Fundamental Study on Ammonia Low-NOx Combustion Using Two-Stage Combustion by Parallel Air Jets

Abstract

Ammonia, which has advantages over hydrogen in terms of storage and transportation, is increasingly expected to become a carbon-free fuel. However, the reduction of fuel NOx emitted from ammonia combustion is an unavoidable challenge. There is the report that two-stage combustion with parallel independent jets could achieve Low-NOx combustion under ammonia/methane co-firing conditions. In order to further improve NOx reduction, we experimentally evaluated the effects of secondary air nozzle parameters, such as nozzle diameter and nozzle locations, on combustion characteristics in two-stage combustion of ammonia/natural gas co-firing using parallel independent jets. As a result of the experiments under various secondary air nozzle conditions, it was found that under the conditions where NOx was significantly reduced, the peak temperature in the furnace was observed at 300–500 mm in the axial direction from the burner, and then the temperature decreased toward the downstream of the furnace. We assumed that this temperature distribution reflected the mixing conditions of the fuel and secondary air and estimated the combustion conditions in the furnace. It was confirmed that the two-stage combustion was effective in reducing NOx by forming a fuel rich region near the downstream of the burner, and the lean combustion of the unburned portion of the first stage combustion with secondary air. We confirmed that the low NOx effects could be achieved by two-stage combustion using independent jets from the same wall under appropriate combustion and air nozzle conditions.

1. Introduction

To reduce carbon dioxide emissions from combustion, the use of green fuels generated from renewable energy sources, such as hydrogen, ammonia and eFuel, is under consideration. Among them, ammonia has several advantages over hydrogen: it can be easily liquefied at about 0.85 MPa at room temperature; it has a high hydrogen storage density per unit volume; and the infrastructure for transportation and storage is well developed [1]. On the other hand, there are many challenges in using ammonia directly as a fuel. For example, its burning velocity is much slower than that of hydrocarbon fuels [2], making it prone to combustion instability. Another issue is that a large amount of Fuel NOx is emitted during combustion [3]. To promote the use of ammonia for combustion, it is necessary to reduce the Fuel NOx derived from nitrogen in the fuel. Many studies on ammonia combustion have been reported for internal combustion, such as gas turbines [4,5,6,7]. Research on the mechanism of NOx formation and NOx reduction has been actively conducted [8,9,10,11,12], and it has been shown that NOx in ammonia combustion can be reduced by two-stage combustion in which the fuel is burned in a rich fuel region in the first stage, similar to hydrocarbon fuels [13,14,15]. In addition, in two-stage combustion of ammonia, unlike two-stage combustion of hydrocarbon fuels, intermediate products, such as NH₂, NH, HNO, and unburned NH₃ formed in the first stage of fuel rich combustion are thought to significantly contribute to the reduction of NO [16,17,18]. It was also reported that there is an optimum temperature for these denitration reactions [19,20]. Ishii et al. reported the possibility that the reduction of NOx by unburned NH₃ and intermediate products occurred in two-stage combustion in pulverized coal/ammonia co-firing [21].

In the case of two-stage combustion in a combustion furnace, it is difficult to supply secondary air into the combustion chamber from the side wall of the combustion chamber because it is common for combustion furnaces to be mounted with multiple burners. Therefore, in the case of combustion furnaces, secondary air nozzles should be mounted on the same wall as the burner, and secondary air should be supplied as parallel independent jets. Akamatsu has shown that two-stage combustion with these parallel independent jets enables Low-NOx combustion under ammonia/methane co-firing conditions [22]. To further reduce NOx emissions, it is necessary to clarify the effects of secondary air nozzle parameters, such as the nozzle diameter and nozzle location, on combustion characteristics in two-stage combustion of ammonia/natural gas mixtures using parallel independent jets.

To estimate the combustion state in a combustion furnace, it is ideal to measure the temperature in the direction of the central axis of the burner using a suction pyrometer (thermocouple). However, Matsushita et al. reported that measuring the temperature of the diffusion-burning flame of a burner using a suction pyrometer in a small experimental furnace resulted in a large measurement error due to the movement of the ignition point caused by reaction inhibition [23]. These effects are more apparent in ammonia combustion, which has low flammability. On the other hand, measuring the ambient temperature near the furnace wall using thermocouples does not interfere with the combustion reaction and is one of the few factors to estimate the combustion state in ammonia combustion with low flammability. In general, the ambient temperature near the inner wall of the furnace is affected by the combustion temperature in the vicinity, and some reports have used it to estimate changes in the combustion state of the burner under various conditions, such as a higher ambient temperature downstream of the furnace with a longer flame length [24,25]. If the combustion state of ammonia combustion can be estimated from the temperature near the inner wall of the furnace, it is expected to be used as useful information for clarifying the mechanism of low NOx emissions.

The purpose of this study is to experimentally evaluate the following two points. One is the NOx emission characteristics of a conventional hydrocarbon-fueled burner with a two-stage combustion system using parallel independent air jets, and the other is the effects of secondary air nozzle parameters on the NOx reduction effect and the ambient temperature distribution near the furnace wall. We examined the possibility of achieving low-NOx ammonia combustion.

2. Experiments and Methods

The configuration of the experimental equipment used in this study is shown in Figure 1. The inside dimensions of the experimental furnace are W300 mm × H300 mm × L1200 mm. The inside of the furnace is surrounded by 100 mm thick ceramic fiber insulation made of alumina and silica (ISOWOOL 1600, Isolite Industries, Osaka, Japan). Six ports for R-thermocouples are installed in the upper part of the furnace, and the ambient temperature at the same height as the ceiling of the furnace was constantly measured every two seconds. The axial positions of the thermocouples are Z = 100 mm, 200 mm, 300 mm, 500 mm, 775 mm, and 1050 mm, with Z = 0 mm at the furnace wall on the burner side. The R-thermocouples are installed in the exhaust gas flue at Z = 1505 mm to measure the exhaust gas temperature near the center of the exhaust gas flue. An exhaust gas sampling hole at Z = 1605 mm is connected to an exhaust gas analyzer and a NH₃ detector (GD-70D, RIKEN KEIKI, Tokyo, Japan). The thermocouple has a strand diameter of 0.45 mm and is covered with a 6 mm diameter ceramic protection tube. An automatic CO/CO₂/O₂ analyzer (EIR-31SS, Anatech Yanaco, Kyoto, Japan) and an automatic NOx analyzer (ECL-88ALite, Anatech Yanaco, Kyoto, Japan) were used to measure the exhaust gas composition. The measurement principles were the magnetic dumbbell method for O₂ and the chemiluminescence method for NOx, respectively.

We used a diffusion combustion burner of the plate-retention type in these experiments. It is one of the most commonly used burners in industrial furnace applications and is also widely used in actual industrial furnaces for the purpose of pilot and main burner applications [26]. In this study, secondary air nozzles were installed symmetrically in a horizontal direction on the same wall of the burner as shown in Figure 2. We conducted a two-stage combustion experiment using the air supplied inside the burner as primary air and the air supplied from the secondary air nozzles as secondary air. The secondary air nozzle ports were installed at three locations on each side of the burner, and the distance between the burner and the nozzles could be adjusted in three steps.

In this experiment, we used a mixture of city gas 13A and NH₃ as fuel. The typical composition of city gas 13A is CH₄ = 88.9%, C₂H₆ = 6.8%, C₃H₈ = 3.1%, and C₄H₁₀ = 1.2% [27]. The lower heating value of city gas 13A is calculated to be 40.65 MJ/Nm³, and that of NH₃ is 14.1 MJ/Nm³. City gas is supplied from a pipeline, NH₃ from a cylinder, and combustion air from a compressor. After being depressurized by a regulator, the mixture is adjusted to the desired ratio by mass flow controllers and supplied at room temperature.

In the experiments, we used the well-stored experimental furnace with a common combustion rate of 11.63 kW and an excess air ratio calculated from fuel flow and total flow of primary and secondary air λ_total = 1.2, and obtained data on the temperature distribution and exhaust gas composition at 1273 K at a point near the center of the furnace (Z = 500 mm) under various conditions. The experimental parameters were the ratio of ammonia in the lower heating value of the fuel mixture “E_NH3”, the secondary air nozzles diameter “D₂”, the distance between the burner and the secondary air nozzle “L”, and the ratio of the primary air flow rate to the secondary air flow rate. At first, E_NH3 was fixed at 10% and we adjusted the primary combustion air ratio “λ₁”, calculated from the fuel and primary air flow rates to 0, 0.2, 0.4, 0.6, 0.8, and 1.0 with secondary air nozzles of various diameters installed. Under the λ₁ condition where the lowest NOx emissions were obtained, we adjusted E_NH3 in 10% intervals, and obtained data up to exclusive NH₃ combustion (E_NH3 = 100%). The results were evaluated by comparing them with the data when using a same burner with normal single-stage combustion (λ₁ = 1.2).

3. Results and Discussion

3.1. Relationship between Primary Combustion Air Ratio “λ₁” and NOx

Figure 3 shows the relationship between the primary combustion air ratio λ₁ and NOx emissions rate for the different secondary air nozzles diameter “D₂”. The NOx emissions rate is defined as 1.0 for single-stage combustion (λ₁ = 1.2). As shown in Figure 3, NOx was reduced the most when the ratio of primary to secondary air flow was set to 1:1 (λ₁ = 0.6) for all conditions of D₂. The NOx emissions rate was reduced the most when D₂ = ϕ6.6 mm, to about 7% of that in single-stage combustion. The trend of the highest NOx reduction at λ₁ = 0.6 is consistent with the results of a previous study conducted by Akamatsu [22].

Figure 4 shows the change in temperature distribution in the furnace for different λ₁. For simplicity, only five conditions are shown in this figure. As shown in Figure 4, the downstream temperature (Z = 775 mm and 1050 mm) of the furnace decreased the most around λ₁ = 0.6 and increased as λ₁ was further decreased. As λ₁ decreased, the temperature at Z = 100 mm and 200 mm increased, but in λ₁ = 0.6, there was a large peak at Z = 300 mm.

3.2. Relationship between NH₃ Mixing Rate “E_NH3” and NOx

Figure 5 shows the relationship between E_NH3 and NOx under the condition of L = 100 mm when λ₁ = 0.6, which resulted in the highest NOx reduction rate. In this figure, results in D₂ = ϕ4.2 mm and ϕ6.6 mm are shown. In the single-stage combustion, the NOx increased as E_NH3 increased, and NOx peaked at more than 4,000 ppm when E_NH3 = 20%. As E_NH3 was further increased, NOx gradually decreased, but the value was still high at 1311 ppm even when E_NH3 = 100%. This tendency for NOx to peak at about 20–30% E_NH3 is also consistent with the results of the numerical analysis by Kunkuma et al [12]. They suggested that this was because the flame temperature and OH concentration decreased as the NH₃ mixing ratio increased.

On the other hand, when two-stage combustion was applied, NOx was significantly reduced. In the case of D₂ = ϕ6.6 mm, NOx remained almost unchanged in the range of E_NH3 = 20–80%, and it peaked at 500 ppm in E_NH3 = 50%. At E_NH3 = 100%, NOx was 300 ppm, but by adjusting λ_total from 1.2 to 1.04, a low NOx performance of 95 ppm was achieved at an O₂ = 11% conversion value, well below the upper limit of 150 ppm set by the Air Pollution Control Law in Japan.

3.3. Relationship between Secondary Nozzles Diameter “D₂” and NOx

Figure 6 shows the relationship between E_NH3 and NOx in λ₁ = 0.6 and L = 100 mm for different D₂s. However, for D₂ = ϕ11.0 mm, E_NH3 is limited to 80% because a misfire occurred when E_NH3 = 90% or more. In the case of D₂ = ϕ4.2 mm and ϕ6.6 mm, NOx tended to decrease above E_NH3 = 60%, and NOx was lower in D₂ = ϕ6.6 mm for all mixing ratios when the two conditions were compared. As D₂ increased further, NOx did not tend to decrease even under higher E_NH3, and to increase more than in the other conditions.

Figure 7 shows the relationship between the NOx emissions rate and D₂/D₁, the ratio of the burner diameter D₁ to the secondary air nozzles diameter D₂ at E_NH3 = 10%, 40%, 60%, and 80%, respectively. However, in this experiment, D₁ was fixed at 27.2 mm, and only D₂ was adjusted by replacing the nozzle with nozzles of a different diameter. The ratio of the secondary air velocity V₂ to the primary mixture velocity V₁ for each D₂/D₁ is V₂/V₁ = 45.3 when D₂/D₁ = 0.09, V₂/V₁ = 15.8 when D₂/D₁ = 0.15, V₂/V₁ = 6.16 when D₂/D₁ = 0.24, V₂/V₁ = 4.54 when D₂/D₁ = 0.29, V₂/V₁ = 2.26 when D₂/D₁ = 0.40. Under the low mixing ratio condition of E_NH3 = 10–40%, NOx emissions rate decreased as D₂/D₁ increased. However, under the high mixing ratio condition of E_NH3 = 60–80%, NOx emission rates did not tend to change significantly and conversely, and increased as D₂/D₁ increased beyond a certain value. This shows that there are different optimum values of D₂/D₁ depending on the E_NH3 value.

Figure 8a–d shows the temperature distribution in the furnace for different D₂/D₁ conditions with E_NH3 =10%, 40%, 60%, and 80%, respectively. When D₂/D₁ = 0.15 and 0.24, there was no significant change in the temperature distribution even when E_NH3 increased. However, under the conditions of D₂/D₁ = 0.29 and 0.40, the downstream temperature of the furnace increased as E_NH3 increased, and the range of increase was larger when D₂/D₁ was larger. In addition, the upstream temperature of the furnace was generally lower than that of the D₂/D₁ = 0.15 and 0.24 conditions.

3.4. Relationship between Distance between Burner and Secondary Nozzles “L” and NOx

Using the position of the primary air supply as the burner’s inner tube diameter D₁ = 27.2 mm and the ratio L/D₁, with the distance L being between the burner and the secondary air nozzle, the NOx emissions rate for different L/D₁ in λ₁ = 0.6 and D₂/D₁ = 0.24 is shown in Figure 9. In the L = 100 mm condition, L/D₁ = 3.68. When L/D₁ = 1.84, the NOx reduction effect became small, and the behavior became closer to single-stage combustion. As L/D₁ becomes smaller, the mixing of fuel and air near the burner is accelerated, and the fuel rich region becomes narrower. Thus the results are as expected. On the other hand, when L/D₁ = 5.51, NOx was further reduced in the range of E_NH3 < 80%, but this was reversed in the range of higher E_NH3, and NOx became higher. This behavior is similar to the behavior when the D₂/D₁ is larger than 0.24, as described in Section 3.3, and it was found that there is an optimum value for L/D₁ depending on the value of E_NH3.

Figure 10a–e shows the temperature distribution in the furnace at different L/D₁ for E_NH3 = 10%, 40%, 60%, 80%, and 100%, respectively. At L/D₁ = 3.68, there was no significant change in the temperature when E_NH3 increased, but at L/D₁ = 1.84, the downstream temperature of the furnace increased as E_NH3 increased in the range of E_NH3 > 40%. At L/D₁ = 5.51, the downstream temperature of the furnace did not change significantly as E_NH3 increased, but the upstream temperature of the furnace became lower.

As described in Section 1, to achieve Low-NOx combustion by two-stage combustion it is important to form a fuel rich region near the burner so that the unburnt portion is completely burned by the secondary air. Hayakawa et al. reported that the mole fraction of NO had a large peak around λ = 1.0–1.2 [28]. Therefore, the amount of NO could be reduced by fuel lean combustion with λ = 1.2 or higher. On the other hand, in the fuel rich region, NO is reduced by the reaction represented by Equations (1)–(4).

$${\mathrm{NH}}_2+\mathrm{NO}\leftrightarrow {\mathrm{N}}_2\mathrm{H}+\mathrm{OH}$$

(1)

$${\mathrm{N}}_2\mathrm{H}+\mathrm{O}\leftrightarrow \mathrm{NH}+\mathrm{NO}$$

(2)

$$\mathrm{NH}+\mathrm{NO}\leftrightarrow {\mathrm{N}}_2\mathrm{O}+\mathrm{H}$$

(3)

$$\mathrm{N}+\mathrm{NO}\leftrightarrow {\mathrm{N}}_2+\mathrm{O}$$

(4)

Therefore, the mixing of fuel and air, and combustion area formed by their mixing, have a great influence on the NOx reduction effect [29]. Figure 11 shows an image of the combustion region formed by two-stage combustion with parallel independent jets. When the mixing of fuel and secondary air takes place near the burner, the fuel rich region is shrunk and the amount of NO production in primary combustion is increased. Therefore, the reduction reactions by unburnt content and intermediate products (NHi), represented by Equations (1)–(4), are also suppressed. On the other hand, if the mixing of NHi and secondary air becomes excessively slow, it is considered that diffusion combustion close to the stoichiometric air ratio continues downstream of the furnace, and the NOx suppression effect of fuel lean combustion is not obtained.

In this study, the experimental results for various secondary air nozzle parameters showed that the conditions with large NOx reduction tended to have a temperature peak at Z = 300–500 mm and relatively low downstream temperatures at Z = 775 mm and 1050 mm in the furnace. We considered that this temperature distribution reflects the mixing condition of fuel and air. When a large NOx reduction effect is obtained, the second-stage combustion is almost complete at Z = 300 mm to 500 mm, and thereafter the temperature appears to decrease toward the downstream of the furnace. On the other hand, when the mixing of NHi and secondary air becomes excessively slow, the diffusion combustion by the secondary air continues to the downstream of the furnace, and it is considered that the temperature distribution shows an increase in temperature at the downstream of the furnace with no temperature peak at the upstream of the furnace.

Figure 12 shows the relationship between NOx and the temperature difference “ΔT” (=T_{z = 1050} (temperature at Z = 1050 mm) − T_{z = 300} (temperature at Z = 300 mm)). In Figure 12, λ_total = 1.2 and λ₁ = 0.6 are common, and all conditions of E_NH3 = 10–100%, D₂/D₁ = 0.15, 0.24, 0.29. 0.40, L/D₁ = 1.84, 3.68, 5.51 are plotted. As shown in Figure 12, there was a common tendency for NOx to be higher when ΔT increased, i.e., when the temperature increased downstream relative to upstream of the furnace.

4. Conclusions

We experimentally evaluated the NOx production characteristics of a two-stage combustion method using parallel independent air jets, supplied from the same wall as the burner, and the effect of secondary air nozzle parameters on the NOx production characteristics, and examined the possibility of realizing Low-NOx ammonia combustion. Two-stage combustion with parallel independent air jets significantly reduces the NOx emissions to below the regulation value of the Air Pollution Control Law in Japan. However, it was found that the NOx emissions characteristics of two-stage combustion significantly depended on the ratio of primary air flow rate to secondary air flow rate, secondary air nozzles diameter, and secondary air jets location. The combustion state in the furnace was estimated from the temperature distribution in the furnace, especially from the difference between the temperature downstream and upstream of the furnace, and it was found that the formation of a fuel rich reduction reaction zone near the downstream of the burner by two-stage combustion and the lean combustion of the unburnt portion of the first stage combustion with secondary air and fuel were effective in reducing NOx. And it was confirmed that these effects could be achieved by two-stage combustion using independent jets from the same wall under appropriate combustion and air nozzle conditions.