Effect of degree of burden mixing on the permeability of ferrous burden layers in the cohesive zone

Abstract

The cohesive zone is the location where the pressure drop is the largest within the blast furnace. Researchers usually adopt the way of increasing the permeability of the softening-melting ore layer to achieve the purpose of improving the permeability of the cohesive zone. In this study, we proposed the concept of the degree of burden mixing (DBM). Through experiments and thermodynamic calculations found that the increase in DBM enhanced the interaction between sinter and pellet, made the burden produce a liquid phase in the cohesive zone, and deteriorated permeability. In addition, the amount of liquid phase in the cohesive zone determined the permeability of the cohesive zone. Therefore, suppressing the interaction between burdens can be considered as one of the effective ways to improve the permeability of the cohesive zone. It is worth noting that the regulation of the DBM can achieve the purpose of suppressing the interaction.

Introduction

The cohesive zone, as the location of the maximum pressure drop of the blast furnace, it can evaluate the condition of the blast furnace operation. The permeability decline of the cohesive zone will lead to the decline of the injected coal amount, which makes the direct reduction and coke rate increase. The above process will deteriorate the economy of the blast furnace ironmaking. Therefore, the improvement of cohesive zone permeability plays an important role in the blast furnace ironmaking.

To improve the cohesive zone permeability, researchers had carried out most studies. Sunahara et al. confirmed that both low position and thin cohesive zone could improve the permeability of the blast furnace [1–4]. Most scholars support that the mixing of coke and ferrous burdens and the addition of coke to the centre of the blast furnace would increase the permeability of the cohesive zone [5,6], but the existing studies fail to resolve the contradiction between the permeability of the cohesive zone and the coke ratio. Because the low position and thin cohesive zone formation require a high reduction degree, it is feasible to adopt low locations and thin cohesive zones in ironmaking production under conditions of special processes, such as oxygen blast furnaces [7–11], hot reduction gas injections [12–14], and utilization metallization burdens [15–19], but special processes remain in the stage of industrial experimentation. The softening and melting of iron ore reduces the porosity of the ferrous burdens layer, so the researchers regard increasing the permeability of iron ore as the main way to improve the permeability of the cohesive zone at a constant coke rate condition. The research results of Wu et al. [20,21] showed that the interaction would increase the liquid phase, which deteriorated the improvement of the blast furnace permeability. The study results of slag behaviour on the coke bed confirmed that the high melting point of solid slag existed in the coke bed and the slag drip has nothing to do with the liquid phase ratio of the slag [22]. This result provides a basis for reducing the thickness of the cohesive zone, but most research until now has been one-sided in investigating the effect of the burden on the permeability of the cohesive zone in the cold state [23]. One major drawback of this research is that it failed to consider the effects of the interaction. We carried out this study to investigate the rational distribution of ferrous burdens and improve the permeability of the ferrous burdens during softening and melting.

Thermodynamic calculations and experimental

Raw material preparations

We used acid oxidized pellets and basic sinter as raw materials in this experiment. Table 1 presents the compositions of raw materials and the corresponding diameters were 10–12.5 mm. To prevent changes in slag during the experiment, we conducted pre-reduction experiments on sinter and pellet respectively, and made the reduction degree reach 70%. Figure 1(a) presents the reduction furnace. The pellet or sinter was placed in a stainless steel reaction tube (the size of the reduction reaction tube is Φ74 mm × 800 mm) and heated to 900 °C under the nitrogen atmosphere. The furnace-heating rate and the nitrogen gas flow were 5 °C min⁻¹ and 5 L min⁻¹. While the gas inlet introduced reduction gas into the reaction tube. When the temperature of the reaction tube reached 900 °C, the reaction tube started to introduce reduction gas (CO:N₂ = 30:70) from the gas inlet and the pellet or sinter began to lose weight gradually as the reduction proceeded. During the reduction process, the balance recorded the weight loss of the pellet or sinter. When the weight loss of the pellet or sinter measured up the target reduction degree, the reaction tube stopped heating and introduced nitrogen (nitrogen flow rate: 5 L min⁻¹) to cool the pellet or sinter to room temperature. Equation (1) lists the relationship between the reduction degree and the weight loss of the pellet or sinter: (1) where R is the reduction degree of reduction t time, M ₀ is the initial weight of the burden, M ₁ is the burdened weight at the reduction beginning, M_t is the burdened weight at time t, W ₁ is the initial FeO content of pellet, W ₂ is the initial total iron content of the burden, 0.11 is the conversion factor of oxygen required for the oxidation of FeO to Fe₂O₃, 0.43 is the conversion factor of oxygen required when the entire iron amount is oxidized to Fe₂O₃.

Figure 1.

(a) Reduction furnace. (b) Softening and melting furnace.

Table 1.

Main components of sinter and pellet used in experiments (before reduction, mass %).

Burden	TFe	FeO	CaO	SiO₂	Al₂O₃	MgO	R
Sinter	54.7	11.2	11.9	5.08	2.71	1.76	2.34
Pellet	61.4	0.87	1.16	6.63	1.27	1.97	0.17

Softening, melting, and dropping experiments of pre-reduced burden

Figure 1(b) presents the schematic diagram of softening and melting furnace. The softening and melting furnace carried out softening, melting, and dripping experiments of pre-reduced burden. To simulate the charging process of the actual blast furnace, the graphite reaction tube (inner diameter: 48 mm) placed the pre-reduction burden and put the 20 g coke layer on the top and bottom of the burden. The total height of the burden layer reached 65 ± 5 mm. In the experiments, to prevent the reducing gas from causing the change in the slag amount, the furnace injected 12 L min⁻¹ nitrogen for protection. When we observed the first drop of burden in the peephole, and the furnace immediately stopped heating. Then we cooled down the burden to room temperature under a nitrogen atmosphere (nitrogen flow rate: 5 L min ⁻¹). The experiment used the graphite crucible to collect the droplets of the burden. The burden layer was subjected to a pressure of 1 kg cm⁻², while displacement metre and U-differential pressure gauge recorded the burden shrinkage and pressure differences. Table 2 presents the index and physical significance of softening, melting, and dripping off the pre-reduction burden.

Table 2.

Index and physical significance of softening, melting and dripping of pre-reduction mixed burden.

Parameters	Physical significance
T_s /°C	Softening start temperature (°C), burden lays contraction of 10%
T_m /°C	Melting start temperature (°C), the steep pressure drop of a burden lies
ΔT_s /°C	Softening zone interval (T_m –T_S ) (°C)
T_d /°C	Dripping temperature (°C), the temperature at which dripping starts.
ΔT_m /°C	Melting zone interval (T_d –T_m ) (°C)
ΔT_c /°C	Cohesive zone interval (T_d – T_s ) (°C)
ΔP _max/Pa	Maximum differential pressure (Pa)
S/kPa °C	Permeability index (kPa °C), namely
	ΔP_T signifies differential pressure at T

In this study, we propose the concept of the degree of burden mixing (DBM) and Equation (2) expressed the DBM. Experiment set the values of DBM to 0%, 25%, 50%, 75%, and 100%. Table 3 and Figure 2 present the compositions of the sinter, pellet, and the mixed burden and the schematic diagram of burden distribution under different DBM. It is worth noting that the experiment randomly mixed the burdens and failed to achieve the distribution in Figure 2. (2) where DBM is the degree of burden mixing, S_m is the mass (g) of the sinter mixed with the pellet, and S is the mass of the sinter.

Figure 2.

Schematic diagram of burden distribution under different DBM.

Table 3.

Main chemical components of pre-reduced sinter, pellet, and mixed burden.

DBM	Burdens	Rate	TFe	CaO	SiO₂	Al₂O₃	MgO	R
–	Sinter	51%	64.35	14.03	5.97	3.19	2.07	2.34
–	Pellet	49%	69.89	1.32	7.54	1.44	2.24	0.17
0	Not mixed		69.89	1.32	7.54	1.44	2.24	0.17
25%		Chemical composition of the mixed burden	68.75	3.93	7.22	1.8	2.20	0.54
50%	Randomly mixed		67.99	5.66	7.00	2.04	2.18	0.8
75%			67.46	6.87	6.85	2.21	2.16	1
100%	Completely mixed		67.06	7.79	6.75	2.33	2.15	1.15

Thermodynamic calculations

Based on the compositions in Table 3, using Factsage7.0 thermodynamic software calculated the liquid phase rate and liquid phase weight of the gangue phase (CaO + SiO₂+ Al₂O₃+ MgO). In thermodynamic calculations, we assumed that the gangue phase composition of the mixed burden forms a uniform slag and discussed the rationality of the assumption. Through equations (3), (4), and (5) conducted the calculations: (3) (4) (5) where is the weight of the gangue phase of the mixed burden, and are weights of pellet and sinter in the mixed burdens respectively, is the liquid rate of the gangue of the mixed burden, is the liquid phase weight of the gangue of the mixed burden, as obtained through thermodynamic calculations, is the weight of the liquid phase before the unmixed sinter and mixed burdens interaction.

We used the oxides database of Factsage7.0 to calculate the liquid phase rate and liquid phase weight of the gangue phase and used the calculation result to explain the experimental phenomenon. The temperature of thermodynamic calculations ranged from 1000 °C to 1450 °C and included the solidus temperature of each slag phase, while the temperature step was 50 °C.

Results and discussion

Figure 3 shows the temperature change result of the ferrous burden layers. It illustrates that as DBM increased, the softening start temperature of the ferrous burden lays gradually decreased and the melting start temperature, and the dropping temperature followed inconsistent regularity. Figure 4 shows the maximum differential pressure (ΔP _max/Pa) and the permeability index (S/KPa °C). Figure 4 observed that the permeability index (S/kPa °C) increased as the value of DBM increased. This result demonstrates that the distribution disorder of the burden deteriorated the permeability of the ferrous burden layers.

Figure 3.

Variation of the cohesive zone with DBM.

Figure 4.

Variation of maximum differential pressure (ΔP _max/Pa) and permeability index (S/KPa·°C) with DBM.

Figure 5 presents the pressure drop and shrinkage at different temperatures. Referring to Figures 3 and 5, the melting zone caused the pressure drop of the cohesive zone to rise. This result is consistent with the results of the softening and melting behaviour of the burden under oxygen blast furnace conditions [10]. Therefore, the melting zone permeability determines the cohesive zone permeability. However, Figures 3-(3) and 5 observed that the temperature interval of the melting zone did not determine the melting zone permeability because the change of gas permeability follows inconsistent with the temperature range of the melting zone. Based on Figures 4–6 calculated the geometry thicknesses (DBM = 0%: 32.17 mm; DBM = 25%: 32.5 mm; DBM = 50%: 31.2 mm; DBM = 75%: 29.12 mm; DBM = 100%: 31.2 mm) of the melting zone intervals. Varieties of permeability follow inconsistent with the change in the geometric thickness of the melting zone. This result also demonstrates that the geometric thickness of the melting zone does not determine the gas permeability. Table 4 listed the main components of different slag. In Table 4, N1 and N2 represent the upper residual slag and the lower residual slag at DBM = 0%. The varieties of CaO and SiO₂ content made the difference between N1 and N2. N3 and N4, N5 and N6, N7 and N8 respectively represent when DBM reaches 25%, 50%, and 75%, the main components of residual and dripping slag. N9 represents the main component of the residual slag at DBM = 100%. The difference in composition between residual and dripping slag also reflected in CaO and SiO₂ contents. Combined with the analysis of Figures 4-(1) and 5-(1), the influence of burden shrinkage on pressure drops appears insignificant.

Figure 5.

Pressure difference and shrinkage of the layers under different temperatures.

Table 4.

Main components of different slags.

N	Status	DBM	FeO	CaO	SiO₂	Al₂O₃	MgO	R₂
1	Residual	0	7.6	7.34	34.9	9.64	9.80	0.21
2	Residual	0	9.36	30.1	22.2	7.75	9.31	1.35
3	Residual	25	6.68	32.11	31.2	9.44	9.12	1.02
4	Dripping	25	5.22	19.7	32.5	10.2	8.29	0.6
5	Residual	50	4.96	28.9	31.9	9.35	9.92	0.91
6	Dripping	50	13.9	10.2	39.4	11.9	3.08	0.25
7	Residual	75	6.97	33.4	32.7	10.4	7.86	1.02
8	Dripping	75	7.50	19.8	41.9	11.1	5.90	0.47
9	Residual	100	6.66	29.5	28.4	9.46	8.55	1.03

Repeat the DBM = 0% and DBM = 100% experiment, when the pressure difference increased sharply, we immediately stop the experiment and observed the morphology of the burden layers, as showed in Figure 6. It presents that when the pressure difference of DBM = 0% steeply increased, the shrinkage rate of the pellet exceeded the sinter shrinkage rate. In DBM = 0%, the shrinkage of the pellet exceeded the shrinkage in DBM = 100%, but we observed a broken sinter at DBM = 100%. When the pressure difference of the layer steeply increased at DBM = 0% and DBM = 100% and the temperature was 1305 and 1306 °C respectively. In addition, the shrinkage values of the corresponding layers were 50.5% and 52%.

Figure 6.

Macroscopic morphology of the layer, when the pressure difference increased sharply at DBM = 0% and DBM = 100%.

Figure 7 shows the residual slag and iron macroscopic morphology. The upper part of the residue (pellet layer in the experiment) and the lower part (layer of sinter in the experiment) exhibited different morphologies in DBM = 0%. The upper slag remained in a molten state and the lower slag maintained the morphology of the sinter. The macroscopic morphology of the residual slag is shown between the pellet and the sintered morphology in DBM = 25%. If DBM ≥ 50%, the residual slag is presented in the molten state.

Figure 7.

Macroscopic morphology of residual slag and iron.

Figure 8 presents the microscopic morphology of the residual slag in DBM = 0%. Two different regions existed in the figure: the light-coloured region and the darker regions. The slag phase of the light-coloured region mainly consisted of uniform CaO-SiO₂-MgO and CaO-SiO₂-Al₂O₃ phases, but the metal iron particles were distributed within the slag below the size of 100 µm. The slag phase of the darker regions was Ca₂SiO₄. The size of metallic iron particles distributed in the slag phase exceeded 200 µm. Figure 9 shows the microscopic morphology of the residual slag while DBM ≥ 25%. When the DBM reached 25%, the main phases of the slag phase were CaO-SiO₂-Al₂O₃ and CaO-SiO₂-MgO phases, but the slag phase would distribute the high amounts of Ca₂SiO₄ and metallic iron particles. When DBM = 50%, 75% and 100%, CaO-SiO₂-Al₂O_3, and CaO-SiO₂-MgO phases formed the residual slag phase. In addition, compared with DBM = 25%, the Ca₂SiO₄ phase amount in the slag was reduced.

Figure 8.

The microstructure of the longitudinal section at DBM = 0%.

Figure 9.

Microscopic morphology of residual slag with DBM ≥ 25%.

Results of thermodynamic calculations

Figures 10 and 11 present the results of the thermodynamic calculations. Figure 10 represents the liquid phase ratio of (CaO + SiO₂+ MgO + Al₂O₃) component of the mixed burden. Figure 11 observes that the mixed portion gangue phase solidus temperature at DBM = 0% behaved as the lowest (as well as the solidus temperature of the pellet gangue phase). As DBM grew, the solidus temperature of the gangue phase gradually increased and the solidus temperature of the sinter reached the highest value. Although the liquid phase ratio of the gangue phase reached high in certain cases of DBM because the ratios of the gangue phases differed, we must calculate the true liquid phase weight. The true liquid phase weight was difficult to calculate due to the interaction of the burden. Consequently, Figure 11 presents the liquid phases (CaO + SiO₂+ MgO + Al₂O₃) sum of the mixed burden and the unmixed sinter. Figure 11 demonstrates that the higher degree of the burdens mixing, the more amount of the liquid slag at the same temperature.

Figure 10.

Value of at different temperatures.

Figure 11.

Value of at different temperatures.

Mechanism of influence of DBM on the permeability of burden layers

The formation of blast furnace slag mainly depends on the interaction between different burdens, but the interaction caused more liquid phases to generate at lower temperatures. Figure 11 reveals this research result and its result keeps consistent with the results of the References [21,22], so this research result demonstrates the correctness of the thermodynamic calculation hypothesis.

Combined with thermodynamic calculations and experiments, we can interpret the reason why the increase in DBM value deteriorated the permeability of the burdens layer. The reasons as the following: (1) Figures 6 and 11 can demonstrate that the pellets produced the initial liquid phase because the slag phase first formed the olivine phase (Fe₂SiO₄) at a lower melting point; (2) Owing to the mixing of the burden, the initial liquid phase of pellet formation would undergo a short distance migration. When the initial liquid phase migrated to the sinter, it interacted with the sinter to form a new phase and the melting point of this phase was between the gangue of the pellet and the sinter. As the temperature further raised, the formation of a new phase melted and increased the amount of liquid phase; (3) Figure 12 shows that the liquid phase migrated to unmixed sinter and occurred interaction. The increase in the liquid phase before dripping played a role in deteriorating permeability.

Figure 12.

Schematic diagram of burden layer structure changes under different DBM.

According to Ref. [22], the high melting point substance in a solid-state could allow retaining in the coke layer, which included a high melting point slag and a molten state iron fixed by it. Since the liquid phase determined the transportation and mixing of substances, the liquid phase controlled the process of interaction reaction [24]. In the case of DBM < 100%, the liquid phase was mainly generated in the mixed burden region and moved downwards under the upper portion pressure. At DBM = 0% and 100%, the pellet produced the initial liquid phase. However, in the case of DBM = 100%, since the pellet and the sinter were completely mixed, the interaction occurred within the entire layer, while the distance of the liquid phase migration was shorter than the distance of DBM = 0%. In DBM = 0%, the liquid phase was mainly produced in the pellet layer, and the interaction proceeded from top to bottom. Figure 8 could observe that the interaction proceeded incompletely until the liquid phase amount dropped. At the same temperature, DBM = 25%, 50%, and 75%, the mixed burden component could produce a more liquid phase compared with DBM = 100%. Moreover, compared with DBM = 0%, the distance of the interaction was shortened and the liquid phase increased, which more benefited to occur the interaction, produce the additional liquid phase, and separated into slag and iron. However, the liquid phase was unfavourable to cohesive zone permeability, which also caused the gas permeability of the burden layer to decrease as the DBM increased. In the case of DBM = 100%, the liquid phase at high temperatures exceeded the liquid phase temperature of DBM < 100% (1383°C). As shown in Figure 13, in Ref. [23], the balls injected with 3 ml of liquid represent pellets, and the balls injected with 0.1 ml of liquid represent sinter. As shown in Figure 13(a) and (b),the experimental results suggest that complete mixing of the burden is more favorable for gas permeability than complete separation. However, this result is contrary to the results of this study. The fundamental reason is that the former can only simulate the liquid phase generation of a single burden, ignoring the interaction to generate more liquid phases in advance. Complete mixing the burden facilitated the interaction to occur and produced the liquid phase.

Figure 13.

Estimated mechanism of the effect of sponge ball arrangement [23]. Note: Liquid 0.3 ml/ball – pellet and Liquid 0.1 ml/ball – sinter.

Conclusion

By the thermodynamic calculations, taking different DBM values as variables made the softening, melting and dropping experiments to obtain the following conclusions:

The increase in DBM deteriorated the permeability of the burdens layer.

The liquid phase ratio before dripping determined the permeability of the ferrous burden layers in the cohesive zone.

In this study, increasing the DBM value could shorten the distance of the interaction, and reducing the DBM value could suppress the interaction. Therefore, under a certain coke ratio, controlling the DBM value within an appropriate range can effectively improve the permeability of the ferrous burden layers in the cohesive zone. This result could directly apply to blast furnace ironmaking.

Footnotes

Disclosure statement

No potential conflict of interest was reported by the author(s).

References

Zhao

Fan

Chen

et al. Metallurgical properties of ferrous burdens in blast furnace. J Cent South Univ (Sci Technol). 2010;41(6):2053.

Sunahara

Natsui

Shizawa

et al. Effect of coke reactivity on sinter softening-melting property by simultaneous evaluation method of carbonaceous and ferrous burdens in blast furnace. ISIJ Int. 2011;51(8):1322.

Ueda

Miki

Murakami

et al. Agenda for low reducing agent operation of blast furnace-reduction and melting phenomena of iron ore. Tetsu-to-Hagané. 2013;99(1):1.

Takeuchi

Iwami

Higuchi

et al. Evaluation of sinter quality for improvement in gas permeability of blast furnace. ISIJ Int. 2013;54(4):791.

Zhou

Liu

et al. Effect of mixed ore/coke burden on permeability in soft melting zone. J Wuhan Iron Steel Univ. 1990;13(2):117.

Zhao

Chen

et al. Mechanism of burden distribution in central coke charging process and exploration on charging pattern in blast furnace. Iron Steel. 2016;51(6):10.

Meijer

Denys

Lasar

et al. ULCOS: ultra-low CO2 steelmaking. Ironmak Steelmak. 2009;3(4):249.

Zuo

Hirsch

The trial of the top gas recycling blast furnace at LKAB s’ EBF and scale-up. Rev Metall. 2009;106(9):387.

Danloy

Berthelemot

Grant

et al. ULCOS - pilot testing of the low-CO₂ blast furnace process at the experimental BF in Luleå. Rev Métall. 2009;106(1):1.

10.

Yan

Gao

et al. Study on industrial test of the oxygen blast furnace. Kang T'ieh/Iron Steel (Chin). 2011;46(3):6.

11.

Wang

Lan

et al. Softening and melting behavior of mixed burden for oxygen blast furnace. J Iron Steel Res. 2013;20(5):11.

12.

Tseitlin

Lazutkin

Styopin

GM.

A flow-chart for iron making on the basis of 100% usage of process oxygen and hot reducing gases injection. ISIJ Int. 1994;34(7):570.

13.

Babich

Gudenau

Formoso

et al. Choice of technological regimes of a blast furnace operation with injection of hot reducing gases. Rev Metall. 2002;38:288.

14.

Desai

Ramna

Sathaye

JM.

Effect of hot reducing gas (HRG) injection on blast furnace operational parameters: theoretical investigation. Ironmak Steelmak. 2008;35(1):43.

15.

Kaushik

Fruehan

RJ.

Mixed burden softening and melting phenomena in blast furnace operation part 1-X-ray observation of ferrous burden. Ironmak Steelmak. 2006;33(6):507.

16.

Kaushik

Fruehan

RJ.

Mixed burden softening and melting phenomena in blast furnace operation Part 2-Mechanism of softening and melting and impact on cohesive zone. Ironmak Steelmak. 2006;33(6):520.

17.

Kaushik

Fruehan

RJ.

Mixed burden softening and melting phenomena in blast furnace operation Part3-Mechanism of burden interaction and melt exudation phenomenon. Ironmak Steelmak. 2007;34(1):10.

18.

She

Wang

Liu

et al. Increasing the mixing rate of metalized pellets in blast furnace based on the high-temperature interactivity of iron bearing materials. ISIJ Int. 2014;54(12):2728.

19.

Chung

Kim

Sohn

DRI from recycled iron bearing wastes for lower carbon in the blast furnace. ISIJ Int. 2015;55:1157.

20.

Tian

YQ.

Evaluation of lump ores for use in modern blast furnaces as part of mixed burden practice. Ironmak Steelmak. 2009;36(1):19.

21.

Han

et al. Increasing lump ores proportion in blast furnace based on the high-temperature interactivity of iron bearing materials. ISIJ Int. 2010;50(5):686.

22.

Ueda

Kon

Miki

et al. Softening, melting, and permeation phenomena of CaO-FeO-SiO₂ oxide on a coke bed. ISIJ Int. 2015;55(10):2098.

23.

Ishii

Murai R

Sumi

et al. Gas permeability in cohesive zone in the ironmaking blast furnace. ISIJ Int. 2017;57(9): 1531.

24.

Jiang

Sun

Liu

et al. Mechanism of sodium sulfate in promoting selective reduction of nickel laterite ore during reduction roasting process. Int J Miner Process. 2013;123:32.