Numerical Simulation of the Combustion Characteristics of a 330 MW Tangentially Fired Boiler with Preheating Combustion Devices Under Various Loads

Abstract

With the rapid development of renewable energy sources in power generation, utility boilers need to perform load regulation over a wide range to maintain the stability of the power supply system. Preheating combustion technology is a potential approach to achieve wide load range operation, improve combustion stability, and lower NO_x emissions from utility boilers. Preheating combustion devices (PCDs) were designed and installed in the reduction zone of a boiler. These devices preheated the coal at an excess air ratio ranging from 0.35 to 0.7 to generate high-temperature gas and char, which effectively reduced NO_x formation in the furnace. Numerical studies were conducted to evaluate the combustion performance and nitrogen oxides emissions of a 330 MW utility boiler retrofitted with PCDs at different loads. The simulations were conducted over a load range of 20% to 100% of the rated load, corresponding to an electrical power of 66 MW to 330 MW. The preheated combustion device’s previous experimental data served as the boundary conditions of the preheated product nozzles. The simulation results demonstrated that the retrofitted boiler could operate stably from 20% to 100% of the rated load, maintaining acceptable combustion efficiency and lower NO_x emissions. The combustion efficiency gradually dropped with decreasing boiler load, reaching a minimum value of 95.6%. As the load declined, the size of the imaginary tangent circle of the boiler shrank, while the ignition distance increased. Additionally, the variation in NO_x concentration with load was complex. The NO_x concentration at the furnace outlet was between 102.7 and 220.3 mg/m³, and the preheated products effectively reduced the nitrogen oxides produced by combustion in the furnace at all loads.

1. Introduction

To facilitate the low-carbon transition of the power system, the proportion of renewable energy generation in the power grid is gradually increasing. However, the instability of renewable energy generation poses a severe challenge to the stable operation of the power grid [1,2]. Coal-fired power plants, as a fundamental energy source of the power grid, need to innovate their operating modes to adapt to the evolving energy situation. Coal-fired boilers achieve optimal efficiency when operating at loads exceeding 70% of their rated capacity. The rapid growth of renewable energy generation has caused fluctuations in power grid load. To accommodate these fluctuations, a coal-fired boiler’s load must be adjusted over a wide range, with the lowest load reaching 30% of the rated load or even lower. Operating the boiler at low loads significantly deviates from optimal conditions, resulting in operational risks during load regulation [3]. The primary issues for coal-fired boilers operating at low loads are ignition delay [4], combustion instability [5], and raised NO_x emissions [6]. Given the context, it is vital to explore the combustion performance and pollutant emissions of utility boilers operating under various loads.

Researchers have conducted experimental studies and numerical simulations on the combustion performance and pollutant emissions of coal-fired boilers operating at various loads. Ma et al. [4] conducted experimental and numerical studies on the combustion stability of a 300 MWe tangentially fired boiler under a wide load range. The results indicated that as the load decreased, the ignition of the pulverized coal stream was delayed, and the imaginary tangential circle gradually diminished. Compared with the rated load, the ignition distance increased from 2.7 m to 5.4 m under low-load conditions. Jiang et al. [7] optimized operating conditions for a 550 MW utility boiler to stabilize combustion and reduce NO_x formation at half-load. Their study examined the effects of burner configuration, overfire air distribution, and the stoichiometric ratio on combustion performance. The simulations indicated that a middle burner configuration and a specific close-coupled overfire air (CCOFA)/separated overfire air (SOFA) distribution maintained efficient and clean combustion. Chang et al. [8] conducted numerical simulations on the flow field, temperature field, and nitrogen oxide formation in a coal-fired boiler under various loads. The results revealed that the reduced load impaired air–fuel reaction and combustion stability, leading to increased O₂ and NO_x contents and decreased CO content. The tilt angles of burners played a significant role in combustion performance and heat transfer. Li et al. [9] performed investigations on boiler efficiency and species distribution of a retrofitted opposite-wall-fired boiler under various loads. The results indicated that flue gas temperature decreased with decreasing boiler load, while the ignition distance increased. After the retrofit, the NO_x concentration at the furnace exit significantly dropped under the same load. Gu et al. [10] evaluated the effects of boiler load, burner configuration, and burner tilt angles on combustion characteristics of a 660 MW utility boiler. They concluded that the temperature and NO concentration decreased with the decrease in boiler load. The optimal combustion performance was achieved by using ABDE burners under the boiler maximum continuous rating (BMCR) scenario.

Previous research on low-load boiler operation has primarily focused on optimizing boiler combustion characteristics by adjusting air distribution and burner arrangement. Preheating combustion, by preheating the fuel to enhance its reactivity and adjusting the distribution of air and fuel within the furnace, strengthens the combustion stability of the boiler. This pulverized coal preheating technology is a pulverized coal pretreatment method that can achieve stable and clean combustion at various boiler loads [11,12]. In the preheating device, pulverized coal is initially ignited and preheated, with an excess air coefficient far less than 1 [13]. Preheated products, consisting of high-temperature reductive gases and coke, are introduced into the furnace chamber, where they ignite rapidly to ensure stable and efficient combustion [14]. Additionally, the reductive gases can reduce nitrogen oxide generation in the burner area of the furnace [15]. Currently, this preheating technology is widely acknowledged as a feasible, efficient, and clean combustion technology. Tang [16] conducted experiments on the combustion characteristics of a burner with a preheating function in a test boiler. A detailed analysis was conducted on the temperature and species distribution in the burner. The results showed that the preheating temperature remained stably above 700 °C, significantly improving the combustion efficiency and reducing NO formation from the boiler. At a thermal load of 25 MW, the NO_x emission at the boiler outlet was 47.9 mg/Nm³. Hui [17] explored the combustion characteristics of coal-preheating combustion using a circulating fluidized bed during load regulation. The results indicated that the preheating treatment enhanced the reactivity of pulverized coal and ensured stable combustion of the system. In numerical studies on preheating technology, Hong [18] performed numerical simulations to study the effects of a new type of burner on the combustion characteristics of a 660 MW utility boiler under low loads. Self-sustaining combustion could be realized within the burner. The results showed that the retrofitted burners improved the rigidity of jet flow, thus enhancing combustion stability. The reductive gas from the burners inhibited the NO generation in the furnace, resulting in low NO_x emissions. Yao [19] performed simulations to study a 600 MW utility boiler with preheating burners. The results showed that the preheating products, such as reductive gas and char, enhanced the combustion stability of the boiler at 50% rated load. During the preheating combustion in the burners, the fuel nitrogen was converted into N₂ with a stoichiometric ratio of less than 1.

Research on preheating combustion has primarily concentrated on laboratory-scale investigations of NO_x formation and reduction during the preheating stage. In previous studies [20], experiments and numerical simulations have verified that preheating combustion devices can reduce NOx emissions from boilers under rated load. However, the impacts of the preheating combustion on the combustion characteristics of coal-fired boilers at various loads have not been evaluated, resulting in a limitation of efficient and clean combustion in industrial applications. In this study, numerical investigations are conducted to evaluate the combustion performance and NO_x emissions of a coal-fired boiler at a range of 20–100% rated load. The retrofitting strategy is to connect the preheating combustion devices’ (PCDs’) outlets directly to the nozzles of the preheated products within the reduction zone. A three-dimensional CFD model was employed using experimentally measured PCD gas properties as inlet conditions. This simulation compared the combustion stability, temperature, and species distributions of the boiler under different loads. The effects of the PCDs on the boiler’s combustion efficiency and NO_x emissions are discussed in detail. The simulation results demonstrate that the retrofitted boiler can realize lower NO_x emissions and higher combustion efficiency at a load range of 66 MW to 330 MW, which provides insight into efficient and clean combustion at various loads.

2. Methodology

2.1. Boiler Description

A PCD test system was established and experimentally investigated in our previous work, and the details about the device, including experimental setup and procedure, are shown in the literature [20]. The operating principle of preheating combustion technology is illustrated in Figure 1. The pulverized coal is ignited inside the PCD at a stoichiometric ratio (SR) of less than 1. The pulverized coal is primarily converted into reductive gases and char through volatile pyrolysis and char gasification reactions. After the boiler retrofit, the furnace chamber is divided into three regions: the primary combustion zone, the reduction zone, and the burnout zone. The primary combustion zone is the region where most of the coal is burned with air, releasing the majority of heat. In the reduction zone, reductive gases and char are injected through the nozzles to mitigate NO formed during primary combustion. Furthermore, efficient combustion can be achieved through the reaction of air with unburned carbon and gases in the burnout zone.

A 330 MW coal-fired retrofitted boiler is used in this study. The retrofit involves integrating PCDs with the boiler by installing preheated product nozzles at each corner of the furnace chamber. Figure 2 presents the retrofitted boiler’s operating mode, burner arrangement, and dimensions. Figure 2a illustrates the boiler’s burner configuration after the retrofit. The burner zone consists of the bottom-three-layer burners (A, B, and C) and top-two-layer burners (D and E), along with the surrounding secondary air (SA) nozzles. The FF-layer nozzles are set as the CCOFA nozzles, which are located at the top of the primary combustion zone. The reduction zone is composed of preheated product nozzles located 2.23 m above the FF-layer nozzles. Four layers of SOFA nozzles are arranged at the top of the furnace chamber to ensure complete combustion. Figure 2b shows the direction of jet flow and cross-section dimensions. The pulverized coal stream and the secondary air are injected in the same direction and form an imaginary circle at the center of the furnace. During boiler operation, the fuel for the PCDs is supplied by a standby coal mill.

2.2. Case Setup and Coal Properties

Numerical simulations were conducted for boiler loads ranging from 20 to 100% rated load, corresponding to 66–330 MW of electric power. To ensure combustion stability and efficiency under various loads, the flow rate of the preheated coal varies almost proportionally with the load.

Table 1. Boiler operating parameters at various loads.

Case Number	Case 1	Case 2	Case 3	Case 4	Case 5
Load (MW)	330	231	165	99	66
Running burners	ABCD	ABCD	ABC	AB	AB
Total coal feed rate (kg/s)	52.22	36.94	26.39	16.39	10.83
Preheating coal feed rate (kg/s)	11.11	8.33	5.56	2.78	2.78
Preheating air mass-flow (kg/s)	27.78	20.83	13.89	6.94	6.94
Primary air (PA) mass-flow (kg/s)	118.17	82.72	59.08	35.44	23.64
Secondary air (SA) mass-flow (kg/s)	190.61	133.44	95.31	57.19	38.11
SOFA mass-flow (kg/s)	102.92	72.06	51.47	30.89	20.58
PA temperature (K)	343	343	343	343	343
SA temperature (K)	615	589	567	537	537
Boiler outlet temperature (K)	643	621	592	554	492

presents the operational parameters and boundary conditions of the retrofitted boiler. The boiler outlet pressure is −80 Pa. Cases 1–5 investigate the effects of various boiler loads on the boiler’s combustion performance and NO formation after the retrofit. Additionally, the influence of preheated products on the gas components, temperature distribution, and combustion stability is discussed in detail. Considering the PCD, experimental studies were performed with a stoichiometric ratio (SR) of 0.42 and the preheating coal amount ranging from 2.78 kg/s to 11.11 kg/s.

Table 2. Boundary conditions of the preheated products’ nozzles.

Feed Rate of Preheated Coal (kg/s)	SR	CO (%)	CO2 (%)	CH4 (%)	H2 (%)	N2 (%)	Flow Rate of Char (kg/s)	Flow Rate of Gas (kg/s)	Temperature of Gas (K)
2.78	0.42	11.68	13.20	0.68	0.65	73.79	0.96	8.77	926
5.56	0.42	12.33	12.74	0.80	0.68	73.45	1.89	17.55	1006
8.33	0.42	12.71	12.41	0.96	0.72	73.20	2.81	26.36	1121
11.11	0.42	13.16	12.01	1.13	0.78	72.92	3.73	35.16	1236

summarizes the composition (wt. %, dry basis), flow rate, and temperature of the preheated products at the PCD outlet, according to the literature [20]. The gaseous composition is presented as volume percentages measured on a dry basis. These parameters are also used as the inputs for the boundary conditions of the preheated products’ nozzles in this simulation.

The heat exchangers at the top of the boiler are simplified as zero-thickness walls [21]. For thermal boundary conditions, wall temperatures are defined as the arithmetic mean of steam inlet/outlet temperatures, while the wall emissivity is prescribed as 0.7 according to the literature [7]. The same bituminous coal is used in both the boiler and PCD.

Table 3. Coal properties.

Ultimate Analysis (Dry Basis, wt. %)						Vdaf (wt. %)	Qar, net (kJ/kg)
C	H	O	N	S	A	38.63	17,220
63.77	3.62	15.38	0.94	0.6	15.69

presents the bituminous coal properties derived from proximate and ultimate analysis. The particle diameter distribution ranges from 6 μm to 230 μm. The Rosin-Rammler distribution function is used to describe the diameter distribution of the pulverized coal particles, with a mean diameter of 69 μm and a spread parameter of 1.1.

2.3. Numerical Models

The governing equations involved in the combustion process are solved using Ansys Fluent(Version 2020 R2) (Canonsburg, United States). The SIMPLE algorithm couples pressure–velocity fields when discretizing the RANS conservation equations for mass, momentum, and energy. Since the pulverized coal combustion involves heat, mass, and momentum transfer, the Eulerian–Lagrangian approach is employed for gas–solid flow simulation [22]. The discrete phase model (DPM) is employed to track the motion of pulverized coal particles [23]. The motion of a single particle is governed by multiple-force interactions, and the force balance equation describing the particle’s motion is expressed as

$${\displaystyle \frac{d{\overrightarrow{u}}_{\mathrm{p}}}{dt}}=F_D(\overrightarrow{u}-{\overrightarrow{u}}_{\mathrm{p}})+\overrightarrow{g}{\displaystyle \frac{({\rho }_p-\rho )}{{\rho }_{\mathrm{p}}}}-{\displaystyle \frac{\nabla p}{{\rho }_{\mathrm{p}}}}$$

(1)

The turbulent flow of the gas phase is simulated by the Realizable k-ε model [24]. The model performs well in calculating complex gas-phase flow, and the transport equations of the model are given as

$${\displaystyle \frac{\partial (\rho k{u}_i)}{\partial x_j}}={\displaystyle \frac{\partial }{\partial x_j}}\left[\left(\mu +{\displaystyle \frac{{\mu }_t}{{\sigma }_k}}\right){\displaystyle \frac{\partial k}{\partial x_j}}\right]+G_k+G_b-\rho \epsilon$$

(2)

$${\displaystyle \frac{\partial (\rho \epsilon u_i)}{\partial x_j}}={\displaystyle \frac{\partial }{\partial x_j}}\left[\left(\mu +{\displaystyle \frac{{\mu }_t}{{\sigma }_{\epsilon }}}\right){\displaystyle \frac{\partial \epsilon }{\partial x_j}}\right]+\rho C_{1}S\epsilon -\rho C_2{\displaystyle \frac{{\epsilon }^2}{k+\sqrt{\nu \epsilon }}}+C_{1\epsilon }{\displaystyle \frac{\epsilon }{k}}{C}_{3\epsilon }{G}_b$$

(3)

In Equation (3), $C_1=\max \left[0.43,{\displaystyle \frac{\eta }{\eta +5}}\right]$ with $\eta =S{\displaystyle \frac{k}{\epsilon }}$, $S=\sqrt{2S_{ij}{S}_{ij}}$ and $S_{ij}={\displaystyle \frac{1}{2}}({\displaystyle \frac{\partial u_j}{\partial x_i}}+{\displaystyle \frac{\partial u_i}{\partial x_j}})$; C₂, C_1ε, and C_3ε are constants [25].

The discrete ordinates (DO) model has been widely applied in radiative heat transfer calculations due to its good computational accuracy. Therefore, radiative heat transfer is solved using the DO model. The gas-phase absorption coefficient is calculated with the weighted-sum-of-gray-gases model, and the particle emissivity is set to 0.9 [26].

The moisture evaporation is considered during coal combustion, and the moisture evaporates at a temperature of 373.15 K [27]. The two-competing-rates model [28] describes the volatile release of pulverized coal. The model considers devolatilization reactions at both high and low temperatures, and the primary and secondary kinetic rates are expressed as

$$R_1=A_1 {\exp }^{(-E_1/R{T}_p)}$$

(4)

$$R_2=A_2 {\exp }^{(-E_2/R{T}_p)}$$

(5)

In the equations above, the parameters for coal devolatilization and combustion are obtained from the literature [29] due to the similar coal properties. Therefore, A₁ and E₁ for the primary reaction rate are set to 2 × 105 s⁻¹ and 1.046 × 108 J/kmol, respectively, while A₂ and E₂ for the secondary reaction rate are set to 1.3 × 107 s⁻¹ and 1.674 × 108 J/kmol.

The devolatilization rate is derived from the weighted combination of the primary and secondary reaction rates, which is given as

$${\displaystyle \frac{m_{\nu }(t)}{\left(1-f_{w,0}\right)m_{p,0}-m_a}}={\int }_0^t\left({\alpha }_1{R}_1+{\alpha }_2{R}_2\right)\exp (-{\int }_0^t(R_1+R_2)dt)dt$$

(6)

The homogeneous combustion rate is obtained by the eddy-dissipation model [30]. Based on the coal analysis data in Table 3, the volatiles are defined as a pseudo-compound (C_3.08H_10.4O_2.95N_0.184S_0.0818). Additionally, the combustion mechanism involves two reaction steps: (1) oxidation of volatile carbon to CO; (2) further oxidation of CO to CO₂. The detailed reaction steps are given as follows [31]:

$${\mathrm{C}}_{3.08}{\mathrm{H}}_{10.4}{\mathrm{O}}_{2.95}{\mathrm{N}}_{0.184}{\mathrm{S}}_{0.0818}{+\mathrm{O}}_2\to 3.08\mathrm{CO}+5{.2\mathrm{H}}_2\mathrm{O}+0{.0937\mathrm{N}}_2+0{.0818\mathrm{SO}}_2$$

(7)

$${\mathrm{CO}+\mathrm{O}}_2\to {\mathrm{CO}}_2$$

(8)

The kinetic/diffusion-limited model is used to calculate the reaction rate of char combustion [32]. The model assumes that the surface reaction rate of char particles is determined by the competition between the kinetic rate and diffusion rate. The diffusion rate coefficient D₀ and the kinetic rate of char particles R_p are given as

$$D_0=C{\displaystyle \frac{{\left[(T_p+T_{\infty })/2\right]}^{0.75}}{d_p}}$$

(9)

$$R_p=A_p{\exp }^{(-E_p/R{T}_p)}$$

(10)

The char combustion rate is determined by weighting the kinetic rate and the diffusion rate coefficient, which is given as follows [27]:

$${\displaystyle \frac{d{m}_p}{dt}}=-\pi d_p^2{p}_{ox}{\displaystyle \frac{D_0{R}_p}{D_0+R_p}}$$

(11)

A post-processing method is chosen to calculate the NO_x formation from coal combustion. The simulation models both thermal and fuel NO_x generation mechanisms during the reaction of nitric oxide. Thermal NO_x is generated via the reaction of nitrogen and oxygen inside the furnace at high temperature, and the Zeldvich mechanism [33] is used to describe this process. Thermal NO_x formation mechanisms are dominated by the following reactions:

$$\mathrm{O}+{\mathrm{N}}_2\underset{k_{r1}}{\overset{k_{f1}}{\rightleftharpoons }}\mathrm{N}+\mathrm{NO}$$

(12)

$$\mathrm{N}+{\mathrm{O}}_2\underset{k_{r2}}{\overset{k_{f2}}{\rightleftharpoons }}\mathrm{NO}+\mathrm{O}$$

(13)

$$\mathrm{N}+\mathrm{OH}\underset{k_{r3}}{\overset{k_{f3}}{\rightleftharpoons }}\mathrm{H}+\mathrm{NO}$$

(14)

In addition, the NO production rate is given below:

$${\displaystyle \frac{d[\mathrm{NO}]}{dt}} = 2 k_{f1}{\left[\mathrm{O}\right][\mathrm{N}}_2] {\displaystyle \frac{\left(1-\frac{k_{r1}{k}_{r2}{\left[\mathrm{NO}\right]}^2}{k_{f1}{[\mathrm{N}}_2] k_{f2}{[\mathrm{O}}_2]}\right)}{1+\frac{k_{r1}\left[\mathrm{NO}\right]}{k_{f2}{[\mathrm{O}}_2]+k_{f3}\left[\mathrm{OH}\right]}}}$$

(15)

Oxidation of fuel-bound nitrogen generates fuel NO_x, and the De Soete model [34] is used to calculate the formation rate of fuel NO_x. Figure 3 exhibits the detailed reaction paths for the fuel-bound nitrogen. During coal combustion, the fuel-bound nitrogen is released as volatile nitrogen by pyrolysis, while the remainder is retained as char nitrogen. During the fuel nitrogen reaction, the volatile N initially converts to intermediates such as HCN and NH₃, with an HCN/NH₃ ratio of 9/1 [35]. Within oxygen-rich zones, NO_x precursors (HCN/NH₃) are oxidized to NO, while these intermediates reduce NO to N₂ in oxygen-lean zones. Meanwhile, char N is completely converted into NO. The presence of reductive gases in the preheated products allows the fuel reburning model to be activated, with CH₄ and H₂ serving as the fuel species.

2.4. Mesh Independence and Model Verification

The structured mesh is established based on a simplified boiler structure using the ICEM preprocessor. To minimize the influence of outlet backflow on the accuracy of the in-furnace calculation, the computational domain extends to the boiler rear pass. Figure 4 shows the mesh of the boiler computational domain and furnace cross-sectional mesh. To ensure that the computational results are not influenced by the mesh size, mesh independence verification is performed on meshes with 1.90, 2.36, 2.89, and 3.34 million cells, respectively. Figure 5 compares the cross-section average gas temperature and O₂ concentration profiles along the furnace height for the four meshes. The simulated results demonstrate no significant deviation among the temperature curves when the mesh number exceeds 1.90 million. The average relative error for the oxygen concentration and temperature profiles between the 2.36-million-cell mesh and the finest mesh is less than 2%. Therefore, a 2.36-million-cell mesh is employed to ensure accuracy and computational efficiency in this simulation.

Table 4. Summary of mesh quality.

Quality Specification	Range	Average Value	Average Standard Value [36]
Equi-angle skewness (QEAS)	0–0.81	0.36	≤0.4
Aspect ratio (QAR)	1–50	4.65	≤5

shows the mesh quality specification. The results show that the average values of Q_EAS and Q_AR are smaller than the average standard value, indicating a high-quality 3D mesh [36].

During the field test, the composition of flue gas was measured at the entrance of the selective catalytic reduction reactor, which is situated at the rear pass of the boiler. The computational accuracy is validated through comparison of the simulated values with the field test data. The field tests were conducted on both the original boiler and the retrofitted boiler at full load.

Table 5. Comparison of measured data and calculated values.

Item	330 MW (Original)			330 MW (Retrofitted)
	Measured	Calculated	Error	Measured	Calculated	Error
O2 (%)	2.98	3.16	6.0%	2.12	2.04	3.8%
NOx (mg/m3, at 6% O2)	185.7	194.5	4.7%	126.7	121.1	4.4%
Gas temperature at the furnace outlet (K)	1279 (design value)	1200.1	6.1%	/	1254.9	/

illustrates the verification results between the calculated results and test values, with the deviation less than 6.1%, which proves that the numerical models and the related parameter settings are reasonable.

Table 6. Convergence criteria.

Parameters	Criteria	Parameters	Criteria	Parameters	Criteria
Continuity	10−6	ε	10−6	CO	10−4
vx	10−6	Energy	10−6	CO2	10−4
vy	10−6	DO	10−6	H2	10−4
vz	10−6	H2O	10−6	CH4	10−4
k	10−6	O2	10−4	vol	10−4

shows the convergence criteria of the numerical simulation.

3. Results and Discussion

3.1. Effects of Load on Combustion Performance of the Boiler

3.1.1. Temperature Distribution

Figure 6 demonstrates the mean temperature distribution in the furnace under various loads. Since the coal feed rate rises with increasing load, the average temperature within the furnace rises significantly as the boiler load increases. In the 66 MW load scenario, the coal feed rate is minimal, drastically reducing the combustion rate and leading to a considerably lower furnace temperature than that in other operational scenarios. At 66 MW and 99 MW, only the bottom burners (A and B) are in operation. As a result, heat release is concentrated in the primary combustion zone, leading to a significant temperature increase in this area. In other operating scenarios, with the upper burners in operation, the continuous supply of fuel and air to the downstream flue gas area results in temperature rises in both the primary combustion zone and the reduction zone. In the reduction zone, preheated combustion products react with residual oxygen from the primary combustion zone, releasing heat through exothermic reactions. However, due to the low combustion temperature in low-load conditions, the heat release from the flue gas to the furnace walls is greater than the heat generated by combustion, leading to a continuous temperature decrease in the reduction zone. The flue gas temperature in the burnout zone rises slightly due to the heat released from the reaction of combustibles with excess air. Additionally, the 66 MW scenario exhibits a much smaller temperature increase compared to the others due to its lower combustion temperature. Since the burner operating mode is identical in both cases, the average temperature distributions at 99 MW and 66 MW are similar to each other. The 165 MW case employs a three-layer burner operating mode. Compared with the 99 MW and 66 MW cases, the three-layer burner operation enables continuous supplementation of fuel and air to the downstream flue gas regions, resulting in a sustained temperature rise in the primary combustion zone and enhanced heat release in the furnace. Exothermic reactions occur between excess oxygen and preheated products in the reduction zone. However, the exothermic process is limited by low oxygen concentrations, causing a gradual temperature decrease. It is worth noting that the decline rate of temperature is slower than that observed in the 99 MW and 66 MW cases. Four-layer burner configurations in both the 231 MW and 330 MW cases exhibit similar temperature distribution trends in the furnace. These burner configurations enable multiple fuel and air injections along the furnace height. In the primary combustion zone, the temperature curves for both cases exhibit an upward trend with fluctuations. The oxygen concentration of flue gas in the reduction zone increases with the injection of the CCOFA. The excess oxygen reacts with the highly reactive preheated products (CO, H₂, CH₄, and char) in the reduction zone, resulting in notable temperature rises observed in the two curves.

3.1.2. Combustion Stability and Efficiency Under Various Loads

To evaluate the combustion stability of the boiler at different loads, Figure 7 compares the temperature distributions in the direction of the pulverized coal stream for the A- and B-layer burners. The ignition distance is a crucial parameter for assessing combustion stability, and the location of the ignition point of the pulverized coal stream is determined using Semenov’s thermal ignition theory. According to the theory, the location of the ignition point is defined as

$${\displaystyle \frac{\mathrm{d}T}{\mathrm{d}x}}⩾0,{\displaystyle \frac{{\mathrm{d}}^{2}T}{\mathrm{d}{x}^2}}=0$$

(16)

In Equation (16), x represents the distance between the pulverized coal stream and the burner nozzle. Correspondingly, the ignition point is identified as the inflection point in the first derivative of the temperature profile. In Figure 7, the x-coordinates corresponding to dashed lines denote the ignition positions, revealing that the ignition distance significantly decreases with the increase in boiler load.

Table 7. Ignition distances under different loads.

	66 MW	99 MW	165 MW	231 MW	330 MW
Ignition distance of burner A/(m)	5.68	3.79	2.61	2.21	2.18
Ignition distance of burner B/(m)	3.89	2.93	2.49	2.17	2.19

shows the values of the ignition distance under different loads. Additionally, the temperature distributions at 330 MW and 231 MW are similar, indicating that under high-load scenarios, with the same burner operating mode, boiler load has minimal effect on the ignition distance. Figure 7b shows that the ignition distance of the B-layer burner is generally less than that of the A-layer burner. The pulverized coal stream receives more radiant heat with the upward flow of flue gas. The increase in flow rate of flue gas intensifies the disturbance inside the furnace, strengthens the mixing of flue gas and pulverized coal gas stream, and is conducive to heating the pulverized coal. As a result, the ignition distance of the upper burner is shorter, and the combustion stability is enhanced.

To comprehensively evaluate the combustion characteristics under various loads, Figure 8 presents the cross-sectional temperature contours of the A and B burners. As the load decreases, the size of the imaginary circles inside the furnace gradually diminishes, resulting in a decrease in the high-temperature area. In addition, the imaginary circles at 66 MW and 99 MW deviate from the geometric center of the furnace, leading to an uneven temperature distribution.

The ignition distance increases as boiler load decreases, which shortens the combustion time and reduces the heat released during combustion. A decline in the flow rate of PA weakens the penetration of the pulverized coal stream into the flue gas, causing a reduction in the flame spreading speed. Furthermore, when the load drops from 330 MW to 66 MW, the flow rate of SA decreases from 190.61 kg/s to 38.11 kg/s. This reduction in SA weakens the interaction between adjacent airflows, which prevents the expansion of the imaginary circle created by the collision of neighboring gas streams. Consequently, the heating effect of the upstream gas on the downstream gas is diminished. As a result, the gas streams from the four corners of a burner layer become more independent, leading to unstable combustion.

Figure 9 exhibits the boiler’s combustion efficiency and the furnace outlet temperature at different loads. The combustion efficiency is expressed as follows [37]:

$${\eta }_{\mathrm{r}}=\left[100-q_3-q_4\right]\%$$

(17)

The combustible gas heat loss q₃ and the unburnt char heat loss q₄ are given by the following expressions:

$$q_3={\displaystyle \frac{V_{\mathrm{gy}}\times 123.36\mathrm{CO}}{Q_{ar,net}}}\times 100\%$$

(18)

$$q_4={\displaystyle \frac{C_f}{1-C_f}}\times {\displaystyle \frac{33727{\mathrm{A}}_{ar}}{Q_{ar,net}}}\times 100\%$$

(19)

In Equation (18), Q_{ar, net} is the low heating value, kJ/kg, and CO means the volumetric fraction of carbon monoxide in flue gas. The boiler’s combustion efficiency and the furnace outlet temperature decrease with a decline in boiler load. As the load decreases, the total coal feed rate of the boiler decreases proportionally, lowering the combustion temperature. The lower combustion temperature results in unstable combustion in the furnace, increasing the ignition distance and thereby reducing combustion efficiency. However, since part of the fuel and air are preheated in the PCDs, the coal feed rate and air flow rate are reduced in the primary combustion zone. This extends the reaction time of pulverized coal in this zone, resulting in more complete combustion of fuel at lower loads. Furthermore, the fuel-rich combustion in the PCDs produces highly reactive char and gas. The preheated products are more likely to react with the oxygen in the burnout zone, improving combustion efficiency. As a result, although the lower-load operation reduces the boiler efficiency, the minimum combustion efficiency remains above 95%.

3.2. Effects of Loads on the O₂ and NO_x Distribution of the Boiler

3.2.1. O₂ Distribution Under Various Loads

Figure 10 demonstrates the mean O₂ distribution in the furnace at varying loads. The boiler employs air-staging technology, with an SR of 0.9 at the reduction zone outlet for all operating conditions. However, due to variations in the preheating coal feeding rate within the PCDs, the stoichiometric ratio in the primary combustion zone also varies under different operating conditions. The stoichiometric ratio significantly influences the combustion and pollutant formation in the furnace, and the local mean stoichiometric ratio (LMSR) [9] is introduced to accurately compare the stoichiometric ratios in the primary combustion zone for different cases. The LMSR is the ratio of the actual air flow rate in the primary combustion zone to the theoretical air flow rate required for complete combustion of pulverized coal. The theoretical air flow rate is calculated from the analysis data and the feed rate of pulverized coal, while the actual air flow rate is the sum of the primary and secondary air flow rates in the burner region. Figure 11 shows the LMSRs of the primary combustion zone for all cases. In the 66 MW case, the highest LMSR in the primary combustion zone leads to an excessive oxygen supply, resulting in the highest oxygen concentration but simultaneously causing low combustion temperature and slow combustion rate. In the reduction zone, residual oxygen is consumed by reacting with preheated products. However, the low temperature hinders the combustion of pulverized coal, resulting in a high oxygen concentration in the reduction zone. As the boiler load decreases, both the coal feed rate and the average temperature decline, resulting in a slow combustion rate in the furnace. Therefore, at the bottom of the primary combustion zone, the oxygen concentration at low loads is higher than that at high loads. Since the 99 MW and 165 MW boilers operate without the top burners D and E, there is no air supply during the upward flow of flue gas, and the oxygen concentration decreases rapidly. Moreover, as the boiler load decreases, the flue gas velocity correspondingly decreases. The reduced velocity significantly increases the residence time of flue gas in the furnace, which in turn promotes more complete reactions between combustibles and oxygen. Moreover, the injected preheated products further decrease the oxygen concentration. Under the 330 MW and 231 MW scenarios, the bottom burners A, B, and C, the upper burner D, and the top CCOFA nozzles FF are in operation. The identical air distribution and burner configuration lead to a similar oxygen concentration distribution in both cases. Given that burner nozzles C and D, as well as D and FF, are separated from each other, the oxygen concentration in the flue gas exhibits an initial increase followed by a subsequent decrease as it passes through the D-layer burners and the CCOFA nozzles. In the reduction zone, the injected preheated products, which consist of highly reactive combustible gas and char, react intensely with the residual oxygen in the flue gas, leading to a rapid decrease in oxygen concentration.

3.2.2. NO_x Distribution Under Various Loads

Figure 12 presents the comparison of the cross-section average NO_x concentration along the furnace height under various loads. Figure 13 shows the NO_x distributions on the furnace’s vertical section at different loads. The results indicate that the peak NO_x concentration is located in the primary combustion zone. Low-load conditions lead to combustion instability and consequently result in a non-uniform NOx concentration distribution. As only the thermal and fuel NO_x are considered in this study, the formation of NO_x in the furnace depends on the combustion temperature and oxygen concentration. In the primary combustion zone, the 330 MW scenario exhibits a lower LMSR compared with the 231 MW scenario, resulting in a lower NO_x concentration. Under low-load conditions, the complexity of NO_x distribution is influenced by the oxygen distribution and burner configuration. The 66 MW scenario demonstrates a high oxygen concentration and the lowest combustion rate at the bottom of the burner zone, which leads to a lower NO_x concentration compared to the 99 MW and 165 MW scenarios during the early stage of combustion. With the upward flow of furnace gases, the temperature in the 66 MW case increases while maintaining a high oxygen concentration, leading to a rise in NO_x concentration. In the reduction zone, the flue gas at 66 MW exhibits the highest oxygen concentration, which inhibits the reduction of NO_x by the preheated products. In the 99 MW scenario, only the bottom two layers of burners are in operation, and the LMSR is the lowest among the cases. The NO_x concentration progressively declines, owing to the low oxygen concentration in both the primary combustion zone and the reduction zone. Compared to the 66 MW and 99 MW boilers, the 165 MW boiler maintains a significantly higher temperature and high oxygen concentration in the primary combustion zone, resulting in the highest NO_x generation in this zone among the low-load cases.

Figure 14 presents the oxygen and NO_x concentrations at the furnace exit at different loads. As boiler load decreases, the NO_x concentration at the furnace outlet exhibits a fluctuating pattern. The formation of NO_x in the furnace is influenced by the preheated products, load, and burner arrangement. Compared to the 330 MW case, the 231 MW case exhibits a higher proportion of preheated fuel compared to the total fuel and a lower combustion temperature. Both factors inhibit the generation of NO_x, resulting in a lower NO_x concentration at 231 MW. At 165 MW, the furnace adopts a three-layer bottom burner configuration. The reduction in load results in a significant drop in combustion temperature in the primary combustion zone. A lower combustion temperature and a higher oxygen concentration weaken the reduction of the preheated products on the NO_x in the reduction zone, leading to a higher NO_x concentration at the outlet of the furnace. The 99 MW boiler operates at a low combustion temperature with high oxygen concentration during the initial stage of combustion. It employs a two-layer burner arrangement, which extends the fuel reaction time in the burner region. Moreover, the LMSR of the 99 MW boiler in the primary combustion zone is the lowest. A low LMSR creates a strongly reducing atmosphere, which effectively suppresses the conversion of fuel nitrogen to NO_x by favoring the formation of molecular nitrogen (N₂) over nitrogen oxides. In the reduction zone, the 99 MW boiler exhibits the lowest oxygen concentration, enabling effective NO_x reduction after the preheated products mix with the furnace gas, which ultimately leads to the lowest NO_x concentration at the furnace outlet. In the 66 MW scenario, the ignition delay and the highest LMSR result in the highest oxygen concentration in the furnace for all operating conditions. Therefore, the high oxygen concentration inhibits the reduction of NO_x by the preheated products in the reduction zone, resulting in the highest NO_x concentration at the furnace outlet. Under the same excess air coefficient, a higher combustion efficiency means that more pulverized coal and oxygen are consumed. Consequently, the oxygen concentration at the furnace outlet decreases with increasing boiler load.

4. Conclusions

Numerical investigations were conducted on a 330 MW coal-fired boiler retrofitted with PCDs. The combustion characteristics and NO_x emissions of the retrofitted boiler were studied at 20–100% rated load. This simulation compares the combustion stability, temperature, and species distributions of the boiler under different loads. Key findings are summarized as follows.

(1)

The boiler with the PCDs can achieve stable operation with an acceptable combustion efficiency over a load range of 66 MW to 330 MW. As the load declines, the temperature distribution in the furnace becomes uneven, while the ignition distance of pulverized coal increases.

(2)

As the load decreases from 330 MW to 66 MW, the boiler’s combustion efficiency gradually declines from 98.7% to 95.6%. The PCDs help achieve complete combustion. Under the 66 MW scenario, boiler combustion stability markedly decreases, accompanied by a significant rise in NOx emissions. Therefore, the PCDs should not be engaged at a boiler load of 66 MW.

(3)

With fluctuations in boiler load, the NOx concentration at the furnace outlet varies from 102.7 mg/Nm3 to 220.3 mg/Nm3. The preheated products effectively reduce NOx generated during the combustion process under most loads, thereby lowering NOx emissions at the boiler outlet.

In future work, it is necessary to investigate the integration of preheating technology with burner arrangement and air distribution, aiming to optimize combustion stability and NO_x emissions during boiler low-load operation.