Study on Hydrogen Direct Injection in RNG Combustion under Various Ignition Timings for Power Generation in a Retrofitted Gas Engine

Abstract

Renewable natural gas (RNG) is attractive for energy policy goals in the world. Therefore, a regional system is designed to explore RNG combustion for power generation in localities. This study investigates a direct injection (DI) engine fueled with hydrogen (H₂) blended into the simulated renewable natural gas, which consists of 50% methane (CH₄) and 50% carbon dioxide (CO₂) in volume. In order to obtain higher efficiency, comparisons between DI and port fuel injection (PFI) of H₂ addition were made. Then, the volume percentage of H₂ was changed from 20% to 100% by keeping the volume ratio of CH₄ and CO₂ at 1:1. Finally, results of power output, brake mean effective pressure (BMEP), brake thermal efficiency (BTE) and brake specific fuel consumption (BSFC) were discussed. Results showed that in contrast to PFI, H₂ DI injection could increase efficiency by 4%. Additionally, H₂ DI could retard the MBT ignition timing at 5 °CA. Compared with CH₄/CH₄ + CO₂ combustion, under stoichiometric combustion, BMEP increases with H₂ addition but BTE decreases significantly. However, by enlarging the excess air ratio (λ) to 1.24, both BMEP and BTE increase obviously with H₂ addition. Moreover, when λ < 1.3, the MBT ignition timing should be advanced from −10 to 15 °CA top dead center (TDC). But the MBT ignition timing is fixed at −25 °CA TDC when λ is larger than 1.3. Furthermore, if efficiency is the priority, 30% H₂ addition with λ at 1.24 (−15 °CA TDC) should be selected. If higher BMEP is preferred, 20% H₂ addition with λ at 0.99 (−10 °CA TDC) should be selected.

1. Background

Nowadays, the reduction in carbon emissions has become the big issue in the world for both the developed and developing countries [1,2,3]. In order to achieve the target of becoming “carbon neutral” in the middle of the 21st century, many efforts should be made, including carbon reduction and absorption from emissions [4,5]. Renewable natural gas (RNG) is one of the most prominent fuels to help reduce emissions and generate power, which attracts the attention of researchers [6]. Now, RNG has become popular in North America and Europe due to its economic potential [7,8,9]. However, the current resources for RNG are insufficient, and a series of issues about how to establish the viability of integrating RNG into the gaseous fuel industry need to be addressed [10]. Therefore, RNG utilization in localities can be regarded as another solution. Moreover, previous reports showed that hydrogen could accelerate and enhance combustion, which could be applied in RNG combustion [11,12].

Inspired by this idea, a regional system is designed for power generation, as shown in Figure 1. In this system, animals can live and produce agricultural products. And their excrements will be used for renewable natural gas (RNG) and ammonia (NH₃) through hydrothermal pretreatment and high temperature biogas fermentation. Then, the RNG can be directly applied for combustion. However, with the consideration of security, H₂ can be produced for combustion through pyrolysis of NH₃. It can be deduced that a kind of carbon self-circulating system could be established through RNG combustion and utilization. This study hopes to make a contribution to the green economic and “carbon neutral” goals in the locality.

2. Introduction

Generally, CH₄ accounts for almost 96% of natural gas, and the characteristics of compressed natural gas (CNG) are determined by CH₄, which is the main component of fuel [13]. More importantly, CH₄ is a greenhouse gas, leading to a serious concern for the release of it through unburned fuel and the natural fermentation of biomass [14], which is the main reason for our design in RNG combustion. Furthermore, with hydrogen enrichment, the greenhouse emissions of natural gas can be decreased sharply [15]. Therefore, H₂ is produced through pyrolysis of NH₃ in this system for RNG combustion. Owing to the fast flame speed of H₂, H₂ enrichment could increase the torque and power output [16]. Additionally, it can also reduce undesirable emissions such as HC, CO and CO₂ due to more hydrogen being available [17]. However, one concern should be noticed; an increase in the amount of nitrogen oxide (NO_x) emissions can be found because of the higher combustion temperature [18]. Subsequently, different methods were proposed to decrease these NO_x emissions, such as larger λ, delaying the ignition timing and exhaust gas recirculation (EGR) [19,20,21]. Back to this system, CO₂ is usually produced with CH₄ at the same time from biomass fermentation. And it can be used for “combustion” to cool the flame. Additionally, this power generation system is much more energy-efficient without separating CO₂ from mixed fuels. Furthermore, lean-burn combustion will also be applied in our study, as it could increase efficiency and decrease NO_x as well. Previously, H₂ PFI for CH₄/CH₄ + CO₂ combustion was analyzed. In order to deeply increase thermal efficiency in one-cylinder engine combustion, a H₂ DI system is developed and utilized in this study [22,23].

Previously, Wimmer et al. [24] claimed that H₂ DI was a highly promising combustion concept to increase efficiency and decrease NO_x. Mohammadi et al. [25] conducted H₂ DI in a single-cylinder engine to test the effects of injection timing and spark timing on engine performance. Additionally, NO_x emissions were investigated under wide engine loads. The results indicated that H₂ DI prevented backfire and elevated thermal efficiency and output power during late compression stroke. Furthermore, NO_x emissions could be decreased under high engine output conditions. Then, Wallner et al. [26] designed different nozzles for H₂ DI and optimized the injector location for higher efficiency. Results showed that compared to nozzle holes, the injector location played a much more important role on H₂ combustion in engines. Mithun et al. [27] studied the H₂ only combustion performance by changing the DI timing. They demonstrated that H₂ injection timing relative to ignition timing was an important parameter to control the combustion characteristics in hydrogen direct injection combustion. Additionally, Li et al. [28] carried out an experiment on the cycle variation of H₂ DI combustion and found that the coefficient of variation (COV) increased with engine speed. Not only experiments, but also simulations were performed to analyze H₂ DI combustion for both single cylinder and multiple cylinder engines by Klepatz et al. [29]. The increased wall heat losses owing to H₂ DI could almost be compensated for by increasing the delivery rate. Therefore, efficiency can be further increased.

Nowadays, many investigations have been conducted on the H₂ addition in biogas fuel combustion because of the good performance of the H₂ DI in engine. Recently, Di Iorio et al. [30] reported different CH₄-H₂ blends in a single-cylinder direct injection spark ignition (DISI) engine under a steady state condition at 2000 revolutions per minute (rpm)—full load and stoichiometric conditions. Results indicated that when under CH₄–H₂ blend combustion, the mixture is more homogeneous, leading to higher efficiency on the propagation flame speed. Moreover, Hagos et al. [31] checked CH₄ combustion with and H₂ DI addition from 1500 to 2400 rpm with an interval of 300 rpm in a DISI engine. They found that the indicated mean effective pressure (IMEP) increased in speed in all tested conditions. Additionally, the COV in the IMEP also increased with engine speed. Furthermore, Sementa et al. [32] compared CH₄ and H₂ DI in an SI engine. Results showed that for both CH₄ and H₂, efficiency at 3000 rpm was high.

As illustrated above, much research has been performed investigating H₂ DI combustion characteristics. However, the investigations about H₂ DI facilitating CH₄ combustion are limited. Different from the above investigations, CH₄ and CO₂ were premixed before combustion in this research and then H₂ was added to obtain higher thermal efficiency for power generation in our study. Until now, few related studies were reported, leading to a research gap in which no reports can be inferred. Therefore, the effects H₂ DI addition in RNG combustion as well as ignition timing were checked in this paper. The power generation process was simulated by H₂, CH₄ and CO₂ co-combustion in a DISI engine. The quantitative analyses were conducted on power output, BTE, BMEP and BSFC. The main objective of this study was to achieve higher thermal efficiency with H₂ DI additions. This study is expected to clarify the combustion performance in an engine fueled with various volumetric fractions of H₂ DI into the cylinder at different ignition timings. The results are expected to provide practical guidance to power generation for lean-burn combustion.

3. Experimental Conditions and Setup

The specifications of the test engine are listed in

Table 1. Engine specifications.

Engine Type	Robin EH12-2DS
Compression ratio	8.5
Displacement volume	121 cc
Engine speed	1500 rpm
Bore and stroke	60 mm & 43 mm
Connecting rod length	73 mm
Ignition timing	−40, −35, −30, −25, −20, −15, −10, −5, 0, 5° CA TDC
Fuel	CH4 + CO2, 20%, 30%, 40%, 60%, 80%, 100% H2
H2 injection pressure	7 MPa
H2 injection timing	0° Compression TDC (closed valve injection)
Throttle opening	100%

. The gas engine is modified from a Robin EH12-2DS gasoline engine with a displacement volume of 121 cc and compress ratio of 8.5. The cylinder bore diameter is 60 mm and the stroke is 43 mm with the connecting rod length being at 110 mm. The engine speed is fixed at 1500 rpm for experiments in this study with the throttle opening at 100% throttle opening. The ignition timing is justified from −40 °CA TDC to 5 °CA TDC. For easy calculation and experimental operation, biogas is substituted by 50% CH₄ and 50% CO₂ in volume fraction. Then, H₂ is injected from 20% to 100% in the volume fraction of this “biogas” to accelerate the combustion flame. Meanwhile, λ is measured for each case, as shown in

Table 2. Specific flowrate and excess air ratio in each case. ( × indicates no experiment under this condition).

	Flowrate (L/min) [λ]
Case #1	CH4	CH4 + CO2	20% H2
	5.0 [1.10]	5.0 + 5.0 [1.10]	5.0 + 5.0 + 2.0 [0.99]
Case #2	CH4	CH4 + CO2	30% H2
	4.6 [1.22]	4.6 + 4.6 [1.22]	4.6 + 4.6 + 2.8 [1.02]
Case #3	CH4	CH4 + CO2	30% H2
	4.0 [1.45]	4.0 + 4.0 [1.45]	4.0 + 4.0 + 2.4 [1.24]
Case #4	CH4	CH4 + CO2	40% H2
	×	×	3.0 + 3.0 + 2.4 [1.62]
Case #5	CH4	CH4 + CO2	60% H2
	×	×	3.0 + 3.0 + 3.6 [1.55]
Case #6	CH4	CH4 + CO2	80% H2
	×	×	3.0 + 3.0 + 4.8 [1.38]
Case #7	CH4	CH4 + CO2	100% H2
	×	×	3.0 + 3.0 + 6.0 [1.31]

. One thing should be noted; the volumetric fraction of the H₂ addition is calculated from the volume of biogas. Therefore, a 100% H₂ addition suggests that the volumetric ratio of H₂ and biogas is 1:1 instead of H₂ only. And details can be seen in

Table 3. Physiochemical properties of methane and hydrogen.

Properties	H2	CH4
Density (kg/m3)	0.084	0.668
Lower heating value (LHV) (J/kg)	1.197 × 108	4.672 × 107
Volumetric LHV (J/m3)	1.022 × 107	3.297 × 107
Stoichiometric air–fuel ratio	34.32	17.16
Laminar flame speed (m/s)	2.65~3.25	0.38
Minimum ignition energy (J)	2.0 × 10−5	2.8 × 10−4
Quenching distance in air (m)	6.4 × 10−4	2.03 × 10−3

. Additionally, H₂ injection timing is controlled at 0° compression TDC (also called closed valve injection) under 7 MPa of injection pressure.

Figure 2 shows the experiment setup. A specially customized injector was used for the direct injection of H₂ into the cylinder. CH₄ and CO₂ were pre-mixed in the intake port and flowed with air into the cylinder. A dynamometer coupled with the tested engine was applied to control the load and speed. Two flowmeters were utilized to measure the flow rates of CH₄ and CO₂. The flowrate of H₂ was calibrated with the help of injection pressure and injection duration. A digital air/fuel ratio meter (LM-2) was utilized to measure λ. One data acquisition system was designed and utilized to collect λ, engine speed, flow rate of fuels and torque in this study. The measurement resolution in engine speed and flow rate are 1% and 2%, respectively. Then, parameters such as BMEP, BSFC, BTE and power output were calculated simultaneously through this system. Furthermore, four thermocouples were set to monitor the temperatures of exhaust emission, intake air, engine main body and engine oil to ensure the correctness and reproducibility of the experiment. And the measurement error for temperature was controlled at ±1℃. It should be noted that the engine was run for 30 min first, until all the above temperatures were constant; then, the experiments were conducted.

Table 2 lists the specific flow rates of CH₄, CO₂ and H₂ for each case in Table 3, with measured λ indicated. The experiments were conducted in sequence, from Case #1 to #7, corresponding with stoichiometric to lean-burn combustion. One thing that should be stated is that in Case #1, #2 and #3, CH₄ was tested first. Then, CO₂ was added at the same flow rate of CH₄ to maintain a 1:1 volume ratio to simulate biogas. Finally, H₂ was injected as presented in Table 3. The lean combustion could be achieved by decreasing the flowrate of CH₄ from Case #1 to Case #3, resulting in a correspondingly smaller λ. It can be seen that, in this order, H₂ injection decreased λ owing to more fuel addition. However, the engine is difficult to start when the flowrate of CH₄ is smaller than 4 L/min. Therefore, from Case #4 on, the experiment was conducted by just decreasing the flowrate of CH₄ and CO₂ at 3 L/min first; then, the flowrate of H₂ was increased, which leads to a larger λ from Case #4 to Case #7. Moreover, physicochemical properties of CH₄ and H₂ under room temperature are listed in Table 3. It is evident that the lower heating value (LHV) of H₂ is much higher than that of CH₄. However, owing to the lower density of H₂, CH₄ has a triple volumetric LHV of H₂. It is of note that the lamina flame speed of H₂ is much larger than that of CH₄, which is the main advantage of lean-burn combustion. In addition, both the minimum ignition energy and the quenching distance of H₂ in air are lower than that of CH₄, which will also be used to explain the results in the next section. Furthermore, the stoichiometric air–fuel ratio of H₂ is twice that of CH₄ owing to the smaller molar mass, and this fixed the theoretical reaction equation with oxygen.

4. Method of Calculation

Generally, only fuel flowrate, excess air ratio (λ), torque and engine speed can be directly obtained by the measurement. Other parameters should be calculated through data acquisition and control systems. The averaged results are calculated 50 times in repeated recordings containing 10 continuous cycles after 30 min running at the steady state. The detailed calculations are presented below.

The engine power output is calculated as follows:

$$P=\frac{n\times M}{9550}$$

(1)

where P represents power output (kW), n represents engine speed (rpm) and M represents torque (N·m).

Then, BMEP can be calculated as follows:

$$p_e=\frac{30\times \tau \times P}{n\times V}$$

(2)

where p_e represents brake mean effective pressure (MPa), τ represents the number of engine strokes and V represents engine displacement (L).

For the BSFC, as only CH₄ and H₂ are applied as “Fuel”, it can be calculated as follows:

$$b_e\left({CH}_4+H_2\right)=\frac{60\times q_{{CH}_4}\times {\rho }_{{CH}_4}+60\times q_{H_2}\times {\rho }_{H_2}}{60\times P}$$

(3)

where b_e represents brake specific fuel consumption (g/kW·h), q represents the flow rate of gas (L/min) and ρ represents fuel density (kg/m³).

Then, the heat release can be calculated as follows:

$$Q_t=1000\times \left(H_{{CH}_4}\times q_{{CH}_4}+H_{H_2}\times q_{H_2}\right)$$

(4)

where Q_t represents total heat release (J/min) and H represents the volumetric lower heating value (J/m³).

Finally, the thermal efficiency can be obtained by the following equation:

$${\eta }_b=\frac{6000\times P}{Q_t}\times 100\%$$

(5)

where η_b is brake thermal efficiency.

5. Results and Discussion

Figure 3 compares the brake thermal efficiency (BTE) of H₂ PFI and DI, with the horizontal axil being ignition timing and the vertical axis being efficiency. For all results, BTE first increases then decreases by advancing the ignition timing. And the highest efficiency among all ignition timings can be selected, which is named the maximum brake torque (MBT) ignition timing. For H₂ PFI, higher efficiency can be seen with larger λ. The fact that H₂ favors lean combustion should be the main reason for this. Additionally, the MBT ignition timing can be seen at −20 °CA TDC. For H₂ DI, a higher than 4% increase in efficiency can be achieved. And this is the main reason why H₂ DI is applied in our study. Moreover, with DI application, the MBT ignition timing is retarded at 5 °CA. In contrast to PFI, H₂ DI can be easily ignited by a spark plug near TDC, leading to higher efficiency without advancing ignition timing. All in all, due to the higher efficiency, H₂ DI with different ignition timings will be discussed in the following part. Additionally, with H₂ DI, the engine can work at the ignition timing after TDC. And this strongly implies that H2 can facilitate combustion in engines.

Firstly, the power output is checked for all cases with various ignition timings, as illustrated in Figure 4. The horizontal axis is the excess air ratio, and the vertical axis is the power output. From Case #1 to #3, λ decreases because of the lower flowrate of CH₄. However, λ increases from Case #4 to #7 owing to more H₂ enrichment. In addition, the maximum value of the power output decreases from Case #1 to #3 and increases from Case #4 to #7. The power output has a direct relationship with fuel consumption. Therefore, lean burn would decrease power output owing to the decrease in CH₄ and increases in H₂. Furthermore, the variation among different ignition timing changes is obvious. Comparing Case #2 to #3, with the same H₂ fraction of 30%, the larger λ in Case #3 shortens the variation because H₂ favors lean combustion well, leading to the stable combustion with different ignition timings. Moreover, by enlarging the H₂ fraction from 40% to 100% by just adding more H₂, a lower variation can also be expected. This suggests that with more H₂ enrichment, power output becomes less sensitive to the ignition timing effect, which also implies that with H₂ addition in biogas combustion, a higher power output can easily be obtained regardless of ignition timings.

Figure 5 shows BMEP and BTE. The left vertical axis is the brake mean effective pressure, and the right vertical axis is the brake thermal efficiency. The ignition timing is set as the horizontal axis. In Case #1, under stoichiometric combustion, BMEP increases with H₂ addition as more fuel consumption. However, BTE decreases significantly. Although H₂ enrichment could accelerate the flame speed to achieve high power, these fuels cannot be fully burned, resulting in a decrease in efficiency. Then, by increasing the H₂ fraction to 30%, a similar phenomenon can also be seen in Case #2, as λ changes less here. However, by enlarging λ to 1.24 in Case #3, an obvious distinction can be seen in Figure 5 (c). In contrast to CH₄/CH₄ + CO₂ combustion, with H₂ addition, both BMEP and BTE increase significantly. Firstly, H₂ addition favors lean combustion to increase efficiency. Additionally, with H₂ addition, more power is generated to increase BMEP, thus increasing both BMEP and BTE. Moreover, BTE is elevated more than 26% when compared to Case #1 and Case #2, which again proves that H₂ favors lean-burn combustion well. However, BMEP decreases gradually from Case #1 to #3 owing to less fuel supply.

Figure 6 depicts BEMP and BTE with H₂ addition at different ignition timings. The results of Case #1–#3 are shown in Figure 6a. Both BMEP and BTE first increase and then decrease with advancing ignition timing. Furthermore, when at stoichiometric combustion, the MBT ignition timing is at −10 °CA TDC for Case #1 and #2. However, for H₂ PFI in our previous results [33], the MBT ignition timing should be set more than −20 °CA TDC, suggesting that H₂ DI could not only increase the MBT (shown in Figure 5) but also delay the MBT ignition timing. When changed to lean-burn combustion in Case #3, the MBT ignition timing switches to −15 °CA TDC, which indicates that H₂ could accelerate the flame speed, and the MBT ignition timing should be advanced simultaneously. Additionally, with more H₂ addition from 20% to 30%, BMEP decreases but efficiency increases. The results of Case #4–#7 are shown in Figure 6b. It is interesting to see that the MBT ignition timing is fixed at −25 °CA TDC when λ is larger than 1.3. It is similar to the PFI results when 20% addition is used. By changing λ between 0.86, 1.01 and 1.22, the MBT ignition timing is fixed at −20 °CA TDC [33]. This suggests that the MBT ignition timing is mainly decided by the air–fuel ratio. When air is rich, there is a limitation that the MBT ignition does not change, even when more H₂ is added. Moreover, different to Figure 6a, with more H₂ addition, the BMEP first decreases slightly then increases. From Case #4 to #7, only H₂ is added with CH₄ being constant, and more fuel supply should increase the BMEP. However, the BTE decreases directly from Case #4 to #7, indicating that there exists a threshold at which H₂ addition cannot further increase the efficiency of lean-burn combustion. And from the current results, the threshold is 40% H₂ addition. When the H₂ fraction is smaller than 40%, with H₂ addition, the BTE increases. However, when the H₂ fraction is larger than 40%, the BTE decreases even more with H₂ addition. By the way, it should be noted that the threshold of 40% is only applicable in this small engine. When changing to a larger size, it can be supposed that this value can be increased as it has the potential to accommodate more fuel.

Figure 7 illustrates the relationship between the BMEP and BTE. It shows that high BMEP can be achieved at stoichiometric combustion. Higher BTE can be obtained at lean-burn combustion but the BMEP decreases correspondingly. Additionally, with larger λ, both the BMEP and BTE decrease significantly. Therefore, if efficiency is the priority, 30% H₂ addition with λ at 1.24 (−15 °CA TDC) should be selected. If higher BMEP is preferred, 20% H₂ addition with λ at 0.99 (−10 °CA TDC) should be selected. Furthermore, back to the power generation working condition, generally, the power output should be almost constant. Therefore, the BMEP at [0.45–0.50 MPa] is adopted. In this region, for higher efficiency, 30% H₂ addition with λ at 1.24 (−20, −15 and −10 °CA TDC) is accepted. And it will provide reference for the design and optimization in this power generation system.

Figure 8 illustrates the relationship between the BSFC and BTE. Different to Figure 7, with more H₂ addition, both efficiency and BSFC decrease, making the curves’ distribution become shorter. Furthermore, the highest BTE and lowest BSFC can be seen in Case #3, suggests that 30% H₂ addition with λ at 1.24 is the best option when considering economy and efficiency. Moreover, roughly speaking, both the BSFC and BTE decrease from Case #1 to #7.

6. Conclusions

A set of combustion experiments were conducted on RNG combustion for power generation. Different fractions of H₂ addition direct-injected into the cylinder under various ignition timings were tested to check the combustion phenomenon. The main conclusions are summarized as follows:

(1)

In contrast to PFI, H₂ DI injection could increase efficiency by 4% in this small engine. Additionally, H₂ DI could retard the MBT ignition timing at 5 °CA.

(2)

Larger λ shortens the variation in power output because H₂ favors lean combustion, leading to stable combustion with different ignition timings. Moreover, power output becomes less sensitive to the ignition timing effect with more H₂ enrichment.

(3)

In contrast to CH₄/CH₄ + CO₂ combustion, under stoichiometric combustion, the BMEP increases with H₂ addition but BTE decreases significantly. However, by enlarging λ to 1.24, both the BMEP and BTE increase significantly with H₂ addition.

(4)

When λ < 1.3, the MBT ignition timing should be advanced from -10 to 15 °CA TDC. However, the MBT ignition timing is fixed at −25 °CA TDC when λ is larger than 1.3.

(5)

There is a threshold at which H₂ addition could not further increase the efficiency at lean-burn combustion. When the H₂ fraction is smaller than 40%, the BTE increases with H₂ addition. However, when the H₂ fraction is larger than 40%, the BTE decreases.

(6)

If efficiency is the priority, 30% H₂ addition with λ at 1.24 (−15 °CA TDC) should be selected. If higher the BMEP is preferred, 20% H₂ addition with λ at 0.99 (−10 °CA TDC) should be selected.

(7)

With more H₂ addition, both efficiency and BSFC decrease. And 30% H₂ addition with λ at 1.24 is the best option when considering economy and efficiency.

Additionally, limited by this one-cylinder small engine, the thermal efficiency is still lower than 30% when different ignition timings and H₂ fractions were tested. Therefore, the relative increase in efficiency by different strategies is the most important issue to be considered instead of the absolute value. In the future, NH₃-CH₄ or CH₄- NH₃-H₂ fusion will be developed and utilized for RNG combustion in this system.