Effects of Iron Ores on the Combustion Behavior of Coke and NOx Emission during Sintering Process

Abstract

Calcium ferrite could promote the CO–NO reduction reaction, and its formation is affected by iron ores during sintering process. In present study, effects of iron ore coating layers on coke combustion rate and NOx emission were investigated in a visualize combustion equipment, and an optimized ore blending structure was proposed by sinter pot test. Due to the melting of iron ore coating layers at high temperature, coke transformed from coated to naked. With increasing of the binary basicity of the iron ore coating layers, the formation of calcium ferrite increased, resulting in increasing of the melt fluidity. The lower the formation temperature of the melts, the sooner coke was exposed, and the peak combustion rate linearly increased with the melt fluidity of iron ore coating layers. Meanwhile, compared to the high-silicon ores, the maximum NOx emission concentration and conversion rate of N element were lower with the low-silicon ores. NOx emission concentration showed an inverted W-shape trend and had an 8-shape relation with coke combustion rate. Due to the difference of the capability of calcium ferrite formation in coating layers, the conversion rate of N element was linearly negative and positive correlated with the basicity of iron ore coating layers and mass of CO emission, respectively. In addition, with the proportions of the low-silicon limonite and hematite increased in sinter mixture, NOx emission gradually decreased. As a consequence, with exclusively using low-silicon lignite and hematite in sinter mixture, NOx emission decreased by about 20% and sinter indexes significantly improved.

1. Introduction

In recent years, with the rapidly increase of the crude steel production in China, NOx emissions of the iron and steel industry increased gradually. NOx was one of the main culprits of the formation of photochemical smog, acid rain and haze, which serious harm to human health. The iron and steel industry, as the major sources of NOx emission, accounts for about 6% of the total NOx emissions of industry in China,^1,2) while the iron ore sintering process occupied approximately 50% of the total NOx emissions in the iron and steel industry.^3,4) Therefore, it is extremely urgent to reduce NOx emissions from the iron ore sintering process under the stricter environmental protection policy.

NOx was mainly generated by the oxidation reaction of nitrogen-containing organic (Fuel-N) in the solid fuel during iron ore sintering process, which was called fuel NOx. NO was the majority of the fuel NOx, accounting for about 95%.^5,6) CO and NO formed simultaneously during the solid fuel combustion process, and Hida et al.^7,8,9) found that the volume concentration ratio of CO and NO was about 150 in the flue gas near the surface of coke. As a result, NO was reduced to N₂ by CO at the boundary layer of coke. However, Zhou¹⁰⁾ simulated the formation and reduction of NOx in sintering process. It showed that about 10% of the generated NOx was reduced by CO. Moreover, Wu and Kasai^11,12) found that the reaction between CO and NO was promoted by calcium ferrite minerals. In order to promote the formation of calcium ferrite minerals during sintering, Pan et al.^13,14,15) reported that NOx emissions decreased by 15% with the increase of the basicity of sinter from 1.9 to 2.4. Morioka¹⁶⁾ proposed that the basicity of adhering fines increased and the formation of calcium ferrite minerals promoted by decreasing the size of limestone to below 0.25 mm, resulting in reduction of NOx emission about 25%. Meanwhile, Katayama¹⁷⁾ studied the effects of the CaO content in the adhering fines on NOx emission. The NOx emission reached the minimum value at the 20% of CaO content. Gan et al.^18,19,20) investigated the influences of the mass fraction and granulating methods of calcium ferrite minerals on the NOx reduction, and they proposed to form a double-layered granular structure and keep 30% diversion ratios for pre-granulation. In addition, Kasai et al.^12,17,21) found the adhering fines on the surface of coke melted at the high temperature, and then the coating coke transformed to naked coke. The high-temperature melting behavior of the coating layer affected the combustion behavior of coke. Ohno^22,23,24) and Zhou^25,26) established the coke combustion rate equation and investigated the effect of the formation of liquid phase on the coke combustion rate.

Although above studies have been carried out to reduce NOx emission in iron ore sintering process, they focused on promoting the formation and distribution of calcium ferrite minerals by regulating the fluxes. However, iron ore was an important factor on the formation and distribution of calcium ferrite minerals in sintered bed, which had different formation capabilities of calcium ferrite minerals. There were few studies about NOx reduction by regulating iron ores. Hence, in this paper, a visualize combustion experimental was used to investigate the effects of iron ores on the coke combustion behavior and NOx emission with a camera and flue gas analyzer, and different ore blending schemes were optimized to reduce NOx. Then, the ore blending structure was optimized to reduce the NOx emission with ensuring the quality index of sinter.

2. Materials and Methods

2.1. Raw Materials

Raw materials were from a sintering plant in China, which contained iron materials (five kinds of iron ores and return fines), fluxes (quicklime, limestone, dolomite and serpentine) and coke. The chemical components of raw materials are shown in Table 1. Ore-A and ore-B were limonite ores from Australia, ore-C was Marra Mamba ore, and ore-D and ore-E were hematite ores from Brazilian. For limonite ores, the SiO₂ content of ore-A was higher than that of ore-B. For hematite ores, the SiO₂ contents of ore-D and ore-E were 5.81% and 2.31%, respectively. Moreover, Table 2 shows the size distribution of raw materials in sinter mixture. The size of ore-A was slightly smaller than that of re-B while the mass fraction of below 0.5 mm of ore-D was obviously more than that of ore-E. The proximate and ultimate analyses of coke are listed in Table 3. The mass fractions of fixed carbon, volatile and N element were 82.19%, 1.25%, and 0.97%, respectively. The net calorific value was 29313 kJ/kg. The ignition temperature was 563°C.

Table 1. Chemical composition of raw materials.

Raw materials	TFe	FeO	SiO2	CaO	Al2O3	MgO	LOI
Ore-A	57.55	0.23	5.47	0.03	1.30	0.04	10.71
Ore-B	58.22	0.21	4.34	0.04	1.56	0.10	10.40
Ore-C	61.43	0.37	3.43	0.03	2.36	0.07	6.67
Ore-D	63.38	1.01	5.81	0.02	0.88	0.04	2.01
Ore-E	65.34	0.18	2.31	0.01	1.14	0.04	1.95
Quicklime	0.00	0.00	3.59	86.59	0.00	0.00	7.10
Limestone	0.00	0.00	0.55	54.93	0.26	0.44	43.53
Dolomite	0.00	0.00	1.43	31.13	0.34	20.23	45.60
Serpentine	0.00	0.00	38.44	2.46	0.94	37.53	13.57
Return fine	57.27	9.07	5.04	9.56	1.80	1.64	0.00
Coke	0.49	0.00	5.96	0.42	3.88	0.11	87.93

Table 2. Size composition of raw materials in sinter mixture/mass%.

Raw materials	>10 mm	8–10 mm	5–8 mm	3–5 mm	2–3 mm	1–2 mm	0.5–1 mm	0.25–0.5 mm	<0.25 mm
Ore-A	8.40	6.21	12.80	15.87	13.55	18.42	13.30	6.31	5.14
Ore-B	7.02	6.68	13.12	13.85	11.40	15.53	13.72	8.79	9.89
Ore-C	2.12	3.40	14.54	15.71	11.05	14.19	14.11	11.37	13.51
Ore-D	3.04	2.62	8.53	9.53	7.10	8.24	6.75	9.39	44.80
Ore-E	5.80	4.24	10.74	13.44	11.76	10.91	11.41	10.73	20.97
Quicklime	0.00	0.00	0.00	0.14	1.33	4.15	6.00	12.28	76.11
Limestone	0.00	0.00	0.13	8.99	11.89	19.67	19.89	8.53	30.88
Dolomite	0.00	0.00	0.00	3.11	13.96	21.83	18.77	11.13	31.21
Serpentine	0.00	0.00	0.00	5.50	14.50	28.45	20.45	12.10	19.00
Return fine	0.49	0.97	20.54	47.61	11.83	9.10	4.79	2.25	2.45
Coke	0.00	0.00	2.60	9.70	8.40	16.60	31.40	15.20	16.10

Table 3. Proximate and ultimate analyses of coke.

Proximate analysis/mass%				Ultimate analysis/mass%			Net calorific value (kJ/kg)	Ignition temperature (°C)
Fixed carbon	Volatile	Ash	Moisture	C	H	N
82.19	1.25	12.63	3.93	85.55	0.11	0.97	29313	563

2.2. Methods

2.2.1. Combustion Test

Hida²⁷⁾ found that there were S, C and P patterns of coke existence state in the sintered bed. The percentage of S-type coke was about 70%, which was described as single coarse coke covered with adhering fines. Hence, the S-type coke was selected the research object in this study. The S-type coke was prepared by coarse coke as core particle and mixtures of iron ores and Ca(OH)₂ as adhering fines. In order to observe the combustion behavior of S-type coke, the size of coarse coke was 6.3–8.0 mm. The mass ratio of iron ore and Ca(OH)₂ reagent in mixtures was 7:3. The chemical components and binary basicity of mixtures are shown in Table 4. It shows that the binary basicity of the ore-E coating layer was highest while that of the ore-D coating layer was lowest. The binary basicity of the ore-B coating layer was higher than that of the ore-A coating layer. In addition, since Al₂O₃ has no effect on the formation of CO and NO during the coke combustion,^12,22) Al₂O₃ powder was used as an alternative material of mixture, which was as a blank control group. The coarse cokes were coated with adhering fines to be 12 mm quasi-particles by a laboratory pelletizer with a little amount of water. The weight of coke and adhering fines was 0.2 g and 0.6 g, respectively.

Table 4. Chemical composition and binary bascity of the coating layers.

Iron ores	Chemical composition/mass%							Basicity -/-
	TFe	FeO	SiO2	CaO	Al2O3	MgO	LOI
Ore-A	40.29	0.16	3.83	22.72	0.91	0.03	14.79	5.93
Ore-B	40.75	0.15	3.04	22.73	1.09	0.07	14.58	7.48
Ore-C	43.00	0.26	2.40	22.72	1.65	0.05	11.97	9.46
Ore-D	44.37	0.71	4.07	22.72	0.62	0.03	8.70	5.59
Ore-E	45.74	0.13	1.62	22.71	0.80	0.03	8.66	14.04

The combustion test of coke was carried by a horizontal furnace. The schematic diagram of the combustion equipment was shown in Fig. 1. The furnace was programmed from room temperature to 1280°C at the heating rate of 10°C/min. The quasi-particle was put into the quartz tube with Φ20×27 mm and placed on some 2.0 mm of alumina balls. Then the quartz tube was moved into the of horizontal furnace at a speed of 7 cm/min by a transmission device for 7 min, remained for 5 min at 1280°C, and then moved out at a speed of 9.8 cm/min. The temperature of the quasi-particle was evidenced in Fig. 2 by thermoscouple-2. The thermocouple-2 was closed to the surface of the quartz tube and about 5 mm away from the surface of quasi-particle. The thermocouple-2 and quasi-particle always moved at the same time, and the temperature of quasi-particle was reflected in real-time. During experiments, the air flow was 2 L/min, the compositions of O₂, CO₂, CO and NO in off gas were continuously analyzed by flue gas analyzer (MRU OPTIMA 7, Germany), and the macro-morphology of the quasi-particle was also continuously photographed by the camera. Since 95% of NOx emission was NO during sintering process, only NO gas was used in this paper.

Fig. 1.

Schematic diagram of the visualize combustion equipment. (Online version in color.)

Fig. 2.

Temperature and atmosphere of the combustion tests.

The fluidity of the melt directly determines the bonding strength of sinter during the sintering process, which is generally used to characterize the bonding extent. Wu²⁸⁾ used the projection area method to measure the fluidity of the melt which was the ratio of the difference projection areas of the adhering layers before and after sintering to the original projection area of the adhering layers. In the present study, the projection area of the coating layer was difficulty to measure. However, when the coating layer melts at high temperature, the melts will bond the alumina balls around the quasi-particle. Okazaki²⁹⁾ et al. also found that the melts bonded and reacted with the un-melt particle. The fluidity of the melt can also be characterized by measuring the mass of the coated layer and the bonded samples before and after combusting. The schematic diagram of the melt fluidity of the coated layer during the coke combustion process was shown in Fig. 3. The masses of the coating layer and coke in quasi-particle were 0.6 g and 0.2 g, respectively. During the combustion process of coke, the coating layer was melting and the quasi-particle transformed from the coated type to the exposed type. At the end of combusting, coke burnt out and the melts filled in the gaps and on the surface of the alumina balls. After cooling, the melts and the bonded alumina balls formed a bonded sample. The mass of the bonded samples were weighted and defined as m_after. Hence, referencing to the projection area method,²⁸⁾ the melt fluidity of coating layer was the ratio of the difference mass of the bonded samples after combusting and the coated layer before combusting to the original mass of the coating layer in quasi-particles, and it was calculated by the Eq. (1). The quantities and viscosities of melts were different with the different iron ores coated layers during the coke combustion process, resulting in different amounts of bonded alumina balls.   

$$MF=\frac{m_{after}-m_{before}}{m_{before}}$$(1)

where, MF is the melt fluidity of the coating layer, m_before (g) is the original mass of the coating layer before combustion test, m_after (g) is the mass of the bonded samples after combustion test.

Fig. 3.

Schematic diagram of the melt fluidity of the coating layer during the coke combustion process. (Online version in color.)

The evaluation indexes of fuel-N’s transformation were mainly the NOx concentration, the total of NOx emission and the conversion rate of N element in coke. The conversion rate of N element η_N was calculated by the ratio of the mass of N element in coke convert to NOx and the mass of N element in coke. The total of NOx emission and conversion rate of N element in coke was calculated by the following Eqs. (2) and (3). In addition, the combustion behavior of coke was described by the combustion rate which was calculated by Eq. (4).   

$${\text{m}}_{\text{NO}}\text{=}{\int }_{\text{0}}^{t_{\text{end}}}\frac{F_{\text{g}}{C}_{\text{t}}^{\text{NO}}}{\text{60}}\times {\text{10}}^{-\text{3}}\text{dt}$$(2)

  

$${\eta }_{\text{N}}=\frac{{\text{m}}_{\text{NO}}{M}_{\text{N}}}{m_{\text{coke}}{\omega }_{\text{N,coke}}{M}_{\text{NO}}}\times \text{100\%}$$(3)

  

$${\upsilon }_{\text{t}}=\frac{F_{\text{g}}\left({\text{C}}_{\text{t}}^{\text{CO}}+{\text{C}}_{\text{t}}^{\text{CO}}\right)M_{\text{C}}}{60m_{\text{coke}}{\omega }_{\text{C,coke}}{\text{V}}_{\text{mol}}}$$(4)

where, η_N (%) is the conversion rate of N element in coke, m_NO (mg) is the mass of NO emission during combustion test, υ_t (s⁻¹) is the mass ratio of C element oxidized to CO₂ and CO in coke at t second during coke combustion process, $C_{\text{t}}^{\text{NO}}$ (mg/m³) is the volume concentration of NO at t second, $C_{\text{t}}^{C{O}_2}$ (vol%) and $C_{\text{t}}^{\text{CO}}$ (vol%) is the volume concentration of CO₂ and CO at t second, respectively. F_g (L/min) is the air flow of combustion test, t_end (s) is the combustion end time, m_coke (mg) is the mass of coke in sample. ω_C,coke (mass%) and ω_N,coke (mass%) is the mass fraction of C and N element in coke, respectively; M_C (g/mol), M_N (g/mol) and M_NO (g/mol) are the molar mass of C, N and NO, respectively. V_mol (mol/L) is the standard molar volume.

2.2.2. Micro-sintering Experiment

In order to clarify the effects of ore blending on NOx emission and strength of sinter, some blending schemes were designed. The proportions of raw materials in sinter mixture at different schemes were shown in Table 5. In Base scheme, the mass ratio of ore-A and ore-B was 1:1, so was ore-D and ore-E. The 50% and 100% mass of ore-A was substituted by ore-B in A-1 and A-2 schemes, respectively. Likewise, 50% and 100% mass of ore-D was gradually replaced by ore-E in B-1 and B-2 schemes. The binary basicity (CaO/SiO₂, mass ratio) of sinter was 1.80, and the SiO₂ and MgO contents of sinter were 4.80% and 1.70% for all schemes, respectively. The proportion of quicklime in all schemes remained 3.50% to ensure the granulation effective. However, the proportions of limestone, dolomite and serpentine had a little changed to keep the same CaO, SiO₂ and MgO contents of sinter. The proportions of return fines and coke in all schemes were 25.00% and 4.00%, respectively.

Table 5. Proportions of raw materials in sinter mixture at different schemes/mass%.

Scheme	Base	A-1	A-2	B-1	B-2	C
Ore-A	14.55	7.26	0.00	14.52	14.49	0.00
Ore-B	14.55	21.79	29.02	14.52	14.49	28.92
Ore-C	8.84	8.83	8.81	8.82	8.80	8.78
Ore-D	11.31	11.29	11.28	5.64	0.00	0.00
Ore-E	11.31	11.29	11.28	16.93	22.53	22.47
Quicklime	3.50	3.50	3.50	3.50	3.50	3.50
Limestone	2.08	2.46	2.83	2.77	3.46	4.22
Dolomite	4.59	3.91	3.26	3.35	2.13	0.79
Serpentine	0.29	0.66	1.01	0.95	1.60	2.32
Return fine	25.00	25.00	25.00	25.00	25.00	25.00
Coke	4.00	4.00	4.00	4.00	4.00	4.00

In the sinter mixtures, larger than 1.0 mm of particles were used as cores and smaller than 0.5 mm of particles were used as adhering fines. Particles with size between 0.5 mm and 1.0 mm can be used as cores and adhering fines. Hence, in order to simulate the sintering process, the larger than 1.0 mm and smaller than 0.5 mm of particles in the sinter mixture were replaced by the particles with size between 1.8 mm and 2.0 mm and smaller than 0.15 mm in the micro-sintering tests. The sinter mixture was granulated to be quasi-particles by a laboratory pelletizer with some water. Then, the 7 g of samples were used in each run, put into quartz tube (Φ20×27 mm) and sintered in the horizontal furnace. The compositions of NO, CO, O₂ and CO₂ in off gas were analyzed by gas analyzer during the whole process and the total of NO emission and conversion rate of N element were calculated by Eqs. (2) and (3). The temperature and atmosphere of micro-sintering experiment were same to the combustion test.

2.2.3. Sinter Pot Test

Raw mixtures were uniformly mixed and granulated in a drum granulator. The granulated mixture was sintered in a sinter pot with 200 mm in diameter and 700 mm in height. The ignition time was 120 seconds and suction pressure was 11.0 kPa during sintering. Moreover, the compositions of NO, CO, O₂ and CO₂ in off gas were analyzed by flue gas analyzer. In addition, the average NOx emission concentration (A_NOx) and total NOx emission per ton of sinter (T_NOx) were used to evaluate the NOx emission during sintering process. The vertical sintering speed, yield, productivity and tumble index of sinter were investigated to evaluate the quality of sinter and calculated by the following Eqs. (5), (6), (7), (8).   

$$\nu =\frac{h}{t}$$(5)

  

$$\eta =\frac{m_{\text{0}}-m_{-\text{5}\text{.0mm}}}{m_{\text{0}}}\times \text{100\%}$$(6)

  

$$\gamma =\frac{\text{240}M}{\pi d^{\text{2}}t}\times {\text{10}}^{\text{3}}$$(7)

  

$$T{I}_{+\text{6}\text{.3mm}}=\frac{m_{+\text{6}\text{.3mm}}}{m}\times \text{100\%}$$(8)

where, ν (mm/min) is vertical sintering speed of sintering process, h (mm) is the height of sinter bed, t (min) is the sintering time, η (%) is the yield of sinter, m₀ (kg) is the weight of sinter except bottom material, m_{−5.0 mm} (kg) is the weight of return fines with size smaller than 5.0 mm, γ (t·m⁻²·h⁻¹) is the productivity of sinter, M (kg) is the weight of finished sinter, d (mm) is the diameter of sinter pot, TI_{+6.3 mm} is the tumble index of sinter, m₀ (kg) is the mass of sample, m_{+6.3 mm} (kg) is the mass of larger than 6.3 mm of tumbled sample.

3. Results

3.1. Combustion Behavior of Coke Covered with Different Coating Layers

Figure 4 shows the macro-morphology of the quasi-particle during coke combustion at different iron ores. It observed that the coating layer melted at high temperature and the S-type coke transformed to naked coke when coarse coke was coated by mixtures of iron ore and Ca(OH)₂ reagent. The combustion behavior of quasi-particle was similar to the report of Kasai.¹²⁾ Meanwhile, the ore-A and ore-B coating layers entirely melted and felled off the surface of coke at 1200°C which were lower than that of other iron ores. However, the ore-C and ore-D coating layers were difficult to melt. The initial melting temperature of the ore-E coating layer was about 1050°C, the surface of coke was partially exposed at 1200°C, and completely exposed at 1250°C. The cokes in all schemes were almost burned out at the end time of 1280°C.

Fig. 4.

Macro-morphology of quasi-particles during combustion process at different coating layers. (Online version in color.)

Figure 5 shows the combustion rate of coke with different coating layer. As shown in Fig. 5, the real-time combustion rates of coke rose sharply and then decreased steeply in the case of the Al₂O₃, ore-A and ore-B coating layers, reaching the peaks at 1128°C, 1202°C and 1230°C, respectively. In case of the ore-C and ore-D coating layers, the combustion rates of coke grew rapidly and then increased slowly, after that decreased gradually. They reached the peaks at 1262°C and 1269°C, respectively. Moreover, when coke was coated by ore-C, the combustion rate showed an inverted W-shape and reached the first and second peaks at 1207°C and 1262°C, respectively. Overall, compared with the Al₂O₃ coating layer, the peak of combustion rate was higher for the ore-A, ore-B and ore-E coating layers, but it was significantly lower for the ore-C and ore-D coating layers. The sooner the coke exposed, the sooner the combustion rate reached the peak.

Fig. 5.

Combustion rate of coke at different coating layers: (a) Al₂O₃; (b) ore-A and ore-B; (c) ore-C and ore-D; (d) ore-E. (Online version in color.)

3.2. Melt Fluidity of the Iron Ore Coating Layers

Figure 6 shows the melt fluidity of the coating layers with different iron ores during the coke combustion process. As can be seen from Fig. 6, the melt fluidity of the ore-A and ore-E coating layers were highest, then the ore-B coating layers, and that of the ore-C and ore-D coating layers were lowest. Although the binary basicity of the ore-B and ore-C coating layers were higher than that of the ore-A coating layer, the Al₂O₃ content of them were also higher than that of the ore-A coating layers, resulting in lower melt fluidity of the ore-B and ore-C coating layers. Zhang³⁰⁾ also found that the fluidity of liquid phase of adhering fines decreased with the Al₂O₃ content increased. The viscosity of melt increased with the increase of Al₂O₃ content.^31,32) Due to the lowest binary basicity of the ore-D coating layer, it had the lowest melt fluidity.

Fig. 6.

Melt fluidity of the different iron ore coating layers. (Online version in color.)

3.3. NOx Emission during Coke Combusting Process

The NOx emission concentrations during coke combustion at different coating layers were shown in Fig. 7. From Fig. 7(a) we can see that NOx emission concentration showed an inverted V-shape for the Al₂O₃ coating layer, which was similar to the combustion rate. However, in the case of the iron ore coating layers, there were two peaks of NOx emission concentrations, reaching the first peak at below 1100°C and the second peak at about 1280°C. The first peak was significantly higher than the second peak. Furthermore, compared with the Al₂O₃ coating layer, the maximum of NOx emission concentration decreased by 44.4%, 56.6%, 47.3%, 38.4% and 62.7% for the ore-A, ore-B, ore-C, ore-D and ore-E coating layers, respectively. Figure 7(a) shows the NOx emission concentration reached the peak at 1001°C while the combustion rate of coke reached the peak at 1128°C. Likewise, Fig. 7(b) illustrates NOx emission concentration reached the bottom at the corresponding temperature of the max combustion rate. As can be seen from Fig. 7(c), in the case of the ore-C and ore-D coating layers, the NOx emission concentration reached the second peak at the corresponding temperature of the max combustion rate. Figure 7(d) shows NOx emission concentration reached the peak at 971°C and declined to the bottom at the corresponding temperature of the first peak of combustion rate in the case of ore-E coating layers.

Fig. 7.

NOx emission concentration during coke combustion process at different coating layers: (a) Al₂O₃; (b) ore-A and ore-B; (c) ore-C and ore-D; (d) ore-E. (Online version in color.)

Figure 8 illustrates the conversion rates of N element during the combustion process of coke at different coating layers. Compared with the Al₂O₃ coating layer, the conversion rates of N element decreased over 23% for the iron ore coating layer. Furthermore, the conversion rate of N element was lowest and decreased by 42.5% for the ore-E coating layer while it was highest and decreased by 23.5% for the ore-D coating layer. For the limonite ore layers, the conversion rate of N element was obviously higher in the case of the ore-B coating layer than that of the ore-A coating layer, which decreased by 40.0% and 28.7%, respectively.

Fig. 8.

Conversion rates of N element during coke combustion process at different coating layers. (Online version in color.)

3.4. CO Emission during Coke Combusting Process

Figure 9 shows CO emission during coke combustion process at different coating layers. CO emission concentration showed the inverted V-shape trends. CO emission concentration reached the peak value at 952°C for Al₂O₃ and ore-D coating layers, that of ore-C coating layer was 1069°C, and it reached the maximum at 1190°C in the case of ore-A, ore-B and ore-C coating layers. In addition, it declined to zero at the begin time of 1280°C. Under the condition of Al₂O₃ coating layer, the peak of CO emission concentration and the mass of CO emission were significantly higher than those of iron ore coating layers. For iron ore coating layers, the mass of CO emission was highest in the case of ore-D coating layer while that of ore-E coating layer was lowest. The masses of CO emission were 0.090 mg, 0.065 mg and 0.059 mg in the case of ore-A, ore-B and ore-C coating layers, respectively.

Fig. 9.

CO emission concentration during coke combustion process at different coating layers: (a) Al₂O₃; (b) iron ores. (Online version in color.)

4. Discussion

4.1. Effect of Melt Fluidity of the Coating Layers on the Peak of Coke Combustion Rate

Figure 10 shows the relationship between peak combustion rate of coke and the melt fluidity of ore coating layers. It indicated that the peak combustion rate of coke linearly increased with the increasing of the melt fluidity of iron ore coating layers. During coke combustion process, the iron ore coating layer melted at the high temperature, and then the coke transformed from coated to naked. The more the coke exposed, the higher the peak of combustion rate.

Fig. 10.

Relation of peak combustion rate of coke and melt fluidity of iron ore coating layers. (Online version in color.)

4.2. Relation between the Combustion Rate and NOx Emission Concentration

The relations of the combustion rate and NOx emission concentration during the combustion process of coke at different coating layers were shown in Fig. 11. Figure 11(a) shows the NOx emission concentration increased linearly and then decreased slowly with the increasing of the combustion rate before the peak of combustion rate, but there was a positive linear correlation between them after the peak of combustion rate. As can be seen from the Fig. 11(b), there was an inverted V-shaped relation between the combustion rate and NOx emission concentration whether before or after the peak of combustion rate. Figure 11(c) illustrates that NOx emission concentration rose markedly and then declined rapidly, after that increased slightly with the combustion rate increased before the peak of combustion rate, while there was also a positive linear correlation between them after the peak of combustion rate. Figure 11(d) shows that before the first peak and after the second peak of the combustion rate, the relation between the combustion rate and NOx emission concentration was similar to an inverted V-shape. During the two peaks of the combustion rate, the relationships of them were two linear positive correlations. As can be seen from Fig. 11, the relationship between NOx emission and coke combustion rate was overall similar to the trends of before and after the peak of NOx emission.

Fig. 11.

Relation between combustion rate and NOx emission concentration at different coating layers: (a) Al₂O₃; (b) ore-A and ore-B; (c) ore-C; (d) ore-D and ore-E. (Online version in color.)

4.3. Effects of the Basicity of the Iron Ore Coating Layers and CO Emission on the Conversion Rate of N Element and the Melt Fluidity

Figure 12(a) shows the relationship of the basicity of the iron ore coating layers and the conversion rate of N element. It indicated that there was a negative linear correlation between the conversion rate of N element and the basicity of the iron ore coating layers. The formation of the calcium ferrite minerals increased with the increase of the binary basicity of the coating layers, and it promoted the reduction reaction between NO and CO.¹²⁾ Furthermore, due to the high Al₂O₃ content in ore-C, the generations of calcium ferrite minerals were inhibited, resulting in a higher conversion rate of N element than that of ore-B. Figure 12(b) shows the relationship of the mass of CO emission and the conversion rate of N element. It indicated that there was a positive linear correlation between the conversion rate of N element and the mass of CO emission. From Fig. 8, we can see that the mass of CO emission obviously decreased in the case of iron ore coating layers compared to the Al₂O₃ coating layer. In the existence of calcium ferrite, more CO was used to reduce NO, resulting in reduction of CO and NO emissions.

Fig. 12.

Correlativity of conversion rate of N element with basicity of adhering fines and CO emission during the combustion process of coke at different coating layers: (a) basicity of adhering fines; (b) CO emission. (Online version in color.)

Figure 13(a) shows the relation between the basicity of the iron ore coating layers and the melt fluidity. The melt fluidity of the coating layer increased with the increase of the basicity of the iron ore coating layers. Because basicity of the iron ore coating layers was contributed for the formation of the calcium ferrite minerals. However, although the binary basicity of ore-A coating layer was lower than that of other iron ores coating layer, the higher SiO₂ contents of it promoted the silicate formation and lower Al₂O₃ contents decreased the viscosity of melt,³⁰⁾ resulting in the higher melt fluidity of ore-A coating layer. Figure 13(b) shows the correlativity of the melts fluidity with CO emission. It can be seen that there was a negative linear correlation between the melt fluidity of the iron ore coating layers.and the mass of CO emission. The higher the melt fluidity of the iron ore coating layers, the more calcium ferrite minerals, which promoted the reaction between CO and NO, resulting in a reduction of CO emissions. For ore-A coating layer, due to the more silicate formation in the melts, less CO was used to react with NO.

Fig. 13.

Relationship of melts fluidity with basicity of adhering fines and CO emission during the combustion process of coke at different coating layers: (a) basicity of adhering fines; (b) CO emission. (Online version in color.)

4.4. Influencing Mechanism of the Iron Ore Coated Layer on the NOx Emission

From Figs. 5, 7 and 9, we can know that NOx emission concentrations showed the inverted W-shapes and the coke combustion behaviour of coke was similar at different iron ore coating layers. Hence, take the ore-A coating layer as an example to analyse the influence mechanism of iron ore coating layer on NOx emissions during the coke combustion process in this study. Figure 14 shows the NOx, CO emission and coke combustion behaviour with the iron ore coating layer. When coke was coated by the iron ore adhering fines, NOx emission mainly goes through the following stages during the coke combustion process:

Fig. 14.

NOx, CO emission and coke combustion rate with the iron ore coating layer. (Online version in color.)

I stage: prevent coke combusting at low temperature. Coke was tightly coated by the iron ore adhering fines and combusted slowly. Due to the lower temperature and less calcium ferrite formation at this time, the reduction reaction between CO and NO was difficult. Hence, with the increasing of the coke combustion rate, both CO and NO emissions increased.

II stage: strengthen NO reduction at medium temperature. Calcium ferrite minerals began to be formed and the coating layer gradually melted. As the coke combustion rated significantly increased, the formations of CO and NO were also obviously increased. However, under the catalysis of calcium ferrite minerals, a large amount of NO was reduced by CO to N₂, resulting in a reduction in NO emissions. The CO formation was greater than its consumption, and then its emissions continued to increase.

III stage: promote coke combusting at high temperature. The coating layer entirely melted and coke was translated from the coated type to the naked type. Coke was completely combusted at high temperatures, and NO formation further increased while CO formation decreased. As the reduction of NO by CO decreased, the NO emissions increased.

IV stage: burn out. The combustion rate of coke gradually decreased, and the formations of NO and CO was also gradually reduced until the coke was burned out.

As a consequence, the main influencing factors of NO reduction were the calcium ferrite minerals formation and CO emission during coke combustion process. With the formation of calcium ferrite minerals increased, it promoted the reduction reaction between NO and CO. During the coke combustion process, if CO emissions increased, it means less CO was used to reduce NO. Hence, in order to decrease NOx emission, the iron ore that formed more calcium ferrite minerals and less CO emissions during coke combustion process should be selected as the coated layer.

5. NOx Reduction Based on Optimizing the Ore Blending Structure

According to the above combustion tests, we can know that the maximum of NOx emission concentration and conversion rate of N element were lowest for the ore-B and ore-E. Hence, some ore blending structures were designed to reduce NOx emission. Figure 15 shows the NOx emission with different ore blending structures. Figure 15(a) illustrates that NOx emission concentration grew and declined rapidly, reaching the peak at about 1150°C and dropping to zero at the beginning of 1280°C. As the proportion of ore-B increased from 50% to 100% in blending ores, the peak of NOx emission concentration decreased slightly and then dramatically, which decreased by about 10% in A-2 scheme. Likewise, compared with base scheme, the peak of NOx emission concentration decreased by 20% in B-2 scheme. Figure 15(b) shows that the conversion rate of N element decreased gradually with increase of the proportion of Ore-B, so was Ore-E. Compared with base scheme, the conversion rate of N element decreased by 25.5% and 26.5% in A-2 and B-2 schemes, respectively.

Fig. 15.

NOx emission at different blending structures of iron ores: (a) NOx emission concentration; (b) conversion rates of N element. (Online version in color.)

6. Sinter Pot Tests

According to above researches, the NOx emission decreased by using ore-B and ore-E to replace ore-A and ore-D, respectively. Hence, in order to investigate the effectiveness and feasibility of the ore blending technique, the base and C schemes were conducted by sinter pot tests. The proportions of raw materials in sinter mixture are shown in Table 5. The mass ratio of ore-A and ore-B was 1:1 in base scheme, so was the mass ratio of ore-D and ore-E. Ore-A and ore-D was replaced by ore-B and ore-E in C scheme, respectively, and the proportion of others remained constantly. The results of sinter pot tests are shown in Table 6. Compared with base scheme, the average NOx emission concentration (A_NOx) and total NOx emission per ton of sinter (T_NOx)decreased by 20.2% and 22.6% in C scheme, respectively. Due to the increase of the basicity of adhering fines in C scheme, and the formation of calcium ferrite minerals increased, thereby it promoted the reduction of NO. In addition, compared with base scheme, the vertical sintering speed improved to 26.67 mm·min⁻¹ in C scheme. Because the average size of ore-E was significantly larger than that of ore-D, and the mass fraction of the fines below 0.5 mm in ore-E was less than that of ore-D. The yield increased from 77.79% to 78.66%. The formation of melting minerals increased since the increasing of the basicity of adhering fines. However, due to permeability improving and yield deterioration, the productivity increased obviously from 1.87 t·m⁻²∙h⁻¹ to 2.11 t·m⁻²∙h⁻¹. The tumble index was slightly decreased since the rise of the vertical sintering speed.

Table 6. Results of sinter pot tests.

Scheme	NOx emission indexes		Sintering indexes
	ANOx/ (mg·m−3)	TNOx/ (kg·t-1 sinter)	Vertical sintering speed/(mm·min−1)	Yield/%	Productivity/ (t·m−2∙h−1)	TI+6.3 mm/%
Base	198	0.53	24.58	77.79	1.87	62.00
C	158	0.41	26.67	78.66	2.11	61.76

7. Conclusions

(1) There was a positive linear correlation between the peaks of coke combustion rate and the melt fluidity of the iron ore coating layers. Meanwhile, The melt fluidity of the iron ore coating layers increased with the binary basicity of them. The melt fluidity of the limonite coating layer was highest.

(2) Due to the melting of the iron ore coating layer, coke transformed from coated to naked. As a result, the relationships between NOx emission concentration and combustion rate were like to 8-type except for two 8-types relationship in ore-E coating layer. Moreover, the formation of calcium ferrite increased with increasing of the binary basicity of iron ore coating layers, which promoted the reaction between CO and NO, resulting in a reduction of CO emissions. Calcium ferrite was easier to form in the low-silicon limonite and hematite coating layers.

(3) As the proportions of high-silicon Ore-A replaced by low-silicon Ore-B or high-silico Ore-D replaced by low-silicon Ore-E increased from 50% to 100% in sinter mixture, the peak of NOx emission concentration decreased slightly and then dramatically while the conversion rate of N element decreased gradually. Increasing the proportion of low silicon iron ores in sinter mixture can promote to reduce NOx emissions.

(4) By using low-silicon of Ore-B and Ore-E to replace high-silicon of Ore-A and Ore-D, the average NOx concentration and the total NOx emission per ton sinter decreased by 20.2% and 22.6%, respectively. At the same time, the vertical sintering speed improved 2.09 mm·min⁻¹, yield of sinter increased 0.87%, productivity rose 0.24 t·m⁻²∙h⁻¹ and tumble index slightly decreased.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (No. 51904127), the Natural Science Foundation of Jiangxi Province (No. 20192BAB216018), the PhD Project (No. 2018-YYB-05) and Generalized System of Preferences-One Type Project (No. 2018-XTPH1-05) of Jiangxi Academy of Sciences.

References

• 1)   J.  Qie,  C.  Zhang,  H.  Wang,  X.  Li and  X.  Shi: Sinter. Pellet., 41 (2016), 59 (in Chinese). https://doi.org/10.13403/j.sjqt.2016.06.078
• 2)   Y.  Su,  X.  Li and  X.  Fan: Sinter. Pellet., 38 (2013), 41 (in Chinese). https://doi.org/10.13403/j.sjqt.2013.06.013
• 3)   S.  Wu,  Z.  Que,  B.  Su and  Y.  Zhang: J. Eng. Stud., 9 (2017), 61 (in Chinese). https://kns.cnki.net/kcms/detail/detail.aspx?dbcode=CJFD&dbname=CJFDLAST2017&filename=GCKG201701009&v=SKWkA3Q1425WMm%25mmd2B0YMyLdron3aGmOZyBWVTS0v9WoFekERuS9Pe02pHVJTA1JzkW
• 4)   X.  Yan,  Y.  Li,  T.  Zhu and  F.  Qi: J. Environ. Eng. Technol., 5 (2015), 85 (in Chinese). https://doi.org/10.3969/j.issn.1674-991X.2015.02.013
• 5)   G.  Suzuki,  R.  Ando,  H.  Yoshikoshi,  Y.  Yamaoka and  S.  Nagaoka: Tetsu-to-Hagané, 61 (1975), 2775 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.61.13_2775
• 6)   W.  Lv,  X.  Fan,  X.  Min,  M.  Gan,  X.  Chen and  Z.  Ji: ISIJ Int., 58 (2018), 236. https://doi.org/10.2355/isijinternational.ISIJINT-2017-370
• 7)   Y.  Hida,  K.  Ito and  M.  Sasaki: Tetsu-to-Hagané, 67 (1981), 1934 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.67.11_1934
• 8)   Y.  Hida,  M.  Sasaki and  K.  Ito: Tetsu-to-Hagané, 66 (1980), 1801 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.66.13_1801
• 9)   Y.  Hida,  M.  Sasaki,  K.  Ito and  T.  Enokido: Tetsu-to-Hagané, 67 (1981), 2625 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.67.16_2625
• 10)   H.  Zhou,  M.  Zhou,  Z.  Liu,  M.  Cheng and  J.  Chen: Fuel, 179 (2016), 322. https://doi.org/10.1016/j.fuel.2016.03.098
• 11)   S.  Wu,  T.  Sugiyama,  K.  Morioka,  E.  Kasai and  Y.  Omori: Tetsu-to-Hagané, 80 (1994), 276 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.80.4_276
• 12)   E.  Kasai,  S.  Wu,  T.  Sugiyama,  S.  Inaba and  Y.  Omori: Tetsu-to-Hagané, 78 (1992), 1005 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.78.7_1005
• 13)   J.  Pan: Ph.D. thesis, Central South University, (2007), https://kns.cnki.net/kcms/detail/detail.aspx?dbcode=CDFD&dbname=CDFD9908&filename=2008166104.nh&v=NKrn37QeoDXA1b0tDRqRNLn%25mmd2BTrMRdZwjFgbqvXbP6T7jA99KSh7VnDWkLAuGfdnR, (accessed 2007-10-01).
• 14)   H.  Zhou,  M.  Cheng,  M.  Zhou,  Z.  Liu,  R.  Liu and  K.  Cen: Appl. Therm. Eng., 94 (2016), 786. https://doi.org/10.1016/j.applthermaleng.2015.09.059
• 15)   H.  Zhou,  Z.  Liu,  M.  Cheng,  M.  Zhou and  R.  Liu: Energy Fuels, 29 (2015), 974. https://doi.org/10.1021/ef502524y
• 16)   K.  Morioka,  S.  Inaba,  M.  Shimizu,  K.  Ano and  T.  Sugiyama: ISIJ Int., 40 (2000), 280. https://doi.org/10.2355/isijinternational.40.280
• 17)   K.  Katayama and  S.  Kasama: ISIJ Int., 56 (2016), 1563. https://doi.org/10.2355/isijinternational.ISIJINT-2015-742
• 18)   M.  Gan,  X.  Fan,  W.  Lv,  X.  Chen,  Z.  Ji,  T.  Jiang,  Z.  Yu and  Y.  Zhou: Powder Technol., 301 (2016), 478. https://doi.org/10.1016/j.powtec.2016.05.043
• 19)   Z.  Yu,  X.  Fan,  M.  Gan and  X.  Chen: J. Iron Steel Res. Int., 24 (2017), 1184. https://doi.org/10.1016/S1006-706X(18)30016-5
• 20)   H.  Zhou,  M.  Cheng,  J.  Zhao,  M.  Zhou and  Z.  Liu: J. Zhejiang Univ. Sci. A, 19 (2018), 479. https://doi.org/10.1631/jzus.A1700193
• 21)   Z.  Que,  S.  Wu and  X.  Ai: Chin. J. Eng., 42 (2020), 163. https://doi.org/10.13374/j.issn2095-9389.2019.02.21.001
• 22)   K.-i.  Ohno,  K.  Noda,  K.  Nishioka,  T.  Maeda and  M.  Shimizu: ISIJ Int., 53 (2013), 1588. https://doi.org/10.2355/isijinternational.53.1588
• 23)   K.-i.  Ohno,  K.  Noda,  K.  Nishioka,  T.  Maeda and  M.  Shimizu: ISIJ Int., 53 (2013), 1642. https://doi.org/10.2355/isijinternational.53.1642
• 24)   K.-i.  Ohno,  K.  Noda,  K.  Nishioka,  T.  Maeda and  M.  Shimizu: Tetsu-to-Hagané, 101 (2015), 184. https://doi.org/10.2355/tetsutohagane.101.184
• 25)   H.  Zhou,  P.  Ma,  M.  Cheng,  M.  Zhou and  Y.  Li: ISIJ Int., 58 (2018), 1650. https://doi.org/10.2355/isijinternational.ISIJINT-2018-185
• 26)   P.  Ma,  M.  Cheng,  M.  Zhou,  Y.  Li and  H.  Zhou: Chin. J. Eng., 41 (2019), 316 (in Chinese). https://doi.org/10.13374/j.issn2095-9389.2019.03.004
• 27)   Y.  Hida,  M.  Sasaki,  T.  Enokido,  Y.  Umezu,  T.  Iida and  S.  Uno: Tetsu-to-Hagané, 68 (1982), 400 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.68.3_400
• 28)   S.  Wu,  B.  Su,  Y.  Qi,  M.  Kou,  Y.  Li and  W.  Zhang: Metall. Mater. Trans. B, 48 (2017), 2469. https://doi.org/10.1007/s11663-017-1059-9
• 29)   J.  Okazaki,  K.  Higuchi,  Y.  Hosotani and  K.  Shinagawa: ISIJ Int., 43 (2003), 1384. https://doi.org/10.2355/isijinternational.43.1384
• 30)   G.  Zhang,  S.  Wu,  S.  Chen,  B.  Su,  Z.  Que and  C.  Hou: Int. J. Miner. Metall. Mater., 21 (2014), 962. https://doi.org/10.1007/s12613-014-0996-4
• 31)   S.  Machida,  K.  Nushiro,  K.  Ichikawa,  H.  Noda and  H.  Sakai: ISIJ Int., 45 (2005), 513. https://doi.org/10.2355/isijinternational.45.513
• 32)   S.  Wu,  J.  Du,  H.  Ma,  Y.  Tian and  H.  Xu: J. Univ. Sci. Technol. Beijing, 27 (2005), 291 (in Chinese). https://doi.org/10.13374/j.issn1001-053x.2005.03.009