Gas-particle flow and rapid load-up characteristics of a novel deep peak regulation burner

Abstract

Existing swirling combustion technology, which relies on faulty coal, is unable to meet deep peak shaving demands without auxiliary methods. This paper developed a deep peak regulation burner (DPRB) to achieve stable combustion at 15%–30% of the boiler’s rated load without auxiliary support. Gas-particle tests, industrial trials, and transient numerical simulations were conducted to evaluate the burner’s performance. At full rated load, the DPRB formed a central recirculation zone (RZ) with a length of 1.5d and a diameter of 0.58d (where d represents the outlet diameter). At 40%, 20%, and 15% rated loads, the RZ became annular, with diameters of 0.30d, 0.40d, and 0.39d, respectively, with a length of 1.0d. At 20% and 15% rated loads, the recirculation peak and the range of particle volume flux were comparable to those at 40% rated load. The prototype burner demonstrated that, without oil support, the gas temperature within 0 to 1.8 m from the primary air outlet remained below 609 °C, insufficient to ignite faulty coal. As the load rate increased from 20% to 30%, the prototype’s central region temperature remained low, with a maximum of 750 °C between 0 and 2.0 m. In contrast, the DPRB’s central region temperature reached 750 °C at around 0.65–0.70 m. At a 3%·min^‒1 load-up rate, when the load increased from 20% to 30%, the prototype burner extinguished after 30 s. However, the DPRB maintained stable combustion throughout the process.

Graphical abstract

Keywords

swirl burner / gas-particle flow (GPF) characteristics / numerical simulation / deep peak regulation / rapid load-up capability

1 Introduction

In recent years, deep peak shaving for thermal power has become a key research area, especially with the rapid growth of renewable energy. The intermittency and unpredictability of renewable sources pose significant challenges to the stability and reliability of the power system [1]. Flexible retrofit policies now require that the minimum power output reach 35% of the rated load, alongside the ability to make rapid load changes [2]. In China, a significant portion of coal resources consists of faulty coal, such as lean coal, anthracite, or high-ash coal [3,4]. Faulty coal accounts for up 40% of the coal utilized in power plant boilers in China [5]. Currently, the minimum stable combustion load (MSCL) rate for boilers fired with faulty coal is approximately 50%, which cannot meet the requirements of deep peak shaving.

Burners are the core components of boilers, and their stable combustion directly determines the boiler’s capability for deep peak shaving. In domestic coal-fired units with a load of 300 MW or more, over 40% of them use swirling burners [6]. Widely used swirl burners include the low NO_x axial swirl burner (LNASB) [7], enhanced ignition dual-adjustment burner [8], IHI dual-swirl burner [9], HT-NR3 burner [10,11], and central feed burner (CFB) [12]. For environmental and economic reasons, local economic and information committees, along with power generation groups, discourage the long-term use of combustion aids such as micro-oil ignition, plasma-assisted, and oxygen-enriched combustion [13]. Without these aids, boilers using swirl burner struggle to maintain stable combustion at or below 35% rated load when burning faulty coal. Presently, there is no mature swirl combustion technology to address this issue. To achieve effective peak shaving, it is essential to enhance stable combustion characteristics of the burners themselves.

The performance of swirling burners largely depends on the recirculation zone (RZ) [14]. The RZ transports reactive chemical radicals and high temperature byproducts backward, providing ignition energy to the incoming fuel [15]. Several factors, including burner structure, fuel distribution, and swirl intensity, can significantly influence the RZ shape and size [16,17]. Zhao et al. [18] developed a petal swirl burner (PSB), and their numerical simulation indicated that the petal structure aided in RZ formation. In a 210 MW boiler burning lean coal (V_daf = 12%–18%), the MSCL with the PSB was 115.5 MW (55% rated load). The petal structure acted as a bluff body, creating a low-pressure area behind it and forming an RZ. However, this bluff body was continually subjected to the abrasive action of high-speed pulverized coal, making it prone to wear. Based on laboratory cold-state tests, industrial tests, and numerical simulations, Li et al. [12] proposed a CFB, which eliminated the central air and bluff body, instead using swirling secondary air to induce straight-flow primary air, forming a central RZ. Coupled with central feed technology, the CFB achieved stable combustion and low NO_x emissions. Practical application showed that in a 220 MW boiler burning lean coal (V_daf = 21.4%), the MSCL was reduced from 139.9 MW (63.6% rated load) to 103.4 MW (47.0% rated load) after retrofitting [19]. Song et al. [20] performed thermal experiments and numerical simulations to investigate the impact of bluff bodies and swirling secondary airflow rates on recirculation flow, temperature distribution, stable combustion characteristics, and NO_x formation in a novel inverse jet combustion burner (IPCB). Zhou et al. [21] studied gas-particle flow (GPF) characteristics of the HT-NR3 burner using numerical simulation, revealing the formation of an annular RZ adjacent to the primary and secondary air. Su et al. [22] conducted a numerical investigation into the effect of the inner secondary air expansion cone on the aerodynamics and burning performance of swirling burners.

In the present era, mainstream swirl combustion technology primarily applies to early power systems. During that period, thermal power plants typically supplied base loads, maintaining relatively stable output power. The achievable minimum load rate was approximately 50%. However, in the new era of large-scale renewable energy integration, early swirl technology faces challenges in combustion stability. In swirl burners using faulty coal, secondary air velocities are generally lower [23], making it difficult to establish a stable RZ. Consequently, many faulty coal-fired burners adopt primary air rotation or central baffles to compensate. Unfortunately, practical experience demonstrates that this approach can lead to premature coal powder entry into the secondary air, resulting in unstable combustion and elevated NO_x emissions. Retrofitting faulty coal-fired burners thus remains a significant challenge. Huang et al. [24], described a low-load combustion stabilization technology applied to burners in a 700 and 350 MW boilers firing faulty coal, comparing GPF of the different burners and assessing low-load stability. However, achieving deep load adjustments necessitates not only stable combustion at ultra-low loads but also the ability to rapidly adapt to changing loads. Previous research, whether conducted in laboratories or through numerical simulations, has predominantly focused on reducing NO_x emissions under stable boiler conditions at full load, neglecting ultra-low load and load-changing processes. The mismatch between the heat required for ignition and the actual heat supply is the core challenge in ultra-low load and load variation operations. The key to addressing this challenge lies in establishing a stable RZ during these processes.

This study focused on the GPF characteristics of the DPRB in a 700 MW boiler at full, 40%, 20%, and 15% rated loads. It also simulated combustion performance of both new and prototype burners during load changes, with load-up rates of 1% min⁻¹, 2% min⁻¹, and 3%·min⁻¹ as the load rate increased from 20% to 30%. The research improved the analysis of the new burner’s flow characteristics and enriched previous studies on rapid load-up capabilities. This work enhanced the flexibility of boilers burning faulty coal, supported renewable energy integration, and contributed to the transition of the energy structure.

2 Overview of burner

Figure 1 depicts the structure of prototype burner, which was designed by Babcock Germany. For a detailed introduction, please refer to Huang et al. [6]. Figure 2 presents a photograph of prototype burner nozzle taken during the boiler’s maintenance shutdown. The 700 MW boiler, fired faulty coal and equipped with these prototype burners has the following issues: the MSCL rate of 50% during normal operation without oil support, and nozzle deformation.

Figure 3 shows the structure of the DPRB. This design retains the prototype’s secondary air and primary air ducts but removes the central air and primary air blades. It introduces direct primary air and swirling gap secondary air, with the gap air blades angle set at 45°. The swirling gap air, positioned between the primary air and inner secondary air, acts as “mediator,” enhancing entrainment to the coal powder. Additionally, the new and old designed primary air nozzles are optimized with a reasonable pre-mixing section length. The new primary air duct incorporates three-stage coal powder concentration rings to enhance the concentration efficiency. In Fig. 3, x is the distance from the burner outlet, r is the distance from the central axis, and d is the inner diameter of the outlet. The DPRB innovatively uses a secondary air staged entrainment of primary air to create the RZ, eliminating the need for a bluff body. As a result, the RZ shape and size are easier to control. This technology has been granted a domestic patent and is applicable to various types of swirling burners [25].

Table 1 presents main parameters of the DPRB. The parameters reflect the operating conditions of the lower burner. As the load decreases, the flow rates of primary air, secondary air, and coal also decrease, though not in direct proportion to the load. This is because, at low loads, peak load regulation is achieved by deactivating the middle or upper layer burners, while the lower layer burner’s output must remain sufficient to ensure stable combustion. Table 2 provides the coal quality analysis.

3 Small-scale GPF experiments

3.1 Experiment guidelines and system

Based on the gas-particle two-phase modeling criteria, a single-burner gas-particle test platform was established in this study. The criteria for the two-phase modeling are as follows [26,27]:

Geometric similarity: The scale of model burner is 1:9 compared to the actual size.

Reynolds (Re) number criterion: The Re number for the airflow through each model nozzle must exceed 10⁴ to ensure flow similarity.

Fluid mixing criterion: The momentum ratio of the airflow between the model and the full-size burner must be consistent.

Similar boundary conditions: A lengthy straight section is incorporated into the primary air duct. This design guarantees a constant ratio between the velocities of solid and gas phases at the primary air outlet, enabling the solid particles to be effectively entrained and accelerated by the airflow.

Resemblance in the physical characteristics of particles: Due to the insufficient reflectivity of the coal for the experimental requirements, glass particles are used as a substitute because they offer better reflective effects. Both glass and coal particles have densities significantly greater than that of air, and the buoyancy difference between them and air is negligible. Therefore, it is reasonable to assume the physical performance of glass particles to resemble that of coal particles.

Similar Froude number (Fr): The Fr reflects the influence of gravity on fluid flow. Given that the phenomenon under investigation involves forced flow, the gravitational effect on airflow is negligible; hence, the effect of Fr on the experimental outcomes can also be considered negligible.

The GPF characteristics were measured by a 3D laser particle dynamic anemometer (PDA). The details of the test system, measurement errors, and the particle size distribution of tracer particles can be found in Huang et al. [24] and will not be repeated here. A global mass balance calibration of the particle volume flux was conducted.

3.2 Experiment parameters and analytical methods

The 700 MW boiler employs a stratified burner wall arrangement, where the GPF characteristics of the lower burner play a key role in maintaining stable combustion. Based on the data in Table 1, the parameters for the PDA test were calculated according to the modeling criterion, outlined in Section 3.1, as shown in Table 3.

1) The calculation formula of swirl number (Sw):

(1)Sw=\frac{G_{\theta }}{R_b{G}_{\chi }},

where G_θ represents the rotational momentum of gas phase (kg·m·s⁻¹), G_{\theta }={\int }_0^R\rho uw{r}^2\mathrm{d}r; G_χ signifies the axial momentum of gas phase (kg·m·s⁻¹), G_{\chi }={\int }_0^R\rho u^{2}r\mathrm{d}r; R_b denotes the inner diameter of burner outlet (m); ρ is the air density (kg·m⁻³); u and w are the axial velocity and tangential velocity measured by PDA (m·s⁻¹), as depicted in Figs. 6 and 8, respectively. The velocity at x/d = 0.1 is selected for Eq. (1).

2) Definition of RZ: On the x/d cross-sectional plane, the locus of points where the axial velocity is zero is designated as the velocity boundary. To avoid sidewall recirculation, this boundary must be located proximate to the burner central axis. By sequentially connecting these points, a demarcation line is formed. Within this boundary, the axial velocity is assumed negative, whereas outside, it is positive. The region within the demarcation line is defined as the RZ. The maximal length and diameter of the RZ are denoted by L and D, respectively.

3) Expansion angle of air and particles: The absolute magnitude of the axial velocity is averaged, and one-tenth of this average value is defined as the minimal axial velocity. By connecting the corresponding loci on the cross-sectional plane at a distance from the burner center, a demarcation line is formed. The angle between this line and the central axis is defined as the expansion angle, as measured by the PDA.

3.3 GPF experiment result of DPRB

3.3.1 Swirl number and expansion angle

All experimental results presented in this section, as well as in the following text, have been tested for repeatability, demonstrating good reproducibility. Table 3 shows the expansion angles and swirl numbers of the DPRB at different loads. At the same load, the gas expansion angle is slightly larger than the particle expansion angle due to the higher inertia of the particles. As the load decreases, both expansion angle and swirl number first increase, then decrease, with the turning point occurring at 40%–30% load. As the load decreases, both primary and secondary air-flows decrease. The reduction in primary air-flow promotes airflow diffusion, while the reduction in secondary airflow hinders it. The optimal point for both primary and secondary airflows is at the 40%–30% load range.

3.3.2 RZ boundary

Figure 4 shows the RZ boundary under different loads. The boundary lines of the two phases almost overlap, exhibiting the same variation patterns. At full load, a central RZ forms, with a length of 1.5d and a diameter of 0.58d. At load rates of 40%, 20%, and 15%, the RZ becomes annular, with diameters of 0.30d, 0.40d, and 0.39d, respectively, and a length of 1.0d in each case. At x/d = 0.1, the lower boundary of the annular RZ is approximately 30 mm from the central axis. The inner diameter of the primary air duct is 81 mm. This indicates that, even at lower loads, the DPRB can still generate a high-temperature gas recirculation in the primary airflow region, aiding in the early ignition of faulty coal. Additionally, due to the existence of the three-stage concentration ring, the majority of coal powder congregates at the primary air center, with only a minor portion dispersed around the periphery. Although the coal powder in the central area is not within the RZ, the coal powder with a high concentration, which has a lower ignition heat, also contributes to ignition.

In Huang et al. [6,28], the single-phase flow characteristics of the prototype burner and the improved burner were studied. In Huang et al. [24], the GPF characteristics of the prototype burner and the new burner were compared. For comparison between the new burner and the other two burners, please refer to the aforementioned literature; this paper will not reiterate those details.

The recirculation ratio quantitatively describes the stable combustion effect, which is defined as the proportion of the recirculated flow to the primary airflow, and calculated according to Eq. (2).

(2)\mathrm{R}\mathrm{e}\mathrm{c}\mathrm{i}\mathrm{r}\mathrm{c}\mathrm{u}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}=\frac{2\pi }{Q_1}{\int }_{r_1}^{r_2}ur\mathrm{d}r,

where Q₁ represents the flow rate of primary air (m³·s⁻¹); r₁ and r₂ are the starting and ending positions of the RZ (m); u is the axial velocity (m·s⁻¹).

Figure 5 shows the recirculation ratio at different loads. In Fig. 5 and the following text, R denotes the load rate. As shown in Fig. 5(a), the changing trend is similar across all four conditions, with the recirculation ratio initially increasing and then decreasing. The maximum recirculation rate at 40%–15% load is higher than at full load. Recirculation results from the interaction of primary and secondary air. As the load decreases, both the primary and secondary airflows reduce. While the reduction in secondary airflow hinders recirculation, the decrease in primary airflow promotes it. For the DPRB, the positive effect of reducing primary airflow outweighs the negative impact of reducing secondary airflow. Therefore, the maximum recirculation ratio at low loads is higher than at full load.

In addition, the position of the peak recirculation ratio, represented by x/d, increases with increasing load. At full load, the peak recirculation ratio appears at x/d = 0.7; at 40% load, it emerges at x/d = 0.5; and at 15% and 20% loads, it occurs at x/d = 0.3. From Fig. 5(b), the total recirculation ratio at full and 40% loads is significantly higher than at 20% and 15% loads. The maximum total recirculation ratio occurs at 40% load. Although the total recirculation ratio at 20% and 15% loads is lower, their peak recirculation ratio (shown in Fig. 5 (a)) occurs closer to the burner outlet. This indicates that, during actual operation, high-flow gas recirculation exists around x/d = 0.3 from the burner outlet, which supports stable combustion.

3.3.3 Velocity fields analysis

Figure 6 illustrates the average axial velocity at different loads. The velocities of gas phase and particle phase are roughly equivalent, with velocity slip occurring only in certain areas, mainly due to their inertia differences. At full load, the axial velocity exhibits a unimodal distribution with the secondary airflow. The peak value rapidly decays, disappearing at x/d = 2.5. At x/d = 0.3–1.0, the axial velocity near r = 0–60 mm is negative, indicating the presence of a central RZ. As the load decreases, the air mass flow of burner drops, resulting in a different velocity distribution compared to full load. At load rates between 40% and 15%, the axial velocity exhibits a bimodal distribution. The first peak near r = 0 mm is due to the flow of primary air/particle, while the peak near r = 100 mm is caused by the flow of secondary airflow. These two peaks merge into a single peak at x/d = 2.5. At x/d = 0.1–1.5, as the load decreases, the velocity peak of primary air significantly reduces, while the value of secondary air remains relatively constant, which facilitates the formation of the RZ. In the r direction, peak positions of the primary air at 40%, 20%, and 15% load rates remain stable, mainly due to the presence of the concentration ring. As r increases, the primary air velocity rapidly decays, and the values at the three loads become very similar near r = 25 mm. From the perspective of axial velocity, the DPRB establishes a relatively stable RZ at low load, attributed to the consistent peak velocity distribution of secondary air and the rapid decay of primary air velocity.

Figure 7 presents the location of the maximum axial velocity, providing an overall description of the airflow. At full load, the maximum velocity points are all situated within the secondary airflow region. When the load rate ranges between 40% and 15%, the mass flows of both primary and secondary air decrease, causing the burner to operate outside its design parameters. In this case, the airflow splits into two parts: the primary air/particle flow and the secondary airflow. At x/d = 0.1–1.0, the maximum velocity of primary air is concentrated near r = 0 mm. By x/d = 1.5–2.5, r increases as the load decreases.

For secondary air, the peak position varies minimally between x/d = 0.1–1.5. By x/d = 2.5, the overall airflow is dominated by primary air. This same pattern is observed at 40%, 20%, and 15% loads, demonstrating that the GPF of this burner at low load remains stable. Due to inertia, solid particles cluster at the center after colliding with the concentration ring, thereby increasing the axial momentum of the primary air near r = 0 mm. As a result, particle concentration increases in the central area. Even though these particles remain separate from the secondary air, the high concentration of particles with low ignition heat is beneficial for stable combustion.

Figure 8 illustrates the distribution of average tangential velocity at different loads. The tangential velocity exhibits a unimodal distribution due to the swirling secondary air. Since the primary air has the characteristics of direct flow, the tangential velocity within the range of r = 0–40 mm is relatively small. As x/d increases, the rotation gradually weakens, and by x/d = 1.5, the peak tangential velocity almost disappears. The peak values for full to 40% loads are significantly larger than those for 20%–15% loads.

For the primary airflow region, the tangential velocity at 40% load is larger than full, 20%, and 15% load at x/d = 0.1–1.0. Additionally, the starting rotation position of primary air is closer to the center at x/d = 0.5–0.7. The rotation of airflow results from the interaction between the non-swirling primary air and the swirling secondary air, making the 40% load more conducive for their synergy.

The data in Table 1 also show that as the load decreases, the swirl number and expansion angle initially rise and then decrease, with the maximum values appearing at 40% load. This nonlinear relationship explains why the expansion angle after 40% load is similar to the full-load condition. A larger expansion angle facilitates better diffusion, which is beneficial for the formation of a RZ.

Figure 9 shows the average radial velocity at different loads. The radial velocity exhibits a unimodal distribution across all conditions, driven by the outward diffusion of the secondary air. A distinct trough near r = 100 mm appears at x/d = 0.1–0.3, where the radial velocity becomes negative. As the airflow progresses, this negative velocity region gradually shifts toward the wall, with the trough moving to r = 200 mm at x/d = 1.5. The presence of negative radial velocity indicates a tendency for particles to move toward the center.

In addition, the velocity distribution at 20% and 15% loads closely resembles that at full load and 40% load at x/d = 0.1–0.7. This suggests that, at ultra-low loads, the radial velocity distribution of the DPRB is similar to that at high loads.

The turbulence intensity, denoted as TI, which characterizes the impact of turbulence fluctuations and variables on flow stability, is calculated from the three-dimensional average velocity and the three-dimensional fluctuation velocity. The formula for calculating TI is

(3)TI=\frac{1}{U}\sqrt{\frac{{u_f}^2+{v_f}^2+{w_f}^2}{3}},

where u_f, v_f, and w_f denote the fluctuation values of axial, radial, and tangential velocities, which are directly measured by the PDA, with an uncertainty of 1%; and U is the resultant velocity of the three-dimensional average velocity.

Turbulent kinetic energy, denoted as k (m²·s⁻²), can be used to analyze the storage and transfer of energy within a fluid, whose formula is

(4)k=\frac{1}{2}\left(u_f^2+v_f^2+w_f^2\right).

The turbulent energy dissipation rate, named as ε (m²·s⁻³), reveals the rate at which k is transformed into internal energy, whose formula is given by Wang et al. [29] as

(5)\epsilon =\frac{1}{2}\mu \overline{{\left(\frac{\mathrm{\partial }{u}_f}{\mathrm{\partial }{c}_i}+\frac{\mathrm{\partial }{v}_f}{\mathrm{\partial }{c}_j}+\frac{\mathrm{\partial }{w}_f}{\mathrm{\partial }{c}_k}\right)}^2},

where μ is the kinematic viscosity (m²/s); \mathrm{\partial }{u}_f/\mathrm{\partial }{c}_i, \mathrm{\partial }{v}_f/\mathrm{\partial }{c}_j, and \mathrm{\partial }{w}_f/\mathrm{\partial }{c}_k are the partial derivatives of the fluctuating velocities with respect to the coordinates, in which i, j, and k correspond to the spatial coordinate directions. In this study, the velocity at the same x/d cross-section is only related to the radial direction.

Figure 10 illustrates the TI at different loads. At full load, a central RZ forms, resulting in a distinct TI distribution with a bimodal characteristic. The peak near r = 10 mm corresponds to the lower boundary of RZ, while the peak near r = 60 mm aligns with the upper boundary. Turbulent activity is most intense near the RZ boundaries. As x/d reaches 0.7, these two peaks merge into one located near r = 80 mm, marking the point where the primary and secondary air converge.

Under other load conditions, the TI distribution remains unimodal, with peak positions close to their respective RZ boundaries. When x/d reaches 1.0, the decay of the airflow weakens turbulent activity, leading to a reduction in TI. Although the peaks at 20% and 15% loads are farther from the center, the positions of increased intensity are the same as at 40% load, both located near r = 10 mm. Moreover, the peak values at 20% and 15% loads are higher than at 40% load, indicating that the DPRB can still ensure sufficient mixing of airflow at ultra-low loads, which is key to stable combustion.

Figure 11 is the turbulent kinetic energy (k) at different loads. For non-isotropic turbulence, where u_f ≠ v_f ≠ w_f, there is no clear linear relationship between k and TI. Therefore, the distribution of k differs from that of TI. In the primary airflow region (r = 0–40 mm), the energy is relatively higher at 40% load, and lower at 20% and 15% loads. Between r = 75–100 mm, the energy is significantly higher at full load. Notably, as k dissipates, by x/d = 1.5–2.5, the TI markedly decreases.

Figure 12 represents the dissipation rate of turbulent kinetic energy (ε) at different loads, which plays a crucial role in turbulence energy loss. The dissipation rates under different conditions each have their respective peaks which are located within the RZ. This indicates that within the RZ, the rate of turbulence energy conversion to internal energy is relatively higher, and material exchange is more intense.

Except for full load, during the initial phase of jet development (x/d = 0.1–0.5), the dissipation rate decreases as the load increases. At full load, however, at x/d = 0.3–0.5, the dissipation rate has a maximum near r = 0 mm, which is precisely located in the central RZ, leading to the intense energy dissipation. From a microscopic perspective, the DPRB can still ensure sufficient energy exchange even under ultra-low load conditions, which is very beneficial for stable combustion.

3.3.4 Particle volume flux analysis

Figure 13 represents the particle volume flux at different loads, which signifies the volume of particles traversing one square meter per second. The burner forms a central RZ at full load, leading to a relatively low volume flux with a larger absolute value of negative flux and a broader recirculation range. For non-full load conditions, the volume flux peaks within the r = 0–10 mm range, with most particles congregating at the burner center.

The primary airflow of DPRB is non-swirling, and it features a three-stage concentration ring. As the coal powder carried by primary air collides with the ring, it gathers at the center, forming a dense distribution in the central area of the burner. In addition, at x/d = 0.3–0.7, the particle volume flux peaks at 20% and 15% loads, along with the recirculation range closely resembling that at 40% load. At 40% and 20% loads, the particle volume flux is higher near r = 0–40 mm at the x/d = 0.7–1.5 cross-section compared to other cases.

As the load decreases, both primary and secondary air velocities drop. Lower primary air velocity helps particles become entrained by the secondary air, but the lower secondary air velocity has the opposite effect. Due to their higher inertia, solid particles are more difficult to entrain into the secondary air. As a result, at these loads, the swirling secondary air cannot sufficiently entrain particles in the center, causing them to accumulate and resulting in a higher particle volume flux.

The positions of the peaks and valleys in the volume flux reflect the trajectory of particles. Comparing changes in these positions provides valuable insights into the solid-phase movement. Figure 14 illustrates these positions at different loads. The valleys at 15% loads and 20% loads are closely aligned, while at 40% load, they fall between those at full load and those at 20%–15% loads. Generally, the valleys across different conditions concentrate around r = 40–70 mm.

Peak positions show minimal variation across conditions, primarily falling within the r = 0–20 mm range at x/d = 0.1–1.0. This suggests that at ultra-low loads (20%–15% load), the trajectory of particle remains similar to that at high loads, further confirming the wide load adaptability of the DPRB.

4 Cold start-up process of prototype burner

The 700 MW boiler features burners arranged on both the front and rear walls, with each wall hosting three layers of burners. The schematic diagram of boiler is shown in Fig. 15. Detailed parameters can be found in Huang et al. [6]. During the cold startup process, the lower layer burners are ignited simultaneously, followed by the sequential activation of the middle layer burners. The load discussed below refers to the boiler load. At the initial stage of startup, the boiler is not connected to grid, so the load is zero. This part of the study focuses on the prototype burner located on the lower layer of the rear wall, with the cold startup process divided into five stages. The main parameters for these stages are listed in Table 4. The whole cold start-up process lasts for 12 h.

A thermocouple is positioned at the burner monitoring port (refer to Fig. 1) to measure the gas temperature. The reference point for measurement is where the monitoring port axis intersects with the outlet plane of primary air duct. This thermocouple is made of nickel-chromium and nickel-silicon, capable of measuring temperatures ranging from −50 to 1300 °C with an accuracy of ±1 °C. The experimental results from Soete [30] showed that the deviation between the measured and actual temperatures inside the furnace, when using a thermocouple, did not exceed 8%. The ignition temperature ranges from 754 to 1090 °C for lean coal and anthracite [31,32], and 570 °C for bituminous coal [33]. For comparative analysis, this study assumes that the coal ignites at 750 °C.

Figure 16 illustrates the gas temperature at different stages of the prototype burner. The horizontal axis represents the distance, denoted as D_p, from the measurement point to the primary air outlet. During stages 1–4, the prototype persistently employs oil to assist combustion, with a mass flow of 1.0 t·h⁻¹. Initially, with only oil, the gas temperature reaches 1000 °C at a distance of 0.6 m. Notably, in stages 2–4, the burner transitions to a stage of mixed combustion involving both coal and oil. Within 0–1.8 m range, the maximum temperature remains below 750 °C. At 376 MW (stage 5), when the oil injection is stopped, the maximum temperature only reaches 609 °C. This observation indicates that the coal dust within the 0 to 1.8 m range did not ignite. The prototype burner demonstrates insufficient combustion stability without oil assistance, as even under oil-assisted combustion, the temperatures did not reach the ignition point of coal dust.

During the process of increasing the boiler load from low to high, the combustion characteristics of the burner at low load and the dynamic load-up process are key areas of study. In the cold start-up process, the furnace temperature is relatively low, and the conditions resemble those of ultra-low load. The data obtained during this phase provide valuable insights into the burner’s combustion characteristics. To further investigate the burner’s rapid load-up capability, numerical simulations are required. The data obtained during the cold start-up process can serve as fundamental input for these subsequent numerical simulations.

5 Numerical calculation of burner climbing performance

5.1 Mesh subdivision

The computational domain is depicted in Fig. 17, with the burner dimensions modeled based on the actual site dimensions. Since each burner can independently operate in real processes, the furnace is simulated as a cuboid with dimensions of 8.5 m (height) × 11.2 m (width) × 15.6 m (length). These dimensions reflect the actual spacing and width between the burners.

5.2 Calculation method and mathematical model

Major equations of the mathematical models used in numerical simulations:

(1) The realizable k-ɛ model [34] is chosen for airflow turbulence calculations

(6)\begin{array}{rl}\frac{\mathrm{\partial }\left(\rho k\right)}{\mathrm{\partial }t}+\frac{\mathrm{\partial }\left(\rho k{u}_i\right)}{\mathrm{\partial }{x}_i}=& \frac{\mathrm{\partial }}{\mathrm{\partial }{x}_j}\left[\left(\mu +\frac{{\mu }_t}{{\sigma }_k}\right)\frac{\mathrm{\partial }k}{\mathrm{\partial }{x}_j}\right]+\\ & G_k+G_b-\rho \epsilon -Y_{\mathrm{m}}+S_k,\end{array}

(7)\begin{array}{rl}\frac{\mathrm{\partial }\left(\rho \epsilon \right)}{\mathrm{\partial }t}+\frac{\mathrm{\partial }\left(\rho \epsilon u_i\right)}{\mathrm{\partial }{x}_i}=& \frac{\mathrm{\partial }}{\mathrm{\partial }{x}_j}\left[\left(\mu +\frac{{\mu }_t}{{\sigma }_{\epsilon }}\right)\frac{\mathrm{\partial }\epsilon }{\mathrm{\partial }{x}_j}\right]+\rho C_{l\epsilon }{S}_{\epsilon }-\\ & \rho C_{2\epsilon }\frac{{\epsilon }^2}{k+\sqrt{\nu \epsilon }}+C_{l\epsilon }\frac{\epsilon }{k}{C}_{3\epsilon }{G}_b+S_{\epsilon },\end{array}

where k and ɛ is turbulent pulsation kinetic energy and its dissipation rate (m²·s⁻³), respectively; μ_t is turbulent viscosity (kg·m⁻¹·s⁻¹); σ_k and σ_ɛ are Planck’s constant for k and ɛ equations, respectively; G_k and G_b are turbulent kinetic energy generated by laminar velocity and buoyancy, respectively (m²·s⁻²); Y_m is fluctuations from transition diffusion in compressible turbulence (m²·s⁻²); S_k and S_ɛ are customised source terms; C_1ɛ, C_2ɛ, and C_3ɛ are empirical constants; and ν is kinematic viscosity (kg·m⁻¹·s⁻¹).

(2) The turbulent dispersion of the discrete phase is calculated using the Lagrangian stochastic trajectory model [35]

(8)\frac{\mathrm{d}{\mathit{u}}_{\mathrm{p}}}{\mathrm{d}t}=\frac{18\mu }{{\rho }_{\mathrm{p}}{d}_{\mathrm{p}}^2}\frac{C_{\mathrm{D}}Re}{24}\left({\mathit{u}}_{\mathrm{g}}-{\mathit{u}}_{\mathrm{p}}\right)+\mathit{g}\frac{\left({\rho }_{\mathrm{p}}-\rho \right)}{{\rho }_{\mathrm{p}}},

where u_g and u_p are the gas and particle velocity, respectively (m·s); ρ and ρ_p are the gas and particle density, respectively (kg·m⁻³); d_p is the particle diameter (m); C_D is the drag coefficient; g is the acceleration of gravity; and Re is the Reynolds number.

(3) Radiation heat transfer is computed using the P1 model [36]

(9)q_r=-\frac{l}{3\left(\alpha +{\sigma }_{\mathrm{s}}\right)-C{\sigma }_{\mathrm{s}}}\mathrm{\nabla }G,

where q_r is radiation heat flux; α and σ_s are absorption and scattering coefficients (m⁻¹); C is linear anisotropic scattering phase function; and G is the projected radiance.

(4) The two-step competition model is used for volatiles precipitation [37]

(10)\begin{array}{rl}& \frac{m_{\nu }(t)}{(1-f_{\mathrm{w},0})m_{\mathrm{p},0}-m_a}=\\ & {\int }_0^t({\alpha }_1{R}_1+{\alpha }_2{R}_2)\exp \left(-{\int }_0^t(R_1+R_2)\mathrm{d}t\right)\mathrm{d}t,\end{array}

(11)R_1=A_1\exp \left[-\left(E_1/R{T}_{\mathrm{p}}\right)\right],

(12)R_2=A_2\exp \left[-\left(E_2/R{T}_{\mathrm{p}}\right)\right],

where f_w,0 is the proportion of initial moisture in particles; T_p is the particle temperature (K); R₁ and R₂ are the primary and secondary reaction rates, respectively (s⁻¹); m_a is the ash mass in particles (kg); m_p,0 is the initial particle mass (kg); A₁ is the pre-exponential factor of the primary reaction (s⁻¹); A₂ is the former factor of the secondary reaction (s⁻¹); E₁ is the activation energy of the primary reaction (J·mol⁻¹); and E₂ is the activation energy of the secondary reaction (J·mol⁻¹).

(5) The kinetic-diffusion model [38] is used for char combustion process

(13)\frac{\mathrm{d}{m}_{\mathrm{p}}}{\mathrm{d}t}=\pi R_{\mathrm{t}}{p}_{\mathrm{o}\mathrm{x}}{d}_{\mathrm{p}}^2,

where m_p is the mass of coke particles (kg); R_t is the total reaction rate constant of coke combustion (kg·m⁻²·s⁻¹·atm⁻¹); and p_ox is the oxygen partial pressure around the particles (atm).

(6) Activation energy and pre-exponential factors are determined based on Zhou et al. [39]. During the pyrolysis of pulverized coal, for volatile release in Reaction I, A₁ = 3.75 × 10⁵ s⁻¹, E₁ = 7.366 × 10⁴ J·mol⁻¹; for volatile release in Reaction II, A₂ = 1.46 × 10¹³ s⁻¹, E₂ = 2.511 × 10⁵ J·mol⁻¹. In the coke combustion reaction, the pre-exponential factor is 0.0016 (kg·m⁻²·s⁻¹·atm⁻¹), and the activation energy is 8.37 × 10⁴ J·mol⁻¹.

5.3 Boundary condition

This section simulates the load variation process during low-load operation, comparing the stability of the prototype and the DPRBs at different load variation rates. The simulation results indicate that the burner will not extinguish due to excessive load reduction rates. However, during load increase, an excessively high load-up rate can cause the burner to extinguish, especially in the initial stages of low-load operation. Therefore, this section focuses on the process of increasing the boiler load rate from 20% to 30%, specifically studying the lower-layer burners.

For the steady-state calculation at 20% rated load, the burner inlets are set as mass flow boundaries based on actual operating parameters, with TI set at 10%. The basic parameters are listed in Table 5. Inlet temperatures match the actual operating air temperatures. The slip factor between coal particle velocity and air velocity is 0.8, and the wall temperature is set to 1000 °C. The coal analysis is based on the actual coal used onsite, as shown in Table 2. The distribution of coal fineness is shown in Fig. 18, with an average particle size of 26.9 μm.

As shown in Table 4, the load-up rates range from 0.24%·min−1 to 1.06%·min⁻¹. To explore the rapid load-up capability of the burner before and after modification, the stable combustion condition of at 20% load is used as the baseline. The load is then increased to 30% at load-up rates of 1% min⁻¹, 2% min⁻¹, and 3%·min⁻¹, as shown in Table 5. The variation in airflow and coal feed over time to achieve the load-up process is implemented using user defined functions (UDF).

5.4 Numerical simulation results

5.4.1 Grid independence verification

To verify grid independence, a single burner model is tested using 1.46 million, 1.64 million, and 2.20 million grids. The simulation conditions are set for the lower prototype burner operating at 63.4% of the boiler’s rated load. At identical parameters, the velocity and temperature distributions along the burner’s central axis are compared, as illustrated in Fig. 19. D_p is the length from the data point to the outlet plane of primary air duct. The models with 1.64 million and 2.20 million grids show similar velocity and temperature distributions, whereas the 1.46 million grid model has significant deviations in some regions. Based on these results, the model with 1.64 million grids is selected for subsequent numerical calculations.

5.4.2 Mathematical model validating

Figure 20 presents the comparison between the temperature data from industrial tests and numerical calculations. The industrial test data, sourced from Huang et al. [6], correspond to the gas temperature of the lower prototype burner at 63.4% of the boiler’s rated load with 60% central air. The parameters such as air and coal mass flow used in the numerical simulation are identical to those of industrial test. As shown in Fig. 20, the numerical simulation results closely align with the measured values obtained from the burner monitoring port and central axis. Table 6 shows that the simulated temperature rise rate in the central region is 50.98 °C·m⁻¹, compared to the measured value of 48.42 °C·m⁻¹, with a relative error of 5.02%. In the monitoring port region, the simulated temperature rise rate is 158.45 °C·m⁻¹, while the measured value is 156.56 °C·m⁻¹, with a relative error of 1.19%. Assuming coal ignition occurs at a gas temperature of 750 °C (as detailed in Section 4), the numerical simulation predicts the ignition position of 2.58 m, while the measured value is 2.47 m, resulting in a relative error of 4.26%. With all maximum relative errors under 5%, the numerical simulation model demonstrates high reliability.

5.4.3 Numerical results analysis

The primary focus of this study is on the load-up process. Therefore, the combustion characteristics of the burner at different loads are not analyzed in detail. Figures 21–23 show the evolution of the temperature field in the prototype burner over time at load-up rates of 1% min⁻¹, 2% min⁻¹, and 3%·min⁻¹, respectively. The cross-section is taken at the center of furnace height. It can be seen that the prototype burner has a strong flame rigidity, with slow radial diffusion. The high-temperature flame starts near the annular primary air outlet, which may lead to nozzle overheating and deformation if the burner operates for prolonged periods. At load-up rates of 1%·min⁻¹ and 2%·min⁻¹, the flame shape and temperature remain significantly over time. However, at a load-up rate of 3%·min⁻¹, the flame temperature significantly decreases within 30 s and eventually extinguishes as time progresses. Therefore, the maximum time in Fig. 23 is only intercepted to 60 s.

Figure 24 illustrates the temperature variation over time in the center and monitoring port regions of the prototype burner at different load-up rates. Across all three rates, the temperature in central region remains low, with the highest temperature within the 0–2 m range not exceeding 750 °C. It indicates that coal in this region struggles to ignite. In the monitoring port region, temperatures generally remain lower at all rates. At a rate of 1%·min⁻¹, the temperature is initially low at 1 s and gradually increases over time. For a rate of 2%·min⁻¹, the low temperature period occurs at 25 s, while for 3%·min⁻¹, it occurs at 5 s. This indicates that as the load increases, the prototype burner faces a risk of flame blowout, with significant temperature fluctuations over time. Notably, at a rate of 3%·min⁻¹, flame extinction occurs between 40 and 60 s.

Figures 25–27 show the changes in temperature field in the DPRB over time at load-up rates of 1%·min⁻¹, 2%·min⁻¹, and 3%·min⁻¹, respectively. The high-temperature gas (above 1300 °C) from the DPRB fills the furnace more thoroughly compared to the prototype burner, indicating better flame coverage. In actual operation, it will enhance furnace temperature stability, supporting sustained combustion. Moreover, the DPRB’s flame starts farther from the nozzle, effectively preventing overheating and slagging issues. Notably, at a load-up rate of 3%·min⁻¹, the DPRB does not experience exhibit flameout.

Figure 28 illustrates the temperature variation over time in the center and monitoring port regions of the DPRB at different load-up rates. At a rate of 1%·min⁻¹, the center temperature begins to rise after 0.5 m, peaking at 1200 °C. As the load-up rate rises, the rate of temperature increase in the center region slows down. The result shows that the initial heating rate is higher than in the later stages. This slower rise is primarily due to the increasing primary airflow over time, which enhances the axial momentum and further delays the mixing of pulverized coal and secondary air. Overall, the center region temperature of the DPRB remains higher than that of the prototype burner.

In the monitoring port region, the temperature distribution across different load-up rates is similar. The temperature rapidly increases after 0.5 m, reaching 750 °C between 0.65 to 0.70 m. It then peaks near 1.0 m before stabilizing, with the highest temperature concentrated around 1200 °C. In contrast, the temperature at the prototype burner monitoring port fluctuates significantly over time, and the ignition position of the pulverized coal (at 750 °C) is unstable. The highest temperature mostly stays below 1200 °C. Therefore, in comparison to the prototype burner, the DPRB demonstrates more stable flames and superior rapid load-up capability.

The maximum load-up rate of the prototype burner is 3%·min⁻¹. To compare the velocity differences, three time points at this rate are analyzed, as shown in Fig. 29. The velocities shown are the three-dimensional combined velocities, which differ from the axial velocity observed in the cold-state results. As time progresses, the high-speed primary airflow of the prototype burner restricts radial diffusion and prevents the flame from expanding. In contrast, the new burner has higher velocity regions in both the primary and secondary air main flow zones. The primary airflow decays more quickly, while the secondary airflow spreads rapidly, forming a low-velocity zone between them. This low-velocity zone facilitates the timely ignition of the coal.

Due to the effects of gas expansion and chemical reactions under hot-state conditions, the velocity field in the hot state cannot perfectly match the velocity measured in the cold-state tests. However, the low-velocity zone in the hot state is closely related to the RZ in the cold state, with both starting near the burner exit.

6 Conclusions

The MSCL without oil support for faulty coal-fired boilers is approximately 50%, which is higher than the 35%‒30% load rate required for deep peak shaving. To address this issue, this study develops a DPRB technology. Through GP experiments, industrial tests, and unsteady numerical simulations, the flexibility of the DPRB is comprehensively evaluated. The results are as follows:

1) Broad load adaptability of DPRB: The prototype has no RZ. At full boiler load, the DPRB outlet features a central RZ, with measured recirculation lengths and diameters of 1.5d and 0.58d, respectively. As the load decreases to 40%, 20%, and 15%, the RZ transitions to an annular shape, with diameters of 0.30d, 0.40d, and 0.39d, respectively, maintaining a consistent length of 1.0d. At x/d = 0.1, the lower boundary of the annular RZ is located in the primary airflow region, only about 30 mm from the burner central axis. As the load decreases, the maximum recirculation ratio gradually approaches the nozzle.

2) Gas-particle flow characteristics: At full load, the axial velocity of the DPRB exhibits a single peak distribution, primarily influenced by the secondary airflow. As the load decreases from 40% to 15%, the peak velocity of primary air significantly reduces. The tangential velocity peaks at full to 40% loads are considerably higher than those at 20%–15% loads. As the load decreases, both the swirl number and the expansion angle initially increase, peaking at 40% load, then decrease. The expansion angles at 15%–20% loads are similar to those at full load, but the TI peaks are higher than at 40% load. The particle volume flux peaks at center at full to 15% loads. At x/d = 0.3–0.7, the flux peaks and range of recirculation at 20%–15% loads are similar to those at 40% load.

3) Simulation versus industrial test results: The difference between the simulation and industrial test results is less than 5.02%. During the load-up process from 20% to 30%, the DPRB shows better flexibility and stability compared to the prototype burner. The flame origin of the prototype burner is located near the circular primary air outlet, while the flame origin of the DPRB is positioned away from the secondary air nozzle, protected by both primary and secondary air. At load-up rates of 1%·min⁻¹, 2%·min⁻¹, and 3%·min⁻¹, the prototype’s central region temperature remains low, with a maximum of 750 °C within 0–2.0 m. In contrast, the DPRB’s central region reaches 750 °C at around 0.65–0.70 m. At a 3%·min⁻¹ load-up rate, the prototype burner extinguishes after 30 s, whereas the DPRB maintains stable combustion. The center temperature of the DPRB is higher than that of the prototype, and the temperature in the monitoring port region fluctuates less over time.

4) Future directions: Future efforts will focus on optimizing burner structure and operation parameters to reduce NO_x emissions and improve combustion stability. Numerical simulations of the entire furnace will be conducted, followed by industrial trials. The numerical simulation model will be further refined based on industrial testing data.

References

[1] Gu Y J , Xu J , Chen D C . . Overall review of peak shaving for coal-fired power units in China. Renewable & Sustainable Energy Reviews, 2016, 54: 723–731

[2] National Energy Administrator . National coal power unit transformation and upgrading implementation plan. 2021-10-29, available at website of the State Council of the People’s Republic of China. , ,

[3] Jun X , Sun X X , Hu S . . An experimental research on boiler combustion performance. Fuel Processing Technology, 2000, 68(2): 139–151

[4] Xiumin J , Chuguang Z , Che Y . . Physical structure and combustion properties of super fine pulverized coal particle. Fuel, 2002, 81(6): 793–797

[5] Bai X F , Ding H , Lian J J . . Coal production in China: Past, present, and future projections. International Geology Review, 2018, 60(5–6): 535–547

[6] Huang C C , Li Z Q , Wang Y F . . Influence of central air on flow and combustion characteristics and low–load stabilization performance of a Babcock burner. Processes, 2023, 11(7): 1916

[7] Zeng L Y , Li Z Q , Zhao G B . . Effect of the vane angle for outer secondary air on the flow and combustion characteristics and NO_x emissions of the low-NO_x axial–swirl coal burner. Numerical Heat Transfer Part A, 2011, 59(1): 43–57

[8] King J L . Low NO_x pulverized fuel burners: Summary of plant experience. In: American Society of Mechanical Engineers, New York, USA, 1995,

[9] You C F , Zhou Y . Effect of operation parameters on the slagging near swirl coal burner throat. Energy & Fuels, 2006, 20(5): 1855–1861

[10] Ochi K , Kiyama K , Yoshizako H . . Latest low-NO_x combustion technology for pulverized-coal-fired boilers. Hitachi Review, 2009, 58(5): 187–193

[11] Liu J Q , Sun B M , Zhang G C . . Numerical simulation and optimization on stable combustion of a 1000 MW ultra-supercritical unit swirl combustion boiler. Proceedings of CSEE, 2012, 08: 19–27 (in Chinese)

[12] Li Z Q , Chen Z C , Sun R . . New low NO_x, low grade coal fired swirl stabilised technology. Journal of the Energy Institute, 2007, 80(3): 123–130

[13] Zhang G C , Zhou K , Lu F . . Discussions on deep peaking technology of coal-fired power plants. Thermal Power Generation, 2014, 46(9): 17–23 (in Chinese)

[14] Sung Y , Lee S , Eom S . . Optical non-intrusive measurements of internal recirculation zone of pulverized coal swirling flames with secondary swirl intensity. Energy, 2016, 103: 61–74

[15] Olivani A , Solero G , Cozzi F . . Near field flow structure of isothermal swirling flows and reacting non-premixed swirling flames. Experimental Thermal and Fluid Science, 2007, 31(5): 427–436

[16] Zhen H S , Leung C W , Cheung C S . A comparison of the thermal, emission and heat transfer characteristics of swirl-stabilized premixed and inverse diffusion flames. Energy Conversion and Management, 2011, 52(2): 1263–1271

[17] Garde R J . Turbulent Flow. 2nd ed. New Delhi: New Age International (P) Limited Publishers, 2000,

[18] Zhao L L , Zhou Q T , Zhao C S . Flame characteristics in a novel petal swirl burner. Combustion and Flame, 2008, 155(1–2): 277–288

[19] Li Z Q , Chen Z C , Sun R . . Central–fuel–rich swirl coal combustion technology in 1025 t/h lean coal fired boiler. Journal of Mechanical Engineering, 2006, 42(3): 221–226

[20] Song M H , Huang Q , Niu F . . Recirculating structures and combustion characteristics in a reverse–jet swirl pulverized coal burner. Fuel, 2020, 270: 117456

[21] Zhou C Y , Wang Y Q , Jin Q Y . . Mechanism analysis on the pulverized coal combustion flame stability and NO_x emission in a swirl burner with deep air staging. Journal of the Energy Institute, 2019, 92(2): 298–310

[22] Su X Q , Fang Q Y , Ma L . . Improving combustion and lowering NO_x emissions of an industrial coal swirl burner by optimizing its nozzle structure. Applied Thermal Engineering, 2018, 5: 119340

[23] He P A , Zhao Z H , Qin Y K . Design and Operation of Pulverized Coal Burner. Beijing: China Machine Press, 1987, 338–339 (in Chinese)

[24] Huang C C , Li Z Q , Lu Y . . Research on a novel universal low-load stable combustion technology. Energy, 2024, 310: 133223

[25] Li Z Q , Huang C C , Wang Y F . . A swirl pulverized coal burner with adjustable recirculation zone and its application method (China Patent CN202311319641.9). Beijing: China National Intellectual Property Administration, 2023,

[26] Yan R , Chen Z C , Guan S . . Influence of mass air flow ratio on gas particle flow characteristics of a swirl burner in a 29 MW pulverized coal boiler. Frontiers in Energy, 2021, 15(1): 68–77

[27] Wang Z Q , Hu Y J , Cheng X X . . Influence of offset angle of mid-secondary air nozzles on gas-particle flow characteristics in a furnace. RSC Advances, 2018, 8(32): 17764–17772

[28] Huang C C , Li Z Q , Wang Y F . . Influence of secondary air blade angle and oxygen-rich combustion characteristics of an improved Babcock swirl burner. Applied Thermal Engineering, 2024, 241: 122383

[29] Wang G , Yang F , Wu K . . Estimation of the dissipation rate of turbulent kinetic energy: A review. Chemical Engineering Science, 2021, 229: 116133

[30] Soete G G . Measurement of flame temperature with a multi-element thermocouple. Journal of the Energy Institute, 1981, 54(419): 113–116

[31] Xie J Y , Chen H W . Distribution of ignition temperature and ignition heat of power coal in China. CIESC Journal, 2015, 66(10): 4170–4176

[32] Li X G , Zeng L Y , Liu H Y . . Industrial-scale investigations on effects of tertiary-air declination angle on combustion and steam temperature characteristics in a 350-MW supercritical down-fired boiler. Frontiers in Energy, 2021, 15(1): 132–142

[33] Guo L , Zhai M , Wang Z T . . Comparison of bituminous coal and lignite during combustion: combustion performance, coking and slagging characteristics. Journal of the Energy Institute, 2019, 92(3): 802–812

[34] Launder B E , Spalding D B . The numerical computation of turbulent flows. Computer Methods in Applied Mechanics and Engineering, 1974, 3(2): 269–289

[35] Zhou H , Yang Y , Liu H Z . . Numerical simulation of the combustion characteristics of a low NO_x swirl burner: Influence of the primary air pipe. Fuel, 2014, 130: 168–176

[36] Yin C G . On gas and particle radiation in pulverized fuel combustion furnaces. Applied Energy, 2015, 157: 554–561

[37] Ma L , Fang Q Y , Lv D Z . . Reducing NO_x emissions for a 600 MWe down-fired pulverized-coal utility boiler by applying a novel combustion system. Environmental Science & Technology, 2015, 49(21): 13040–13049

[38] Field M A . Rate of combustion of size-graded fractions of char from a low-rank coal between 1200 K and 2000 K. Combustion and Flame, 1969, 13(3): 237–252

[39] Zhou C , Zhang S F , Wen L Y . . Anthracite combustion kinetics study by thermal analysis. Journal of China Coal Society, 2011, 36(8): 1370–1374 (in Chinese)

RIGHTS & PERMISSIONS

Higher Education Press 2025