFD-DEM models, CFD BF process models are more suitable for studying the flow and thermochemical behaviors and overall performance of BFs at an industrial scale.[27,30,38] These models have been used to analyze shaft gas injection technology under various operating conditions. For instance, Zhang et al.[39] investigated how shaft gas injection influenced the inner states of OBF and observed an uneven gas concentration distribution in the radial direction. A few investigators studied the effects of various shaft-injected gases, e.g., COG, top off-gas, hydrogen, etc., on the inner profiles and global performance of BF.[14,16,40,41,42,43] Li et al.[6] quantified the penetration depths referring to different inner state variables in the OBFs with belly injection of COG and/or hot burden charging. All these CFD studies analyzed the behavior of shaft-injected gas lumped with hearth-generated gas, herein considering the collective effect on BF performance. Because of the lack of effective tools, how the two gases contribute to a BF differently has not been investigated. However, this information is highly desired to explore the shaft injection technology.

The other problem associated with the previous CFD studies of shaft gas injection is the use of nontrivial model simplifications. For example, most numerical models were 2D.[6,14,16,39,40,43] However, shaft gas injection is often realized via a few discrete tuyeres. This inevitably generates 3D BF inner states, especially in the injection region. To date, it is unclear how these 3D characteristics affect BF performance. In addition, a few models did not model the liquid flow,[16,39] affecting the prediction of the effect of shaft gas injection.

In this paper, shaft gas injection operations are studied with respect to three variables related to gas penetration: shaft-injected gas (SIG) flow rate, shaft tuyere number, and horizontal cavity depth. This is done using a validated 3D parallel CFD process BF model.[44,45] Based on this model, a method is proposed to track SIG and hearth-generated gas (HGG). Via this method, the SIG penetration effect is directly quantified. Furthermore, the contributions of SIG and HGG to iron ore reduction are isolated and linked with the change in overall performance. Apparently, these efforts are different from our previous study,[43] which studied the effect of SIG on BF performance using a 2D process model. The current study elucidates the roles of SIG and HGG and their interactions to underpin the development of shaft injection technology.

Model Description

A validated 3D steady-state multi-fluid model is used and further developed in this work. The model is detailed elsewhere[44,45] and outlined here for brevity. This model simulates the region between the burden and slag surfaces for the flow and thermochemical phenomena of gas, solid, and liquid phases. Each phase comprised of one or more components is treated as a continuous fluid. Figure 1 presents the framework of the model, integrated with various modeling efforts[6,44,45,46,47,48] to be realistic. For example, the modeling of stockline variation enables the burden charge rate to be automatically adjusted based on the carbon balance. Thus, productivity, as the simulation output depending on BF conditions, can be predicted directly.[6,49]

Fig. 1
figure 1

Framework of (a) the CFD BF process model and (b) the sub-model determining the deadman profile

Full size image

A new method is proposed here to track the components of SIG and HGG and quantify their respective contributions to BF performance. To this effect, extra scalar equations are solved to consider each component of SIG and HGG. This is different from the previous CFD process models, where the same component in SIG and HGG shares the same scalar equation. In total, nine scalar equations are solved in the new gas tracking method to consider CO, CO2, H2, H2O, and N2 in the HGG and CO, CO2, H2, and H2O in the SIG:

$$ \nabla \cdot \left( {\varepsilon_{i} \rho_{i} {\mathbf{u}}_{i} \varphi_{i,m} } \right) - \nabla \cdot \left( {\varepsilon_{i} \Gamma_{i} \nabla \varphi_{i,m} } \right) = \sum\nolimits_{k} {\alpha_{i,m,k} R_{k}^{*} } r_{{{\text{SIG}} - {\text{HGG}},m}}, $$
(1)

where \(r_{{{\text{SIG}} - {\text{HGG}},m}}\) represents the fraction of the kth reaction contributing to the mth component in the SIG, and 1-\(r_{{{\text{SIG}} - {\text{HGG}}}}\) is the fraction contributing to the same component in the HGG. As done in all the existing process BF models, SIG and HGG are assumed to be well mixed in a CFD cell. Therefore, \(r_{{{\text{SIG}} - {\text{HGG}}}}\) can be given by

$$ r_{{{\text{SIG}} - {\text{HGG}},m}} = C_{m} /(C_{{{\text{SIG}}}} + C_{{{\text{HGG}}}} ), $$
(2)

where \(C_{m}\) is the volume fraction of the mth component in the SIG in the CFD cell, \(C_{{{\text{SIG}}}}\) and \(C_{{{\text{HGG}}}}\) are the SIG and HGG volume fractions. Equations [1] and [2] are solved using a postprocessor to avoid the structure change of the in-house code. It also helps alleviate the computational loadings. This gas tracking method is different from the solution of gas components in the main code, which considers only five scalar equations for gas components. It allows for tracking SIG and HGG and quantifying their respective roles.

Boundary Conditions

As detailed previously, this study considers an industrial BF (380-m3) with 14 tuyeres.[50] This is an furnace used by our industrial partners to develop OBF, whose productivity is high and thus can achieve the similar production achieved by a much larger conventional BF. The computational domain, a half-section of the BF including seven hearth tuyeres, is shown in Figure 2. The number of shaft tuyeres varies from 1 to 5. In the model, the axial plane is symmetric. Like the previous studies,[44,50,51] a center-developed burden distribution is applied at the burden surface, where more coke particles are loaded at the center of the furnace. The burden profile keeps the same in the circumferential direction. As a result, the central region in the bed has better permeability for gas flow than the peripheral region.

Fig. 2
figure 2

Computational domain in the present study

Full size image

As listed in Table I, the blast injected through the hearth tuyeres into the BF has a temperature of 1273 K (1000 °C). The oxygen enrichment is 35 pct, selected according to industrial operations. Such oxygen enrichment is often considered for OBF, where shaft injection technology offers a possible solution to the thermal shortage problem. The combustion of coal and coke with the oxygen-enriched hot air in the raceway region leads to the HGG (i.e., bosh gas). The HGG mass flow rate, temperature, and components are used as the inlet gas conditions at the hearth tuyeres. This study determines this information by the local mass and heat balance method with given conditions, as done previously.[6,44,45,46,48,49,50] The HGG flow rate and temperature are fixed. In other words, the same flame temperature and bosh gas volume and components are maintained. Pure pre-heated CO with a temperature of 1073 K (800 °C) is injected from the shaft tuyeres into the BF at 5.2 m in height.

Table I BF Operating Conditions in This Study
Full size table

Table I lists the range of the three variables related to SIG penetration into the particle bed considered in this work: SIG flow rate, shaft tuyere number, and horizontal cavity depth. The previous studies[31,32] suggest that a cavity in front of the shaft tuyere is small. Therefore, the cavity depth is set to zero in the base case. However, the effect of cavity depth is examined in the work. In addition, a uniform solid velocity in the radial direction is applied to the solid surface in the throat. The solid inlet velocity is set based on the mass balance under the given coke rate, burden distribution, productivity, and BF throat dimensions. Note that as the variable value changes, the BF inner states change. Correspondingly, the solid inlet velocity varies in a simulation to adapt to this change, leading to different productivity.

The model is solved using an in-house Fortran F90 code with the above boundary conditions, as detailed elsewhere.[44,45,48,50] Each simulation is allocated 140 CPUs and takes about 72 hour to generate converged results. The results are obtained by obeying CFD convergence criteria. Moreover, the variation of the CZ profile in two consecutive iterations, quantified by the ratio of the volumes of lumpy and dripping zones, needs to be under 0.1 pct.

Results and Discussion

Model Applicability

The current model has been validated through various applications, as found elsewhere.[44,45,48] The model validation is here extended for the 380-m3 industrial BF. Figure 3 shows the measured and calculated results in good agreement, with prediction errors of less than 1.3 pct.

Fig. 3
figure 3

Comparisons of measured and calculated overall performance indicators for an experimental 380-m3 BF

Full size image

The proposed gas tacking method is also validated in this work. As described in Section II, two methods are used to simulate gas components. One is the conventional method used in the main code, where five scalar equations describing gas components are solved. In this method, the same component in the HGG and SIG cannot be identified separately, the same as in all other models. The other is the new gas tacking method in the postprocessor, where nine scalar equations are solved, tracking each HGG/SIG component. These two methods must predict the same distribution of the same gas component, which is the base for validating the gas tracking method. Figure 4 shows that the two methods indeed give nearly the same distributions of total CO (Figures 4(c), (d)). More importantly, the distributions of the CO in SIG and HGG can be identified by the gas tracking method, as demonstrated in Figures 4(a) and (b). Moreover, it can demarcate the regions with and without SIG contribution to BF performance. Figure 5 shows the marked regions (colored in red) according to the values of COSIG concentration when they are larger than 0.01. The SIG contribution is negligible when the COSIG concentration is below 0.01.

Fig. 4
figure 4

Spatial distributions of COSIG (a), COHGG (b), and total CO (= COSIG + COHGG) given by (c) the gas tracking method and (d) the conventional method

Full size image
Fig. 5
figure 5

Marked regions where the SIG exists to contribute to BF performance

Full size image

Effect of SIG Flow Rate

Figure 6 shows how the SIG flow rate affects overall BF performance. Here, the bosh gas volume and compositions, as well as flame temperature, are fixed, so the gas and energy supply in the raceway region is the same in all cases. The HM temperature increases with increasing SIG flow rate; however, it slows down when the flow rate exceeds 5 Nm3/s (Figure 6(a)). This increase represents the potential of coke saving because a high HM temperature allows coke reduction.[49] In contrast, the change in productivity is opposite to that of HM temperature, which decreases first and then slows down (Figure 6(b)). Differently, an almost linear change is observed for top gas temperature, top gas utilization factor, and tuyere pressure drop. Specifically, a higher SIG flow rate leads to a higher top gas temperature, a reduced gas utilization factor, and a higher tuyere pressure drop (Figures 6(c), (d)). Based on the results in Figure 6, an optimum SIG flow rate (i.e., 5 Nm3/s here) can be identified, which leads to a relatively high HM temperature and a relatively low tuyere pressure. Note that a relatively low tuyere pressure is helpful to maintain smooth operations.

Fig. 6
figure 6

(a) Hot metal temperature, (b) productivity, (c) gas utilization factor, top gas temperature, and (d) tuyere pressure as a function of SIG flow rate

Full size image

The inner states governing the overall BF performance are analyzed to understand the shaft injection technology. Two representative SIG rates are considered to show the results for clarity, as given in Figure 7. It demonstrates that the CO concentration, contributed by both SIG and HGG, is enriched near the furnace wall, especially at the shaft injection level. It decreases due to the indirect reduction when the gas ascends toward the burden surface (Figure 7(a)). The SIG penetration is more considerable at a larger SIG flow rate, which is qualitatively consistent with the CFD-DEM results reported elsewhere.[31,32] Figures 7(b) and (c) show that the solid and gas temperature contours have similar distributions. The solid/gas temperature rises as the SIG flow rate increases, especially near the furnace wall. Corresponding to the temperature variation, the CZ, within a solid temperature range between 1473 K (1200 °C) and 1673 K (1400 °C), becomes higher at a higher SIG flow rate (Figure 7(d)). Overall, the CZ head elevates more significantly than the CZ root. Thus, the total CZ length becomes longer. Both the increased CZ length and SIG flow rate account for the increase in the tuyere pressure drop. In addition, at a higher SIG flow rate, the CZ layer in the peripheral region becomes shorter (horizontal) and thinner (vertical); however, it is thicker at the radial positions away from the furnace wall.

Fig. 7
figure 7

(a) CO concentration, (b) gas and (c) solid temperatures, and (d) gas pressure and cohesive zone at various SIG flow rates

Full size image

The zone above the shaft tuyere injection level is divided into two parts to examine the pre-heating effect of shaft injection at different heights. Here, the dry zone targeted by shaft injection is considered, i.e., the zone above the CZ. The average solid temperature is calculated for each part and plotted as a function of the SIG flow rate, as presented in Figure 8. The average temperature increases as the SIG flow rate increases. The increase in the upper part is largely linear. It is initially the case in the upper part, followed by slower growth. In other words, the pre-heating effect in the lower part reaches the limit at relatively large SIG flow rates. This should be because both the increased gas velocities and the reduced cross section of the dry zone reduce the SIG residence time in the lower part.

Fig. 8
figure 8

Average solid temperature in the upper and lower parts of the zone above the shaft injection tuyere level as a function of SIG flow rate

Full size image

Figure 9 shows the profiles of the mass flow rate and temperature of the liquid for the SIG flow rate of 2 Nm3/s and 6 Nm3/s. Here, the liquid flow is predicted by the force balance model coupled with a liquid dispersion model. An “icicle” flow pattern beneath the cohesive ore layer and a large mass flow rate near the wall can be observed. With the increase of the SIG flow rate, the icicle flow becomes more significant. Subsequently, the mass flow rate in the peripheral region becomes weaker. This phenomenon is attributed to the different portions of the ore layer in the CZ (see Figure 7). Owing to the increased liquid flow rate, the liquid temperature decreases in the central region at a higher SIG flow rate. Correspondingly, the temperature increases in the peripheral region. The higher liquid temperature near the wall at a more significant SIG flow rate accounts for the increased HM temperature observed in Figure 6(a).

Fig. 9
figure 9

(a) mass flow rate and (b) temperature of the liquid at different SIG flow rates

Full size image

The nonuniformity of liquid mass flow rate and temperature in the radial direction is analyzed quantitatively. In doing so, peripheral and central regions are identified by dividing the BF according to half of the normalized radius distance. The average values in these two regions at the slag surface are calculated. Overall, the mass flow rate ratio of liquid in the central to peripheral regions, as shown in Figure 10, is much smaller than one. This result means that the liquid flows more significantly in the peripheral region, which is weakened at a higher SIG flow rate, as suggested by the increased flow rate ratio. Conversely, the liquid temperature ratio is larger than one, meaning higher liquid temperature in the central region. This ratio decreases as the SIG flow rate increases.

Fig. 10
figure 10

Ratios of the liquid mass flow rates and temperatures in the central to peripheral regions at various SIG flow rates

Full size image

The proposed gas tracking method is used to analyze SIG penetration and the difference between SIG and HGG contributions to the pre-reduction. Figure 11 presents the thus obtained spatial distributions of indirect reduction rates by shaft-injected CO (COSIG), hearth-generated CO (COHGG), and total CO (= COSIG + COHGG) at two representative SIG flow rates. It reveals that the COSIG indirect reduction reaction occurs mainly near the furnace wall, where the COSIG concentration is relatively high (e.g., > 0.1). As the SIG flow rate increases, this reduction region becomes deeper in the radial direction and higher in the axial direction. Overall, the COHGG indirect reduction rate is relatively small in regions having relatively high COSIG concentrations. All these results suggest that the penetration of the SIG into the bed column is essential to shaft injection effectiveness. This is consistent with the previous study of gas injection,[6] where HGG and SIG were not separated. Without the separation of HGG and SIG, all the previous studies of shaft gas injection did not define SIG penetration depth precisely. They could neither quantify the contributions of SIG and HGG to BF performance to clarify their roles. As demonstrated below, these problems can be overcome by the new gas tracking method, generating new understandings of shaft injection technology.

Fig. 11
figure 11

Indirect reduction rates by COSIG (left) COHGG (middle) and total CO (right) at the SIG flow rate of (a) 2 Nm3/s and (b) 6 Nm3/s

Full size image

The SIG penetration depth relative to the furnace radius at a specific height is defined according to the value of COSIG. Moreover, the values of COSIG are divided into six groups: 0 to 0.01, 0.01 to 0.2, 0.2 to 0.4, 0.4 to 0.6, 0.6 to 0.8, and 0.8 to 1.0. This information is used to divide the BF into six regions. The first region (0 to 0.01) represents the zone where the effect of COSIG is zero or negligible. In contrast, the other five regions are where the SIG can reach and contribute to the BF performance to different extents. The penetration depth is calculated for each region with SIG contributions. The resulting results, as given in Figure 12, consider the two height levels (see Figure 8). The penetration depth largely increases linearly with the increase of COSIG at different SIG flow rates. This trend does not hold in the central region near the shaft tuyere level (H = 5.2 m) and the peripheral region at a relatively high height (e.g., H = 10.2 m here). This result suggests that the SIG concentration reduces nearly at the same rate. However, this rate is increased by the upward flow stream in the central region near the shaft tuyere level. Meanwhile, it is decreased by the wall resistance near the peripheral region at relatively high heights and relatively low SIG flow rates. Overall, with the increase of SIG flow rate, the maximum penetration depth increases and subsequently slows down. This trend is observed at both height levels.

Fig. 12
figure 12

Penetration depth as a function of COSIG at two different height levels for various SIG flow rates

Full size image

The indirect reduction rates contributed by the SIG and HGG are quantified for each region defined above, so the roles of the two gases can be examined spatially. As given in Figure 13(a), when the COSIG concentration increases, the COSIG indirect reduction rate increases before reaching a maximum and then decreases. The maximum rate occurs at COSIG concentrations between 0.2 and 0.8 and does not appear at the highest COSIG concentration. This should be attributed to the relatively low solid temperature near the wall. For this reason, the maximum COSIG indirect reduction rate moves toward the center at a lower SIG flow rate because less energy is brought in to heat the solid near the wall. Similarly, when the COSIG concentration increases, the COHGG indirect reduction rate increases first, then drops drastically, and finally decreases gradually (Figure 13(b)). However, the maximum value appears at the COSIG concentration of about 0.2, located close to the central region. Overall, the SIG contributes more significantly to the indirect reduction rate than the HGG when the COSIG concentration exceeds 0.2, which reverses at concentrations smaller than 0.2. This result quantitatively demonstrates how the COHGG indirect reduction reaction is impeded in the region with relatively high COSIG concentrations.

Fig. 13
figure 13

Variations of indirect reduction rates due to (a) COSIG and (b) COHGG with COSIG concentration at various SIG flow rates

Full size image

Figure 14 shows how the SIG flow rate affects the COSIG and COHGG indirect reaction rates in the regions with and without COSIG, so the roles of SIG and HGG are understood further. Here, the regions with COSIG concentration larger than 0.01 are thought to have COSIG (see, e.g., Figure 6), referred to as SI for convenience. The COSIG indirect reduction rate outside SI is nearly zero, according to the results given in Figure 13. The COSIG indirect reduction rate inside SI increases when the SIG flow rate increases. Conversely, the COHGG indirect reduction rates inside and outside SI become smaller with increasing SI flow rates. This result indicates that shaft gas injection impedes the HGG indirect reduction reaction in the entire region above the CZ, although this impeding effect is more significant inside SI. Generally, COHGG leads to larger indirect reduction rates than COSIG. The impeding effect does not change this situation. As a result, with increasing the SIG flow rate, the pre-reduction in terms of the COSIG indirect reduction rate intensifies initially and then slows down.

Fig. 14
figure 14

COSIG and COHGG indirect reaction rates inside and outside SI as a function of SIG flow rate. SI represents the regions with COSIG larger than 0.01

Full size image

The indirect reduction rate is also quantified in the circumferential direction over 20 sub-sectors (see Figure 15) to understand the SIG impact further. Here, the COSIG indirect reduction rate is calculated for each sub-sector. This consideration applies to the upper and lower parts (see Figure 8) of the region above the shaft injection tuyere level. The resulting results are given in Figure 16. Generally, the sector containing a shaft tuyere has the maximum COSIG indirect reduction rate, which drops rapidly as the circumferential distance away from the sector increases. Moreover, the COSIG indirect reduction rate in the upper part increases as the SIG flow rate increases. Conversely, in the lower part, the COSIG indirect reduction rate profiles are close to each other for all the SIG flow rates except for the smallest one. This feature somewhat limits the effectiveness of shaft injection at relatively high SIG flow rates.

Fig. 15
figure 15

Sub-sectors for circumferential results analysis

Full size image
Fig. 16
figure 16

Circumferential profiles of COSIG indirect reduction rate in the (a) upper and (b) lower parts of the region above the shaft injection tuyere level

Full size image

Effects of Shaft Tuyere Number and Horizontal Cavity Depth

Shaft injection could be implemented with various tuyere numbers. All the previous studies treated all shaft tuyeres as one because of the 2D models used. To date, it is unclear how the shaft tuyere number could affect the effectiveness of shaft gas injection. This effect is examined using the current 3D model at a SIG flow rate of 3 Nm3/s. The shaft tuyere number changes from 1 to 5. The tuyeres are evenly distributed in the circumferential direction. Figure 17 demonstrates that the tuyere number affects the 3D spatial distribution of COSIG concentration. When the tuyere number increases, the SIG flow rate through each tuyere decreases, which reduces the SIG penetration depth in both radial and circumferential directions, leading to smaller regions with relatively high COSIG concentrations.

Fig. 17
figure 17

Spatial distributions of COSIG concentration at various shaft tuyere numbers (STN)

Full size image

How the shaft tuyere number affects overall BF performance is shown in Figure 18. Different from the apparent differences in COSIG concentration contours, the overall BF performance has little difference at various shaft tuyere numbers. To understand this result, the COSIG indirect reduction rates in the circumferential direction are quantified for two regions above the shaft injection level. Figure 19 shows that in the upper and lower parts, the COSIG indirect reduction rates become more uniform in the circumferential direction. However, their values decrease with the increase of shaft tuyere number. These two phenomena counteract each other and explain why a larger shaft number does not improve the effect of shaft injection.

Fig. 18
figure 18

(a) Productivity, (b) hot metal temperature, (c) top gas temperature, and (d) gas utilization factor at various shaft tuyere numbers

Full size image
Fig. 19
figure 19

Circumferential profile of COSIG indirect reduction rate in the (a) upper and (b) lower parts of the region above the shaft injection level for various shaft tuyere numbers

Full size image

According to the previous CFD-DEM studies,[31,32,33,34] the cavity created by shaft injection is generally small. Therefore, the effectiveness of shaft injection may be constrained by shallow gas penetration. Via the current model, this understanding is examined, provided that some new technology allows shaft gas injection to achieve any expected penetration. In doing so, a cavity with different horizontal depths is considered in simulations. The horizontal cavity depth (r/R) varies between 0 and 25 (see Figure 20), where R is the furnace radius at the shaft injection level, and r is the depth of the horizontal cavity. Here, the cavity is assumed to have a square cross-section to facilitate mesh generation. Figure 20 shows the contours of COSIG concentration, which reveals that as the horizontal cavity depth increases, the COSIG is more distributed in the radial direction; however, the distribution becomes narrower in the circumferential direction.

Fig. 20
figure 20

Spatial distributions of COSIG concentration at various cavity depths for a SIG flow rate of 3 Nm3/s

Full size image

Figure 21 compares the overall performance indicators at different horizontal cavity depths. As the horizontal cavity depth increases from 0 to 25 pct, the hot metal temperature at the slag surface increases from 1968 K (1695 °C) to 1978 K (1705 °C) (Figure 21(b)). Note that the temperature at the slag surface is generally 50 K higher than at the taphole. As discussed above, this hot metal temperature increase means the potential for coke savings. The increased top gas temperature and gas utilization factor at a deeper cavity depth (Figures 21(c), (d)) suggest the better utilization of thermal and chemical energy, accounting for the increase of HM temperature. This is further confirmed by the circumferential profiles of the COSIG indirect reduction rate given in Figure 22. It shows that the COSIG indirect reduction rates increase with the increase of cavity depth but mainly in the sector containing the shaft tuyere and the neighboring sectors. The phenomena become more apparent in the upper part. The increased indirect reduction rate reduces coke consumption. For this, the burden charging rate drops to bring less coke into the BF to maintain the constant coke rate, decreasing productivity (Figure 21(a)). Overall, the horizontal cavity depth has a much less significant effect on BF performance than the SIG flow rate.

Fig. 21
figure 21

(a) Productivity, (b) hot metal temperature, (c) top gas temperature, and (d) gas utilization factor at various horizontal cavity depths

Full size image
Fig. 22
figure 22

Circumferential profiles of COSIG indirect reduction rate in the (a) upper and (b) lower parts of the region above the shaft injection level for various horizontal cavity depths

Full size image

Conclusions

Shaft gas injection helps reduce energy consumption and CO2 emissions of ironmaking BF. This work studies this technology with respect to three key variables related to SIG (shaft-injected gas) penetration into the particle bed: SIG flow rate, shaft tuyere number, and horizontal cavity depth. This is done using a recently developed 3D CFD process model under constant flame temperature, bosh gas volume and compositions, and coke rate. The resulting major conclusions are summarized below:

  1. (1)

    A new gas tracking method is proposed and incorporated into the 3D CFD process model. This development enables the identification of the BF regions where the SIG can reach and the quantification of the respective contributions of SIG and HGG (hearth-generated gas) to BF performance.

  2. (2)

    Increasing the SIG flow rate increases the gas penetration into the particle bed, which helps increase the solid temperature, lift the cohesive zone, and improve the indirect reduction rate by SIG in the BF shaft. However, it hinders the pe-reduction by HGG and increases the gas pressure resulting from the lengthened and thickened cohesive zone and increased gas velocities. An optimum SIG flow rate is identified with a relatively high HM temperature and a relatively low tuyere pressure, which helps achieve coke savings and maintain smooth BF operations.

  3. (3)

    Under a constant SIG flow rate, the shaft tuyere number has a minor effect on BF performance, while increasing horizontal cavity depth improves BF performance. This improvement is small compared with that achieved by increasing the SIG flow rate. This is because varying shaft tuyere number or horizontal cavity depth changes the gas penetration in the radial and circumferential directions oppositely, leading to similar SIG penetration areas where the iron ore reduction is improved.

Comprehensively, this study assesses the effectiveness of shaft gas injection, focusing on three injection variables. However, the injection operation is also affected by other variables and may be implemented with various new BF ironmaking technologies like high oxygen/hydrogen enrichment and new burden material charging. This model will help develop new BF processes to achieve massive CO2 emissions. Some works in this respect are undergoing.