An Experimental Study on the Combustion Characteristics of a Methane Diffusion Flame within a Confined Space under Sub-Atmospheric Pressure

Abstract

Gas-fired boilers, gas stoves, and wall-mounted gas boilers are the main consumers of gas fuel, but they generally encounter problems when operating at high altitudes, such as reduced thermal efficiency and increased pollutant emissions. Previous studies on gas combustion characteristics under sub-atmospheric pressure were mostly carried out in a large space, which is quite different from chamber combustion equipment. Therefore, it is insufficient to guide the design and operation optimization of plateau gas equipment. In this paper, experimentations were carried out to explore the characteristics of a methane diffusion flame under sub-atmospheric pressures. The mass flow rates of methane and air remain consistent under different pressure conditions. The centerline temperature ($T_c$) distribution, flame appearance, smoke point, CO emission, and NO_x emission under different pressures (ranging from 61.66 to 97.75 kPa) were examined under both fuel rich and lean conditions. The results show that $T_c$ at the rear and front of furnace variation with pressure is opposite under fuel-lean and -rich combustion. The $T_c$ at the front of furnace decreases with decreasing pressure, whereas $T_c$ at the rear of furnace increases with decreasing pressure. With decreasing pressure, flame length decreases under lean combustion, but increases under rich combustion. The smoke point fuel flow rate, flame length, and residence time increases with decreasing pressure, following the law of negative exponent. The CO emission decreases with decreasing pressure, which indicates that the reduced pressure makes methane combustion more complete. For NO emission, the reduced pressure results in an opposite tendency under fuel-lean and -rich combustion. With decreasing pressure, the NO emission decreases under fuel-lean combustion but increases under fuel-rich combustion.

1. Introduction

In recent years, to protect the ecological environment and reduce carbon emissions at high altitudes, the proportion of natural gas used has increased year by year [1]. However, as the main utilization equipment of natural gas, gas-fired boilers, gas stoves, and wall-mounted gas boilers generally encounter problems such as reduced thermal efficiency and increased pollutant emissions when operating at high-altitude areas [2,3,4,5]. This is mainly due to the environmental characteristics of low atmospheric pressure and low oxygen mass concentration at high-altitude areas, which reduce the burning reaction rate and the flame radiation ability. In addition, reports on methane combustion characteristics are mainly under atmospheric pressure and high pressure (0.1~6 MPa) [6,7]. It was demonstrated that pressure influence on heat release rate and chemical reaction rate is essentially non-linear [8]. It is not trivial to scale information gathered from measurements at atmospheric and high-pressure flames to sub-atmospheric pressure flames. Therefore, it is urgent to study the combustion characteristics of natural gas by conducting targeted experiments under sub-atmospheric pressure, which is the scientific basis for improving the operational optimization and design of plateau gas utilization equipment.

Due to the demand for improving performance of gas stoves and fire prevention at high altitudes, many scholars have conducted experimental studies on combustion characteristics of gas diffusion flames under low-atmospheric pressure outside the furnace, mainly distributed in temperature, flame appearance (flame length, flame width, flame color, and flame brightness), and CO emission. Yao et al. [9] conducted an acetylene free jet combustion experiment in Hefei and Lhasa, and found that the flame temperature decreased with decreasing atmospheric pressure, while the flame plume temperature increased with decreasing atmospheric pressure. Hu et al. [10] also observed the consistent conclusion. They pointed out that the variation in flame temperature with pressure was attributable to the change in combustion reaction rate and that in plume temperature to the change in air entrainment rate. However, compared to previous studies, which were conducted in an open environment, the interplay between flame radiation and flame temperature within the furnace may lead to opposite conclusions for flame temperature and plume temperature. A series of previous experimental studies indicated that the flame length was obviously affected by atmospheric pressure and fuel jet parameters. Zeng et al. [11] found that for momentum-dominated jet flame, flame length was independent on atmospheric pressure but linearly dependent on mass flow rate. The flame length of a buoyancy-dominated jet flame increased with decreasing pressure. However, Yang et al. [9] observed that the flame length of the buoyancy-dominated jet flame decreased with decreasing pressure. Kim et al. [12] measured the flame structure of acetylene-fueled laminar jet diffusion flames burning in the air at pressures of 10~100 kPa, and observed the same flame heights under the atmospheric pressure and sub-atmospheric pressure. At the same time, increased flame width and darkened flame brightness were observed with decreasing pressure [13,14]. Some scholars found the opposite conclusion about CO emission under sub-atmosphere pressure. Zhou et al. [2] investigated the influence of altitude on a natural gas stove’s CO emission in Chongqing and Lhasa, and found that oxygen-lacking conditions in Lhasa promoted the formation of CO. But Yang [15] observed the opposite conclusion about CO.

Smoke emission is a signature of incomplete combustion, and the minimum fuel flow rate (or flame height) is defined as the smoke point at which the flame just begins to emit smoke. There are limited experimental studies published regarding the smoke points of gaseous fuels, and even fewer publications have focused on the effects of sub-atmospheric pressure. Berry et al. [16] tested the smoke point of methane over a range of pressures from 101 to 1616 kPa, and found that the flame height at the smoke point decreased with increasing pressure. Dai et al. [17] found that smoke point flame lengths and fuel flow rates can be increased by increasing air/fuel-stream velocity ratios, and the effect was most pronounced at low pressures. Urban et al. [18] investigated ethylene and propane laminar jet diffusion flames burning in still and slightly vitiated air at pressures of 35–130 kPa. The smoke point flame height is roughly inversely proportional to pressure. Li et al. [19] investigated the smoke point of acetylene laminar jet diffusion flames under sub-atmospheric pressures. The measured smoke point flame height, fuel flow rate, and residence time were found to increase with decreasing pressure. However, due to the low carbon content of methane, little literature considered how sub-atmospheric pressure affects the methane smoke point.

To sum up, although the above studies have investigated the influences of low air pressure on gas combustion in an open environment, the relevant work for methane combustion behavior within a confined space under sub-atmospheric pressure is yet to be reported in the literature. Compared to methane combustion in an open environment, O₂ is accurately controlled, and the heat loss is smaller within a confined space. In addition, the radiation heat transfer between flame and wall within a confined space is the main heat-transfer method, and the flame heat transfer has a significant impact on the flow and chemical reaction process. The reduced flame radiation capacity under low pressure will inevitably lead to different combustion characteristics than those in an open environment. Hence, the guidance previous studies can provide is limited, and it is significant to further study how sub-atmospheric pressure affects gas combustion characteristics within a confined space.

To study the influence of sub-atmospheric pressure on combustion behavior of combustible substances, it is necessary to obtain the low-atmospheric pressure required for experiments. Some scholars usually adopt the following two methods: (1) Establish the same experimental platform at high- and low-altitude areas [4,10,18]. (2) Simulate a low-atmospheric pressure at low-altitude areas [1,12,19,20]. The latter is widely used because of its convenience in conducting repetitive experiments, lower experimental costs, and the ability to achieve a wider pressure range. In addition, the similar experimental results were obtained by using the above two methods. Therefore, by using the second method, we designed and built a simulated low-air-pressure experimental platform in Xi’an.

With the above considerations, the objective of the current work is to optically study the effects of sub-atmospheric pressure on combustion characteristics of a methane diffusion flame within a furnace, especially the similarities and differences between fuel rich and lean conditions. Specifically, the effects of pressure (ranging from 61.66 to 97.75 kPa) and excess air coefficients (α, ranging from 0.8 to 1.3) on temperature distribution, flame appearance, smoke point, CO emission, and NO_x emission are compared. This paper is organized as follows: Section 2 describes the experimental system and detailed case setting; Section 3 shows a detailed comparison of methane combustion under different pressures and α; and finally, the main conclusions of the study are presented in Section 4.

2. Experimental Methodology

The schematic of the overall combustion system and its external devices are shown in Figure 1. In this work, the sub-atmospheric pressure combustion environment within the furnace was achieved by three oil-free diaphragm vacuum pumps with capacities of 10 L·min⁻¹, 15 L·min⁻¹, and 20 L·min⁻¹, which was proven to be scientifically sound and feasible in many previous studies [20,21]. The low air chamber has a minimum working pressure of 50 kPa with an accuracy of ±50 Pa, a cylindrical quartz tube, an internal height of 1000 mm, and an internal diameter of 116 mm. To protect the oil-free diaphragm vacuum pump, the water and carbon black of the flue gas were filtered by the condensing-west tube, joint flask, and filter. The air was supplied through two air compressors. The CH₄ is supplied through a high-pressure gas cylinder (with the purity of higher than 99.999%) and a reducing value. The other species in the gas cylinder are mainly C₂H₆, N₂, and H₂, which will not affect the CH₄ reactivity with such a low concentration. The burner consisted of two concentric stainless-steel tubes and consists of a central fuel-jet port (external diameter 8 mm and internal diameter 5 mm) and an outer air-jet port (external diameter 25 mm and internal diameter 21 mm). A metal sintered filter was placed in the air nozzle to minimize the instabilities in the initial air flow and create a uniform exit velocity, as shown in Figure 2. A quartz observation window was embedded on the heating furnace wall to facilitate visual observation during the test process. During the fuel combustion, the furnace wall is heated electrically, maintaining a constant temperature of 673 K to avoid the ambient temperature impact. In addition, the higher wall temperature can effectively stabilize combustion and avoid flue gas condensation. The centerline temperature was measured by using a 0.5 mm diameter, butt-welded, uncoated, K-type thermocouple. A gas analyzer (Testo 335) was adopted to measure the compositions of flue gas with a focus on determining the CO and NO concentrations. The main experimental equipment type specification and uncertainty are shown in

Table 1. Experimental equipment type specification and uncertainty.

Equipment Type	Application	Measurement Span	Uncertainty
Mass flowmeter	Air/methane	Air: 10 L·min−1 Methane: 1 L·min−1	±2%
K-type thermocouple	Centerline temperature	0~1523 K	±0.5%
Pressure gauge	Reactor pressure	0~110 kPa	±0.1%
Testo 335	O2	0~25%	±0.1% under 0~4.99%; ±0.5%, 5~25%
	CO	0~10,000 ppm	±10%, 0~200 ppm; ±5%, 201~2000 ppm
	NO	0~2000 ppm	±10%, 0~99 ppm; ±5%, 100~1999 ppm

.

The environmental characteristics of low-atmospheric pressure and low-oxygen-mass concentration at high-altitude areas have a great impact on the aerodynamic parameters of gas-fired boilers. Some boilers are designed with fan selection based on low-altitude areas, which leads to a decrease in air volume. This greatly affects the efficient operation of the boiler. In recent years, the influence of low-atmospheric pressure on the operating performance of fans is gradually recognized and valued by boiler design and manufacturing enterprises. To ensure that the amount of oxygen mass concentration fed into the furnace remains unchanged and meets the fuel combustion needs, the fan power will generally scale according to the pressure. The impact of pressure on the combustion characteristics of gas within confined spaces was only investigated in our work. Therefore, the experiment was conducted under different pressure conditions by using the same air and gas mass flow rates. The specific experimental conditions selected in this work are shown in

Table 2. The specific experimental conditions under different pressures.

Case	P/kPa	$\mathbf{A}\mathbf{l}\mathbf{t}\mathbf{i}\mathbf{t}\mathbf{u}\mathbf{d}\mathbf{e}$/m	${\mathbf{V}}_{\mathbf{f}\mathbf{u}\mathbf{e}\mathbf{l}}$/m·s−1	α	${\mathbf{m}}_{\mathbf{a}\mathbf{i}\mathbf{r}}$/mg·s−1
1	97.75	300	0.074	0.8	61.63
2				0.9	69.33
3				1.0	77.03
4				1.1	84.74
5				1.2	92.44
6				1.3	100.14
8	76.62	2295	0.094	0.8	61.63
9				0.9	69.33
10				1.0	77.03
11				1.1	84.74
12				1.2	92.44
13				1.3	100.14
15	61.66	4000	0.012	0.8	61.63
16				0.9	69.33
17				1.0	77.03
18				1.1	84.74
19				1.2	92.44
20				1.3	100.14

. The fuel flow was 4.44 mg/s, the pressure within the furnace varied from 61.66 to 97.75 kPa (about equal to 0~4000 m altitude), and the α was varied from 0.8 to 1.3. In addition, the air flow rate was fixed at 5, 6, and 7 L·min⁻¹ and the fuel flow rate was slowly adjusted to the smoke point, at which the flame just begins to emit smoke when testing the smoke point. Each case was repeated three times and the root-mean-square error was used to represent the test error.

3. Results and Discussion

3.1. Temperature Distribution

Figure 3 shows the dependence of centerline temperature ($T_c$) on pressure under fuel-lean and -rich combustion. As the pressure decreases from 97.75 to 76.62 kPa and further to 61.66 kPa, there is an increase in the $T_c$ at the front of furnace. This phenomenon can be attributable to two reasons. First, the fuel burning rate decreases with decreasing pressure [22]. Second, the reduced pressure reduces the oxygen and methane mass concentration gradient at the fuel-rich region [23]. As a result, the combined effect of the two leads to a decrease in the combustion intensity and heat release rate per unit volume. At the rear of the furnace, the combustion reaction ended. Yang et al. [9] presented a correlation for predicting the buoyancy-driven flame fuel gas centerline temperature rise above the ambient temperature, as shown in Equation (1). The $T_c$ at the rear of furnace is determined by the convective heat release rate and density of fuel gas. The convective heat release rate and density of fuel gas increase with decreasing pressure. Hence, compared with $T_c$ at the front of furnace, a reverse effect of pressure at the rear of furnace on $T_c$ is observed, and the value of $T_c$ is higher with smaller pressure after point B. Interestingly, $T_c$ is not dependent on pressure at an appropriate axial height, as shown in point B in Figure 3. This may be the coupling effect of radiation and convection heat. During the experiment, the wall temperature is fixed. According to Equation (2), the heat released per unit flame volume is related to the flame temperature or flue gas temperature and the absorption coefficient of flame. As the pressure decreases, the spectral absorption bands of the triatomic gas spectrum (CO₂, H₂O) become narrow, and the absorption coefficient of the triatomic gas spectrum decreases [24]. For point A, $T_{c,A}$ decreases with decreasing pressure. The radiant heat transfer between the flue gas and the wall will be reduced. The reduced radiation heat transfer leads to an increase in $T_c$. In addition, from Equation (1), it is shown that the reduced convection heat transfer also leads to an increase in $T_c$, and such a tendency of temperature was also observed in previous methane diffusion flames and four configurations of cardboard box fires under a sub-atmospheric pressure [9,10,25]. Whether fuel lean or rich combustion, the trends of $T_c$ at the front and rear of furnace with pressure are both opposite. In addition, it is also shown in Figure 3 that as the pressure increases from 97.75 kPa to 61.66 kPa, the combustion intensity and heat release rate per unit volume decreases, and the maximum value of centerline temperature ($T_{c,max}$) in our work decreases by about 20~50 K at different α.

$$T_{FG}=9.1{\left(\frac{T_w}{g{c}_p{\rho }_{\infty }}\right)}^{\frac{1}{3}}{Q_c}^{\frac{2}{3}}{∆z}^{-\frac{3}{5}}+T_W$$

(1)

$$q=\sigma (T_F^4-T_W^4)(1-\exp \left(-\kappa L\right))$$

(2)

where $T_{FG}$ is the fuel gas temperature, $Q_c$ is the convective heat release rate of flue gas, $g$ is the gravitational acceleration, $c_p$ is the specific heat of fuel gas, ${\rho }_{\infty }$ is the density of fuel gas, and $∆z$ is the difference between the axial height and location of flame virtual origin. $q$ is the heat released per unit flame volume, $\sigma$ is the Boltzmann constant, $T_F$ is the flame temperature, $T_W$ is the wall temperature, $\kappa$ is the absorption coefficient of flame, and $L$ is the average flame optical distance.

3.2. Flame Appearance

Figure 4 shows the photographs of methane diffusion flames under different pressures. The width of flame is defined as the maximum radial width of flame envelope. As the pressure decreases, the flame width increases, providing an overall shrinking appearance to the flame as noted previously [9,11,21,24,25,26]. On the one hand, the reduced pressure leads to an increase in the mass diffusion coefficient of combustible gases, which makes it more easily diffusible and results in an increase in the combustion radius. On the other hand, the yellow carbon black, due to its higher density than gaseous products, cannot diffuse as easily as gaseous products. Therefore, soot particles are lifted upward, mainly due to the buoyancy force created by the hot gas, which leads to a stubby flame under sub-atmospheric pressure [23,27]. However, the height of maximum radial width of flame envelope increases with decreasing pressure. This is because the combustion reaction is delayed under sub-atmospheric pressure. As the pressure drops from 97.75 kPa to 61.66 kPa, the flame brightness darkens under fuel-rich combustion. It can also be observed that the rise jet velocity and reduced burning rate result in an increase in the blue region height of the flame under sub-atmospheric pressure. Figure 5 shows variations of flame length with pressure and α. It is found that the flame length under fuel-lean and -rich combustion shows a reverse trend with pressure. Under fuel-lean combustion, the α of flame region is larger than 1, the air/methane mixing at the flame region increases and the buoyancy effect decreases with decreasing pressure, which leads to a decrease in the flame length under sub-atmospheric pressure. However, under fuel-rich combustion, more unburned components are drawn into the flame area, deepening the lack of oxygen in the flame. A large amount of smoke generates to escape the flame, resulting in a decrease in flame length. The behavior we observe in the flame height variation with pressure conflicts with previous studies in an open environment [9,11,12].

3.3. Smoke Point

The fuel flow rate, flame length and residence time of smoke point under different pressures are compared, as the results show in Figure 6. As the pressure decreases from 97.75 to 60 kPa, more air is drawn into the flame region, leading to an increase in the α of the flame region under lower pressure. Hence, when the flame begins to emit smoke, the fuel flow rate of the smoke point also increases with decreasing pressure. As aforementioned, the flame length decreases with decreasing pressure under fuel-lean combustion, as the results show in Figure 4. However, it is observed that the smoke point flame length increases with decreasing pressure in Figure 6b. The soot formation increases with increasing pressure, which will reduce the smoke point flame length at a higher pressure. In a word, the increase of the smoke point flame length with decreasing pressure may be attributed to the decrease in oxygen mass per unit volume of air. In our work, the residence time is calculated here by dividing the flame height at the smoke point by the volumetric fuel flow rate [16], as shown in Figure 6c. Similarly, the residence time at the smoke point increases with decreasing pressure under sub-atmospheric pressure conditions. But other researchers observed the opposite conclusion at elevated pressures. This indicates that under sub-atmospheric pressure, the pressure-sensitive reaction pathways, which are governing the overall balance between smoke production and oxidation, are different with elevated pressure. In Figure 6, it is observed that the smoke point flame height, fuel flow rate, and residence time are found to vary with pressure to power laws, which is similar to the change patterns of ethylene, n-heptane and wood flames obtained by other researchers in an open environment [19,28]. Meanwhile, the fuel flow rates, flame lengths and residence time of the smoke point can be increased by increasing air velocity, and the effect still exists under sub-atmospheric pressures. Additionally, experimental results also show that the effect of air to the fuel velocity ratio on the smoke point is important. Further tests of the smoke point at different pressures and nozzle structures need to be conducted in the future.

3.4. CO and NO_x Emission

The Testo 335 was adopted to measure the CO of flue gas, but the accurate range of the CO sensor was 0~2000 ppm. Hence, CO emission under fuel-lean combustion was only tested and compared, as the results show in Figure 7a. It is shown that as the pressure decreases from 97.75 kPa to 61.66 kPa, CO emission decreases from 219.34 mg·m⁻³ to 81.17 mg·m⁻³ when α = 1.1, from 63.83 mg·m⁻³ to 3.75 mg·m⁻³ when α = 1.2, and from 11.55 mg·m⁻³ to 0 mg·m⁻³ when α = 1.3. This trend of CO emission is similar to that reported by Yang [15]. The lower CO emission under smaller pressure can be because, with a fixed air mass flow rate, a smaller pressure produces a higher jet flow velocity, and therefore a larger amount of ambient air is entrained into the flame to support combustion. Meanwhile, the trend of CO emission indicates that the reduced pressure promotes more complete combustion. However, Kim et al. [29] found that CO emission increases with decreasing pressure when α = 1.25. It may be caused by the poor flame stability under sub-atmospheric pressure. In addition, in Figure 7a, it is also shown that the variation of CO emission with decreasing α generally follows the same trend for all pressures.

Besides CO, the NO emission under sub-atmospheric pressure is also taken into consideration because, generally, a significantly higher level of NO emission for utilization equipment of natural gas was found when operating at high altitudes [2,5]. Figure 6b shows the dependence of NO emission on pressure. As the pressure decreases from 97.75 kPa to 61.66 kPa, NO emission decreases under fuel-lean combustion. In Figure 3 and Figure 4, the flame length and temperature at the front furnace decrease with decreasing pressure, which is unfavorable for NO_x formation via the thermal mechanism. Kim et al. [30] have observed the same change pattern for methane premixed flames. However, the NO emission appears opposite trend under stoichiometric and fuel-rich combustion. On the one hand, the flame length increased with decreasing pressure under stoichiometric and fuel-rich combustion, which promotes NO generation. On the other hand, the reduced pressure leads to improved fuel-air mixing, which promotes active groups’ (OH, O and H et al.) formation. The combined effect of the above two aspects causes the NO emission to increase with decreasing pressure under stoichiometric and fuel-rich combustion. Therefore, for those low-NO_x burners designed for low altitudes, operation at high altitudes cannot probably comply with the integrated emission standard of air pollutants. When designing low-NO_x burners, we should fully consider the influence of altitude and air distribution.

4. Discussion

Experimental investigations were carried out to explore the combustion characteristics of a methane diffusion flame within a furnace under sub-atmospheric pressure. The temperature distribution, flame appearance, smoke point, CO emission, and NO_x emission under different pressures (ranging from 61.65 to 97.75 kPa) were examined and compared. The results are summarized as follows:

(1)

The $T_c$ at the front of the furnace decreases with decreasing pressure, whereas $T_c$ at the rear of the furnace increases with decreasing pressure.

(2)

Under fuel-lean combustion, the flame length decreases with decreasing pressure. However, the flame length appears as a reverse trend under fuel-rich combustion. In addition, the rise jet velocity and reduced burning rate result in an increase in the blue region height of the flame under sub-atmospheric pressure.

(3)

The smoke point fuel flow rate, flame length and residence time increase with decreasing pressure, following the law of the negative exponent.

(4)

The CO emission decreases with decreasing pressure, which indicates that the reduced pressure makes methane combustion more complete. For NO emission, the reduced pressure results in an opposite tendency under fuel-lean and -rich combustion. With decreased pressure, the NO emission decreases under fuel-lean combustion but increases under fuel-rich combustion.