Numerical Study of the Influence of Regenerator Structure on the Performance of Hot Blast Stoves

Abstract

The properties and arrangement of checker bricks in regenerators are crucial for the heat exchange process of hot blast stoves. In this study, a 3D fluid flow heat transfer model is established to analyze the influence of three regenerator layered structures on the combustion and air supply performance of hot blast stoves. The results show that a “silica bricks–high-alumina bricks–clay bricks” three-layer arrangement in regenerators produces a “thermal conduction hindrance effect” at the interface between silica and high-alumina bricks during the 2 h combustion period, which raises the local temperature and improves air supply performance. Compared to the “silica bricks–clay bricks” and “high-alumina bricks–silica bricks–clay bricks” structures, this setup increases the maximum air supply temperature difference to 23 and 64 K, respectively, and extends the effective air supply time by 80 and 320 s, respectively. However, the “thermal conduction hindrance effect” diminishes over longer combustion periods, and by 4 h, the performance across all structures becomes increasingly consistent. Additionally, the study suggests that the temperature level and distribution in the upper part of the regenerator are the key factors determining the air supply performance of hot blast stoves.

1. Introduction

With the continuous development of blast furnace ironmaking technology towards modernization, greening, and high efficiency, the performance improvement of hot blast stoves, which are important auxiliary equipment for blast furnaces, plays a crucial role in optimizing the entire ironmaking process. The primary function of hot blast stoves is to continuously provide high-temperature air to blast furnaces through a heat-exchange process. High-temperature air not only helps to increase the temperature of blast furnaces and promote the reduction reaction of iron ore but also effectively reduces energy consumption in the ironmaking process and improves production efficiency [1,2,3,4]. Regenerators of hot blast stoves are composed of stacked checker bricks that act as carriers for heat exchange in hot blast stoves. During the combustion period, the high-temperature flue gas generated by the combustion of gas and combustion air heats checker bricks through convection and radiation heat transfer; during the air supply period, checker bricks transfer the heat stored inside to the cold air through convection heat transfer, resulting in high-temperature air used for blast furnace ironmaking. Evidently, the heat storage and release efficiencies of checker bricks directly affect the performance of hot blast stoves [5,6,7].

Many factors affect the heat storage and release efficiencies of regenerators, including the thermophysical parameters and arrangement of checker bricks, as well as the flow characteristics and temperature distribution of the flue gas inside the regenerators. Therefore, a detailed analysis of these influencing factors is the basis for improving the structure of hot blast stoves, optimizing operating parameters, and enhancing thermal efficiency. In recent years, with rapid advancements in computer software and simulation techniques, numerical studies have been widely used to simulate the performance of hot blast stoves [8,9,10,11,12,13,14,15,16,17,18]. In these studies, the primary focus on regenerators was on their heat transfer characteristics and the impact of the checker bricks’ physical parameters on performance. Zhong et al. [19] analyzed the heat exchange characteristics between checker bricks and flue gas in the regenerator of a self-preheating hot blast stove by employing a 3D unsteady heat conduction model. They concluded that the self-preheating process effectively enhanced the heat storage capacity of the regenerator, subsequently increasing the air supply temperature. Guo et al. [20] studied the influence of physical parameters, such as thermal conductivity, density, and specific heat capacity of checker bricks, on the temperature distribution and heat storage capacity of the regenerator based on an established mathematical model of the heat transfer process and temperature distribution calculation program. Kimura et al. [21] developed a 3D mathematical model to analyze the transient mass and heat transfer in a hot blast stove. They studied the turbulent mixing, combustion, and heat transfer performance of gases within the combustion chamber and also calculated the optimal layout and operating conditions for the checker bricks in the regenerator. This approach combines the combustion chamber and regenerator, making the calculation results more consistent with the actual situation. Chen et al. [6,22,23] established a 2D temperature field mathematical model for checker bricks in a regenerator, analyzed the heat transfer parameters of gases within different checker bricks, and proposed optimization designs for the pore structure of the checker bricks. Hou et al. [24] used numerical simulation methods to study the influence of a regenerator’s top structure on airflow distribution and found that airflow uniformity improved when the top was modified to a convex shape.

However, the research on the regenerators mentioned above did not consider the influence of checker brick arrangements with different materials on the performance of hot blast stoves. The commonly used refractory materials for checker bricks in hot blast stoves include silica, high-alumina, and clay bricks. Among them, silica or high-alumina bricks are often used in the high-temperature areas of the regenerator’s upper part, whereas clay bricks are generally used in medium- and low-temperature areas because of their lower softening temperature under load. Owing to the different mechanical and thermal properties of refractory materials, their selection and arrangement can significantly affect the heat exchange performance of hot blast stoves during the combustion and supply periods. In this study, based on a top combustion hot blast stove, a 3D fluid flow heat transfer model was established to systematically study the influence of different material checker brick arrangements on the combustion and air supply characteristics of hot blast stoves. By comparing the distribution characteristics of the temperature fields, the relationship between the regenerator’s structure and the performance of the hot blast stove was determined, and a reasonable combination of heat storage materials was identified. This provides valuable research results for optimizing the design of hot blast stoves.

2. Model Design and Condition

2.1. 3D Computational Model

A 3D physical model of a top combustion hot blast stove equipped with a 2350 m³ blast furnace is shown in Figure 1. It consists of four parts: the pre-combustion chamber, combustion chamber, regenerator, and furnace grill. The stove’s height was 46.9 m, and the bottom diameter of the combustion chamber was 8.8 m. In turn, the regenerator’s diameter and height were 8.3 and 26 m, respectively. The regenerator is composed of two layers of checker bricks, with silica bricks at the top and clay bricks at the bottom. To analyze the impact of the stratified structure of checker bricks with different material on the performance of hot blast stoves, a layer of low-creep high-alumina bricks was added to the existing two-layer checker brick structure without altering the dimensions of the regenerator. This resulted in the creation of three distinct regenerator structural models, as depicted in Figure 2, where the two- and three-layers checker brick structures are referred to as the 2L and 3L structures, respectively. The physical parameters of the checker bricks are listed in

Table 1. Physical parameters of checker bricks.

Checker Brick	Density/(kgm−3)	Thermal Conductivity/(Wm−1K−1)	Specific Heat/(Jkg−1K−1)
Silica brick	1900	0.93 + 0.0007 t 1	794.0 + 0.251 t
High-alumina brick	2500	1.51 − 0.00019 t	836.8 + 0.234 t
Clay brick	2070	0.84 + 0.00052 t	836.8 + 0.263 t

¹ The unit for “t” is K.

[25].

Considering the medium characteristics in different areas of the hot blast stove, the computational domain of each model was divided into two main parts: the fluid domain, including the pre-combustion chamber, combustion chamber, and furnace grill, and the porous medium domain, namely the regenerator. The fluid domains were discretized using a tetrahedral mesh, whereas the regenerator was segmented using a hexahedral mesh. The overall model consisted of 1,320,464 (2L) and 1,318,356 (3L) grid cells.

In addition, the numerical model used in this study has been validated through on-site testing data [15] and has undergone mesh independence verification [26] as parts of prior studies. Consequently, the computational method utilized for analyzing the regenerator structure is deemed to be credible and dependable.

2.2. Mathematical Model

The operation of a hot blast stove involves various intricate physical and chemical processes. By employing ANSYS FLUENT 19.2 software, appropriate mathematical models were developed for different computational domains within the system.

2.2.1. Turbulence Model

The turbulent behavior of the fluid within the hot blast stove was modeled using the realizable k-ε model, known for its stability and computational efficiency, as depicted by the following equations [27,28]:

$${\displaystyle \frac{\partial (\rho k)}{\partial t}}+\nabla \cdot \left(\rho k{\overrightarrow{v}}_i\right)=\nabla \cdot \left[\left(\mu +{\displaystyle \frac{{\mu }_t}{{\sigma }_k}}\right)\nabla k\right]+G_k-\rho \epsilon ,$$

(1)

and

$${\displaystyle \frac{\partial (\rho \epsilon )}{\partial t}}+\nabla \cdot \left(\rho \epsilon {\overrightarrow{v}}_i\right)=\nabla \cdot \left[\left(\mu +{\displaystyle \frac{{\mu }_t}{{\sigma }_{\epsilon }}}\right)\nabla \epsilon \right]+C_{1\epsilon }{\displaystyle \frac{\epsilon }{k}}{G}_k-C_{2\epsilon }\rho {\displaystyle \frac{{\epsilon }^2}{k}},$$

(2)

where ρ, $\overrightarrow{v}$, and μ correspond to the fluid’ density, velocity vector, and dynamic viscosity, respectively. k and ε denote the turbulence kinetic energy and dissipation rate, respectively. The symbol t signifies time, whereas G_k indicates the turbulence kinetic energy arising from the mean velocity gradient. In the proposed model, the constants C_1ε and C_2ε are set to 1.44 and 1.9, respectively. σ_k =1.0 and σ_ε =1.2 refer to the Prandtl numbers associated with the k and ε equations, respectively.

2.2.2. Combustion Model

An eddy dissipation model was selected to simulate the mixing and combustion processes within the hot blast stove, assuming that the chemical reaction rate was much higher than the turbulent mixing rate. In other words, the speed-controlling step of the combustion reaction is turbulent mixing. Consequently, the ultimate reaction rate was determined by the mass fractions of the reactants and products. The net rate R_i,r of the component i produced via the reaction r is contingent on the lower value of the following pair of equations:

$$R_{i,r}={\vartheta }_{i,r}^{\prime }{M}_{w,i}A\rho ({\displaystyle \frac{\epsilon }{k}}){\min }_R\left({\displaystyle \frac{Y_R}{{\vartheta }_{R,r}^{\prime }{M}_{w,R}}}\right),$$

(3)

and

$$R_{i,r}={\vartheta }_{i,r}^{\prime }{M}_{w,i}AB\rho ({\displaystyle \frac{\epsilon }{k}}){\displaystyle \frac{{\displaystyle \sum _p{Y}_p}}{{\displaystyle \sum _j^N{\vartheta }_{j,r}^{″}}{M}_{w,j}}},$$

(4)

where ${\vartheta }_{i,r}^{\prime }$ and ${\vartheta }_{i,r}^{″}$ denote the stoichiometric coefficients of the reactant and product, respectively, involved in the chemical reaction r pertaining for component i. M_w,i signifies the molecular weight of component i, whereas ρ indicates the mixture density. N refers to the number of species in the system. Y_R and Y_P correspond to the mass fractions of the reactant R and product P, respectively. Empirical constants are represented by A = 4.0 and B = 5.0.

2.2.3. Radiation Model

The discrete ordinate (DO) method was adopted to model the radiative transfer inside the hot blast stove, which is formulated as follows [29,30]:

$$\nabla \cdot (I(\overrightarrow{r},\overrightarrow{s})\overrightarrow{s})+\left(a+{\sigma }_S\right)I(\overrightarrow{r},\overrightarrow{s})=a{n}^2{\displaystyle \frac{\sigma T^4}{\pi }}+{\displaystyle \frac{{\sigma }_S}{4\pi }}{\displaystyle {\int }_0^{4\pi }I\left(\overrightarrow{r},{\overrightarrow{s}}^{\prime }\right)\mathsf{\Phi }\left(\overrightarrow{s},{\overrightarrow{s}}^{\prime }\right)}d{\Omega }^{\prime },$$

(5)

where the terms $\overrightarrow{r}$, $\overrightarrow{s}$, and ${\overrightarrow{s}}^{\prime }$ correspond to the vectors that denote the location, orientation, and scattering direction, respectively. α, σ_s, and n signify the coefficients for spectral absorption, scattering, and refraction, respectively. I refers to the radiation intensity, T indicates the local temperature, Φ denotes the phase function, ${\mathsf{\Omega }}^{\prime }$ stands for the solid angle, and σ represents the Stefan–Boltzmann constant.

2.2.4. Porous Media Model

The model designed for porous media was used to simulate the heat transfer process between the checker bricks within the regenerator, high-temperature flue gas, and cold air. For a single phase passing through a porous medium, the conservation equations for momentum and energy are as follows:

$${\displaystyle \frac{\partial }{\partial t}}(\gamma \rho \overrightarrow{v})+\nabla \cdot (\gamma \rho \overrightarrow{v}\overrightarrow{v})+\gamma \nabla p-\nabla \cdot \left[\gamma \mu \left(\nabla \overrightarrow{v}+\nabla {\overrightarrow{v}}^T\right){\displaystyle \frac{2}{3}}\gamma \nabla \cdot \overrightarrow{v}\right]-\left({\displaystyle \frac{\mu }{\beta }}+{\displaystyle \frac{C_2\rho }{2}}\left|\overrightarrow{v}\right|\overrightarrow{v}\right)=0,$$

(6)

and

$${\displaystyle \frac{\partial }{\partial t}}\left(\gamma {\rho }_f{E}_f+(1-\gamma ){\rho }_S{E}_S\right)+\nabla \cdot \left[\overrightarrow{v}\left({\rho }_f{E}_f+P\right)\right]=\nabla \cdot \left(k_{eff}\nabla T-{\displaystyle \sum _i{h}_i{\overrightarrow{J}}_i}\right),$$

(7)

where γ denotes the porosity and β and C₂ correspond to the viscous and inertia resistance coefficients, respectively. E_f and E_S signify the total fluid and solid medium energy, respectively. ρ_f and ρ_S indicate the fluid and solid densities, respectively. K_eff denotes the effective thermal conductivity. h_i and $\overrightarrow{J_i}$ denote the enthalpy and diffusion flux of the component i, respectively [31].

2.3. Solver Settings and Convergence Criteria

In this study, the Fluent solver was configured using the SIMPLE algorithm. The residual for the continuity equation was set to 10⁻³, while the residual for the energy equation was set to 10⁻⁶. Additionally, the residuals for the turbulence kinetic energy k and its dissipation rate ε were both set to 10⁻³. Monitors were added to track the variation in outlet temperature. All these configurations ensure that the solver can achieve accurate and stable solutions for the given computational fluid dynamics problem.

2.4. Boundary Condition

All boundary conditions were based on the actual operating parameters of a plant. The fuel used was blast furnace gas, the composition of which is listed in

Table 2. Composition of blast furnace gas.

Component	CO	CO2	H2O	H2	N2
Percentage composition	0.24	0.19	0.023	0.01	0.537

[15].

Table 3. Operational parameters of the hot blast stove.

Fluid Type	Inlet Flow Rate/(Nm3h−1)	Inlet Temperature/(K)
Gas	118,300	423
Combustion air	77,400	473
Cold air	334,200	473

lists the operational parameters of the hot blast stove. During the combustion period, the gas and air inlets were configured as velocity boundaries, and the two flue gas outlets were set as pressure boundaries at 0 Pa. During the air supply period, the cold air input was used as the velocity boundary, and the hot air output was considered the pressure boundary at 0 Pa. All surface walls are defined as adiabatic and no-slip walls.

3. Results and Discussion

3.1. Temperature Field in the Combustion Period

3.1.1. Temperature Distribution in Relevant Sections

Figure 3 illustrates the temperature distribution across the vertical sections of the three types of regenerator structures of hot blast stoves at the end of the combustion periods of 2 h and 4 h, respectively. As depicted, the overall distribution of the temperature field for the three structures was similar: in the combustion chamber, the temperature distribution exhibited a “saddle-like” structure with a central region significantly lower than that near the walls. This is mainly due to the “Coanda effect” generated when the combustion gases in the prechamber flowed towards the combustion chamber; that is, when fluid flows over the surface of an object, it tends to “adhere” to the surface and flow along the surface of the object. In the regenerator, the upper region had a higher temperature due to its proximity to the combustion chamber, whereas the temperature of the lower region gradually decreased as the distance from the heat source increased. Simultaneously, the characteristics of the combustion chamber temperature field are inherited in the regenerator. Taking the 2L structure as an example, the temperature distributions of different horizontal sections of the regenerator (see Figure 3a for the section positions) during the combustion periods of 2 h and 4 h are shown in Figure 4. The figure indicates that the temperature across each section exhibits a “circular” pattern with a lower central temperature corresponding to the temperature distribution of the combustion chamber.

In addition, there were significant differences in the temperature distribution among the three regenerator structures. During the combustion periods of 2 h and 4 h, although the highest temperatures across the three structures were relatively similar, the extent of the high-temperature areas covered showed some variation. For example, during the 2 h combustion period, the total area of the vertical sections in the high-temperature region (T ≥ 1500 K) for the three structures is illustrated in Figure 5. Here, the area of the high temperature zone for structure 3L(I) was reduced by 22.43% compared to structure 2L, while the corresponding area for structure 3L(II) increased by 8.05%. These data indicate that the introduction of a high-alumina brick layer and the adjustment of its relative positions affect the internal temperature distribution within the hot blast stove.

3.1.2. Temperature Distribution in the Regenerator

As shown in Figure 2, the three-layer regenerator structure came in two configurations, namely 3L(I) and 3L(II). The main difference between them lay in the arrangement of high-alumina and silica bricks: in 3L(I), the high-alumina bricks were located above the silica bricks, whereas in 3L(II), the situation was completely opposite. Given that the bottom layers of the three regenerator structures in this study were composed of clay bricks, it was only necessary to focus on how the arrangement of silica and high-alumina bricks affected the performance of the regenerator.

The thermal performance of checker bricks primarily depends on their physical properties, including density, thermal conductivity, and specific heat capacity. According to the data in Table 1, the density of high-alumina bricks is 2500 kgm⁻³, which is significantly higher than that of silica bricks at 1900 kgm⁻³. By examining the changes in thermal conductivity and specific heat capacity with temperature (as shown in Figure 6 and Figure 7), it was found that within the operating temperature range of hot blast stoves, the difference in specific heat capacity between silica and high-alumina bricks was relatively small, and the specific heat capacity of high-alumina bricks was only 1%–4% higher than that of silica bricks. However, there was a significant difference in their thermal conductivities, particularly in high-temperature areas. For instance, the thermal conductivities of silica and high-alumina bricks were 1.84 and 1.26 Wm⁻¹K⁻¹, respectively, at 1300 K, indicating that the thermal conductivity of the high-alumina bricks was 31.52% lower than that of the silica bricks. An increase in material density means that more heat can be stored in the same volume; a higher thermal conductivity indicates a faster rate of heat transfer, which helps transport heat more efficiently. Therefore, when evaluating the comprehensive performance of checker bricks, it is important to consider the impact of these thermophysical parameters on both the heat transfer efficiency and heat storage capacity of the carrier fluid within the regenerator. This evaluation should be based on numerical simulation data from the air supply process conducted under the same conditions.

• The 2 h combustion period

Figure 8 illustrates the temperature distribution along the design centerline of the regenerator for the three distinct structures. After the 2 h combustion period, the temperature within the heat storage body, along the centerline, was significantly lower for the 3L(I) structure compared to both the 2L and 3L(II) structures (Figure 8a). This variation is primarily attributed to the thermal conductivity of the high-alumina bricks in the upper layer, which under high-temperature conditions is markedly lower than that of the silica bricks (as depicted in Figure 6). Additionally, the density of the high-alumina bricks significantly exceeded that of the silica bricks. These combined factors impede the heat transfer from the surface to the interior of the checker bricks, leading to a comparatively lower temperature within the 3L(I) regenerator.

Above the interface where the silica bricks meet the high-alumina bricks, the temperature of the 3L(II) heat-storage body experienced a notable rise, surpassing the temperature of the 2L structure at the same location. This advantage persists in areas close to the junction. However, as the distance from the junction area along the centerline increased, the temperature of the 3L(II) heat-storage body dropped sharply, eventually falling below that of the 2L structure and approaching the temperature levels of the 3L(I) structure. The distinctive temperature distribution within the checker bricks is attributed to the thermophysical properties of the heat-storing material. During the 2 h combustion period, the temperature at the interface of the silica and high-alumina bricks was approximately 1300 K (Figure 8a). At this temperature, the specific heat capacities of the high-alumina and silica bricks were similar. However, the thermal conductivity of the high-alumina bricks was reduced by 31.52% compared with that of the silica bricks. This reduction indicates a higher thermal resistance within the high-alumina brick layer, which results in a decelerated heat transfer rate when the high-temperature flue gas reaches the interface. Consequently, for the 3L(II) structure, the flue gas temperature increased, leading to a higher temperature in the silica brick region above the junction. Essentially, the “thermal conduction hindrance effect” of the high-alumina bricks contributed to the increased temperature in the silica brick area. This increase in temperature also extends to the shallow region of the high-alumina bricks adjacent to the silica bricks, enhancing their thermal storage capacities. With densities of 2500 kgm⁻³ for high-alumina bricks and 1900 kgm⁻³ for silica bricks, the density of high-alumina bricks is 31.58% higher than that of silica bricks. This implies that under the same temperature conditions, a unit area of high-alumina bricks can store more heat energy. Consequently, during the air supply phase, this area released more heat. In summary, the temperature variation at the interface of silica and high-alumina bricks in the 3L(II) structure is beneficial for improving the air supply temperature. However, in the lower region of the regenerator, this structure may affect the heat exchange efficiency between the cold air entering from the bottom of the hot blast stove and the low-temperature area of the checker bricks. To evaluate the comprehensive impact of this structure, it was necessary to verify it by comparing its air supply characteristics.

• The 4 h combustion period

Upon examining Figure 8b, it is evident that the 3L(II) structure exhibited a considerable variation in regenerator temperature distribution between the 2 h and 4 h combustion periods: at the interface of the silica and high-alumina bricks, there was no occurrence of increased temperature in the adjacent heat storage body due to the “thermal conduction hindrance effect” of the high-alumina bricks. The temperature distribution curves of the 2L and 3L(II) structures coincide from the upper part of the heat-storage body to the interface. This indicates that when the combustion period was extended to 4 h, the heat exchange process between the checker bricks and high-temperature flue gas was relatively complete, gradually forming a stable temperature field in this area. In other words, although the “thermal conduction hindrance effect” of high-alumina bricks reduces the heat transfer rate, this can be partially compensated by extending the combustion period. Consequently, no significant temperature increase was observed at the interface area of the silica and high-alumina bricks. However, in the lower half of the heat storage body, after fully entering the area of the high alumina bricks, the temperature of the 3L(II) structure was significantly lower than that of the 2L structure. This reduction in the temperature was attributed to the increased thermal resistance of the high-alumina bricks, which resulted in a decrease in the heat transfer rate. For the 3L(I) structure, as the combustion period increased, the high-alumina bricks stored more heat, and the rate of heat transfer was enhanced, causing the temperature in the upper region of the regenerator to approach that of the 2L and 3L(II) structures. In the lower region, the temperature is higher than that of the 3L(II) structure.

3.2. Characteristic Analysis of Air Supply Period

Hot blast stoves provide hot air to blast furnaces by introducing cold air from the bottom of the stoves and exchanging heat with checker bricks. Therefore, at the end of the combustion period, the temperature level and distribution in the heat storage body directly affect the performance of the air supply. The two main criteria for evaluating the air supply performance of hot blast stoves are the air supply temperature and effective air supply time (i.e., the time required from the start of the air supply to the temperature drop to 1273 K). Usually, the higher the air supply temperature or the longer the effective supply time, the better the performance of hot blast stoves.

Figure 9 shows the temperature curves at the hot air outlet during the air supply period for the three types of regenerator structures over time. As the air supply time increased, the hot-air temperature decreased parabolically. At the beginning of the air supply, when the hot air flow remained constant, the temperature decreased at a relatively slow rate because the heat storage body had a higher temperature and stored a large amount of thermal energy. In the middle stage of the air supply, the temperature decreased more quickly. In the later stage of the air supply, as the temperature of the heat storage body continued to decrease, the temperature of the hot air decreased sharply. When the remaining thermal energy in the heat storage body was insufficient to maintain the minimum effective air supply temperature (1273 K), the end of the air supply period was marked.

Upon examining the air supply curve during the 2 h combustion period, as shown in Figure 9a, it is observed that the initial air supply temperature for the 3L(II) structure was 5 and 15 K higher than that of the 2L and 3L(I) structures, respectively. In the middle stage of the air supply, this difference became apparent, with the maximum temperature difference between the 3L(II) and both the 2L and the 3L(I) structures, reaching 23 and 64 K, respectively. Additionally, the effective air supply time for the 3L(II) structure was extended by 80 and 320 s compared to the 2L and 3L(I) structures, respectively. This result is consistent with the temperature distribution shown in Figure 8a for the regenerator, and it also confirms that the “thermal conduction hindrance effect” occurring at the interface of the middle high-alumina brick layer increased the temperature in nearby areas, which was beneficial for increasing the air supply temperature.

As shown in Figure 9b, during the 4 h combustion period, the air supply temperature curves for the three structures converged, unlike in the 2 h period. Notably, the curve for the 2L structure nearly overlapped with that of the 3L(II) structure. This phenomenon indicates that as the combustion period extended, the heat exchange process became more efficient, leading to less significant differences in air supply performance among the 2L, 3L(I), and 3L(II) structures.

A thorough analysis of the regenerators’ temperature distribution, as presented in Figure 8, reveals that the temperature in the lower half of the 3L(II) structure was consistently lower than that of the 2L structure, regardless of the combustion period’s length. Notably, during the 4 h combustion period, the temperature of the 3L(II) structure was observed to be even lower than that of the 3L(I) structure. This difference is primarily attributed to the distinct thermophysical parameters and arrangements of the structures. However, according to the air supply curve in Figure 9, the temperature distribution in the lower half of the regenerator did not fundamentally impact the air supply performance of hot blast stoves. Therefore, it can be inferred that the temperature level and distribution in the upper part of the regenerators are key factors that determine the air supply performance of hot blast stoves.

4. Conclusions

Based on a top combustion hot blast stove equipped with a 2350 m³ blast furnace, three models with different layered structures of the regenerator were constructed, including a double-layer structure 2L (silica brick–clay brick) and two three-layer structures, namely 3L(I) (high-alumina brick–silica brick–clay brick) and 3L(II) (silica brick–high-alumina brick–clay brick). Utilizing the established 3D fluid flow heat transfer model, the influence of three distinct regenerator layered structures on the performance of hot blast stoves was studied, leading to the following conclusions.

(1)

When the regenerator adopted the 3L(II) structure, the “thermal conduction hindrance effect” occurred at the junction area between the silica and the high-alumina bricks during the 2 h combustion period. This effect effectively increased the temperature of that area, thereby raising the air supply temperature and prolonging the effective air supply time. Compared to the 2L and 3L(I) structures, this setup increased the maximum air supply temperature difference to 23 and 64 K, respectively, and extended the effective air supply time by 80 and 320 s, respectively.

(2)

The “thermal conduction hindrance effect” was time-sensitive. As the combustion period extended, the checker bricks had sufficient time to exchange heat with the high-temperature flue gas, gradually establishing a stable temperature field at the junction area between the silica and high-alumina bricks. Consequently, the “thermal conduction hindrance effect” diminished over time, resulting in the air supply performances of the 2L, 3L(I), and 3L(II) structures becoming consistent during the 4 h combustion period.

(3)

Among the three regenerator structures, the temperature distribution in the lower half was primarily influenced by their thermal properties and configurations, regardless of the length of the combustion period. However, these variations did not significantly affect the air supply performance of hot blast stoves. This suggests that the temperature level and distribution in the upper part of the regenerator are the key factors determining the air supply performance of hot blast stoves.