Effect of rotating gliding discharges on the lean blow-off limit of biogas flames

1.   Introduction

The latest reports of the 2022 Conference of the Parties (COP 27) state that to limit global warming to 1.5 °C, the greenhouse gases produced by humans should be significantly minimized. Specialists indicate that CO₂ emissions should be reduced by 40% by 2030. However, global energy demand is still increasing, and it is necessary to find sustainable solutions. The use of biogas for combustion is considered as a net zero-carbon process. Biogas is produced by the anaerobic digestion of organic matter (animal manure, food scraps, wastewater, and sewage) by bacteria. It primarily consists of methane diluted with inert gases, including CO₂. It offers advantages, including waste valorization, fossil fuel substitution, and reduced carbon emissions [1].

However, high CO₂ dilution reduces flame temperature, increases heat capacity, and reduces burning velocity [2–6]. Yang et al [7] demonstrated that the flame front narrowed as the diluent concentration increased. In addition, the nature of the diluent was studied, and it was concluded that the CO₂-diluted flame had a narrower flame front and a lower chemical reaction compared to the N₂-diluted flame for the same dilution. To improve biogas combustion, Sivri et al [8] demonstrated that the flame temperature increased when the flow fields were changed (high swirl number) and when hydrogen was added to the mixture. Akhtar et al [9] studied various biogas mixtures in a low-swirl burner. The results revealed that an increase in the CO₂ concentration in the fresh mixture resulted in the shrinking of the burner stability map. Reduced fuel reactivity is associated with changes in the flame shape [9, 10]. Many methods, such as oxygen enrichment and hydrogen or syngas addition, have been implemented to improve biogas fuels’ burning velocity and flammability [10, 11]. However, fuel mixing leads to increased carbon emissions and fuel storage costs.

Another strategy involves increasing flammability by adding plasma actuation in the design to achieve plasma-assisted combustion (PAC) [12–18]. The use of nonequilibrium plasmas to enhance combustion has been widely studied [19–28]. The main advantages include low energy consumption and a variety of plasmas adapted for different requirements (e.g., pulsed corona, microwave, dielectric barrier discharges). Nonequilibrium plasmas have been applied to laminar and turbulent flames to extend flammability limits and improve flame stability [29–31].

In particular, the use of rotating gliding arcs (RGA) was investigated to improve the performances of burners with different fuels. Tang et al [32] showed that the blow-off limits of ammonia-air swirling flames can be extended from 0.8 to 0.4 using AC-powered gliding arc discharges. You et al [33] extended the blast furnace gas fuel extinction limits with gliding arc plasma from 0.78 to 0.44, and the amplitude of the heat released pulsation was decreased by 75%. Feng et al [34] demonstrated that the primary reason for RGA plasma extending the lean blow-off limit was because RGA-reignition replaced auto-reignition, increasing the probability of reignition at lean equivalence ratios. The RGA seems to be a promising approach to overcome the challenges of using biogas in combustion systems. Recently, Bykov et al [35] performed visible emission spectroscopy of CH₄-CO₂ mixtures with RGA and determined that plasma frequency plays a role in the emission of excited species. Ghabi et al [36] showed that a gliding arc can stabilize a biogas flame while decreasing the CO production but increasing the NO_x production. However, more studies are necessary to evaluate the effect of RGA on biogas combustion. Both the technical and fundamental aspects need to be studied.

In this work, the lean blow-off limits are determined for biogas combustion assisted by rotating gliding discharges. Three blends and two plasma powers are investigated to assess the effectiveness of rotating gliding discharges at reducing the lean blow-off limits for a low amount of added energy. A spectroscopic study is conducted to determine the dominant effects of the discharges for biogas combustion compared to pure methane. Two hypotheses are investigated: the thermal reignition proposed by Feng et al [34] and the kinetic excitation suggested by Bykov et al [35] that was demonstrated for DBD discharges [37–39].

2.   Experimental setup and methodology

2.1   Experimental setup

The experimental setup consisted of an unconfined vertically mounted low-swirl burner combined with a nonequilibrium plasma system, as schematically shown in figure 1. A detailed description of the low-swirl burner geometry and performance is available in reference [9].

Figure  1.  Schematic of the burner equipped with the high-voltage electrode.

DownLoad: Full-Size Img PowerPoint

Three mass flow controllers MFCs (MKS Instruments, 1559 series) were used to regulate the flow rates of air, methane, and carbon dioxide. Fresh gases were mixed in a 500-mm-long cylindrical mixer. Two mesh screens were placed at the inlet and exit of the mixer with a mesh size of 1 mm² to break down the large-scale turbulent structures.

A 10-mm diameter stainless-steel ring-shaped electrode was installed at the center of the burner exit. This electrode was the high-voltage (HV) electrode while the burner was grounded. The diameter of the burner rim was 33.7 mm, and the distance between the burner rim and the HV electrode was 7.3 mm. The HV electrode was connected to the power supply output (Trek 30/20 A). In this study, the rotating gliding discharges were obtained by applying a high voltage of 4 kV at a frequency, f, of 4 kHz. Voltage measurements were conducted using a high-voltage probe (P6015A), and the currents were measured directly from the Trek Power supply output. An oscilloscope (Tektronix DPO 7254C) was used to record voltage and current waveforms.

2.2   Experimental conditions

Experiments were carried out using three different fuel compositions: pure methane (CH₄), a blend of 70% CH₄ and 30% CO₂ (BG_70/30), and a blend of 50% CH₄ and 50% CO₂ (BG_50/50), under ambient conditions. The mass flow rates for the fuel–air blends were within the standard litre per minute (SLPM). To modify the equivalence ratio, the mass flow rate of the fuel was kept constant while the air flow rate was increased. Thus, the equivalence ratio, $\phi$, was varied from 0.8 to a leaner value at which the flame was blown off for a given bulk velocity. The range of bulk flow velocity studied was 2–11 m/s. A summary of the experimental test conditions is presented in table 1.

Table  1.  Experimental test conditions.

Fuel	Range of equivalence ratios $\phi$	Stoichiometric adiabatic flame temperature (K)	Lower heating value (MJ/kg)
CH4	0.49–0.70	2235	50.04
BG70/30	0.47–0.75	2154	35.02
BG50/50	0.47–0.86	2055	25.02

 | Show Table

DownLoad: CSV

Finally, two plasma conditions were tested. In both cases, the setting voltage and frequency were kept constant at 4 kV and 4 kHz, respectively, and the current was adjusted to change the average electrical power deposited in the discharges.

2.3   Emission spectroscopy

Optical emission spectroscopy measurements were performed using a spectrometer with a 750-mm focal length, equipped with a grating of 1800 lines/mm (Princeton Instruments, Acton SP2750) and an ICCD camera (Princeton Instruments, PI-MAX 3). The gain of the intensifier was set to the maximum, the exposure time was 1 ms, and the spectra were averaged over 50 measurements. Due to the rotating nature of the discharge, it was not possible to realize space-resolved measurements. The light from the electrode gap region was collected using a short focal lens to have the discharge in the field of view systematically. The recordings were acquired after a fixed delay of 10 ms from breakdown to avoid a transient state and work at constant power. Four “steady state” periods were recorded during the 1 ms exposure time. Therefore, the obtained results are not quantitative but constitute a qualitative basis for comparing the plasma effects and mixture composition on the lean blow-off (LBO) limits.

3.   Results and discussions

3.1   Electrical characterization of the discharges

Examples of the temporal evolutions of current and voltage corresponding to average discharge powers (P) of 22 and 59 W are presented in figures 2(a) and (b), respectively. Note that even though the setting voltage was kept constant, the measured voltage differed depending on the current. Therefore, the average power and applied voltage differed for the two plasma conditions. The measurements were performed for a steady-state regime far from the breakdown events. The average power was the same for all fuels, showing no direct effect of gas composition on the electrical parameters of the discharge. Under this study’s experimental conditions, the flame’s thermal power ranged between 10 and 17 kW. Thus, the discharge power represented less than 1% of the flame thermal power for all the cases.

Figure  2.  Voltage and current waveforms measured for (a) P = 22 W, f = 4 kHz, and (b) P = 59 W, f = 4 kHz.

DownLoad: Full-Size Img PowerPoint

3.2   Lean blow-off limits

Images of the flame shape and plasma morphology for the CH₄-air and BG_50/50-air mixtures with and without plasma (59 W) are shown in figure 3. The pictures were obtained for an equivalence ratio of 0.72 and a bulk flow velocity of 3.4 m/s.

Figure  3.  Direct visualizations of swirling CH₄ and BG_50/50 flames with and without plasma at f = 4 kHz and P = 59 W.

DownLoad: Full-Size Img PowerPoint

Figures 3(a) and (c) were obtained without plasma for the CH₄-air and BG_50/50-air mixtures, respectively. A V-shaped flame was attached to the ring of the high-voltage electrode for pure methane. When diluted in CO₂, the flame is stretched, lifted, and attached to the upper part of the high-voltage electrode, demonstrating a reduction in the overall burning velocity.

Figures 3(b) and (d) were obtained with plasma and an average discharge power of 59 W for the CH₄-air and BG_50/50-air mixtures, respectively. In both cases, the resulting plasma was a gliding discharge that followed the swirling motion of the flow around the high-voltage electrode. One to three filaments were observed for each experiment without transition to a thermal arc. The number of filaments and their locations were random, and no anchoring of the discharge was observed. The flame developed closer to the burner rim for both mixtures, demonstrating an increase in the overall burning velocity due to plasma action, which agrees with previous findings for DBD [37].

To further investigate the effect of plasma on the flame, a study of the LBO limits was conducted for the CH₄-air, BG_70/50-air, and BG_50/50-air mixtures without and with plasma (P = 22 and 59 W). The results are displayed in figures 4(a)–(c) as the equivalence ratio at blow-off versus bulk flow velocity. The x-axis error bars represent the errors from the MFCs readings when the airflow is increased. The y-axis error bars correspond to the repeatability of lean blow-off for a given bulk flow velocity. For the three mixtures, the empty symbols represent the case without plasma. As expected, dilution by CO₂ increased the LBO limits, particularly when the bulk flow velocity increased. These results agree with [40] and highlight that sustaining a biogas flame at high bulk flow velocities becomes challenging.

Figure  4.  Equivalence ratio at blow-off as a function of the bulk flow velocity for the three mixtures considered in this study: (a) CH₄-air, (b) BG_70/30-air, and (c) BG_50/50-air.

DownLoad: Full-Size Img PowerPoint

The half-filled and solid symbols represent cases with plasma P = 22 and 59 W, respectively. The general trend for the three mixtures was that plasma significantly enhanced the LBO limits. However, it is interesting to note that regardless of dilution, the equivalence ratio at blow-off is similar, within the uncertainties, for the same bulk flow velocity for the three mixtures when the plasma is applied. The final LBO limits obtained for the three mixtures were similar, and the CH₄-air and BG_50/50-air flames were sustained at the same equivalence ratios. For example, for a bulk flow velocity of approximately 7 m/s without plasma, the lean blow-off limit was obtained for an equivalence ratio of 0.62 for CH₄-air and 0.85 for BG_50/50. When an RGA with an average power of 59 W was applied, the equivalence ratio at blow-off was approximately 0.53 for both mixtures.

The gain that expresses the variations of the equivalence ratio at blow-off without and with plasma at P = 59 W, as a function of the bulk flow velocity, is presented in figure 5. The gain is calculated as follows:

Figure  5.  Gain of equivalence ratio by plasma actuation versus bulk flow velocity for CH₄-air, BG_70/30-air, and BG_50/50-air mixtures.

DownLoad: Full-Size Img PowerPoint

where ${\phi }_{\mathrm{w}}$ is the equivalence ratio with plasma and ${\phi }_{\mathrm{w}\mathrm{p}}$ is the equivalence ratio without plasma. If we disregard the results shown in figure 4, it seems that the plasma efficiency increases as more CO₂ is added to the mixture. However, the equivalence ratios at blow-off obtained for the three mixtures with plasma are approximately the same. From an engineering point of view, this result shows that regardless of the dilution with carbon dioxide (up to 50%), the discharges could extend the LBO down to an equivalence ratio of 0.45–0.55 for the range of bulk flow velocity studied. Optical emission spectroscopy measurements of flames with plasma were performed to understand this result.

3.3   Emission spectroscopy

In this section, we attempt to explain the presented results using spectroscopy by discussing the two hypotheses of the plasma effect on biogas flame: a thermal effect due to changes in the mixture composition or a kinetic effect due to changes in the reduced electric field.

The optical emission spectra of CH₄-air plasma + flame and BG_50/50-air plasma + flame were obtained via spectroscopy in the ultraviolet (UV) range (275–435 nm). The results were obtained for a steady-state regime, with light collection of the entire system so that the discharges were always in the field of view. Note that temporally resolved measurements could not be performed because the time of the discharges was not controlled (AC high-voltage). The two mixtures investigated are very similar in composition as the BG_50/50-air is diluted with an addition of 8% CO₂ in the blend compared to the CH₄-air mixture. Therefore, the spectra obtained for the two mixtures are similar over the entire range of wavelengths studied. However, it is interesting to compare the relative intensities of the bands to evaluate the population of excited levels, specifically for N₂ and ${\mathrm{N}}_{2}^{+}$. Note that the quenching rates with CH₄ are usually 1.5 times larger than those with CO₂ [41–47]. Thus, the following results only present qualitative information of the maximum discharge emission over 1 ms exposure time.

Figure 6 shows the spectra obtained in the 345–360 nm range for both CH₄-air plasma + flame and BG_50/50-air plasma + flame systems. The black and red lines are experimental and simulation results, respectively. The vibrational and rotational temperatures of N₂(C) can be estimated by fitting the experimental results with simulations using the Specair software [48]. Because N₂(C) is mainly populated by electron impact, it is a good representation of the fundamental level of N₂ (Frank Condon principle). Figure 6 presents the experimental and simulated bands corresponding to the 1–2 and 0–1 transitions from N₂(C³Π_u–B³Π_g) for both the CH₄-air (figure 6(a)) and BG_50/50-air (figure 6(b)) mixtures. The vibrational temperatures were estimated by fitting the relative intensities of the band heads. The rotational temperatures were estimated by fitting the wings of the bands. The uncertainties were determined by framing the rotational and vibrational levels with spectra in a wide range of temperatures (±500 K), as shown in figure 6(b). Note that the other species (molecular bands) emitting in this region were not considered in the simulations. The vibrational and rotational temperatures for CH₄-air (T_vib = 2800 K, T_rot = 2450 K) and BG_50/50-air mixtures (T_vib = 2750 K, T_rot = 2300 K) are similar, and the differences are within the uncertainties (±100 K). In other words, the thermal hypothesis cannot explain the plasma effect observed in the previous section.

Figure  6.  Experimental and simulated bands corresponding to the transitions 1–2 and 0–1 from N₂(C³Π_u–B³Π_g) for (a) CH₄-air and (b) BG_50/50-air.

DownLoad: Full-Size Img PowerPoint

The second hypothesis is that the plasma-flame chemistry is effectively being altered by the CO₂ addition. It is possible to estimate the reduced electric field in the plasma (E/N) by comparing the intensities of two lines at 391 nm for ${\mathrm{N}}^+ _2$(B²∑$^+ _{\mathrm{u}}$–X²∑$^+ _{\mathrm{g}}$) and 394 nm for N₂(C³Π_u–B³Π_g) using the work of Paris et al [49]:

where ${R}_{391/394}$ is the ratio of intensity between the 391 nm and 394 nm lines and E/N is the reduced electric field in Townsend. This method was proposed and validated for plasmas of air, where nitrogen molecules are dominantly excited from the ground state by direct electron impact [49]. In the present study, the same technique is used as the concentrations of CH₄ and CO₂ were relatively small compared to the N₂ concentration. The relative intensities of the observed lines were corrected to include the quenching of the bands by CH₄ and CO₂ using the Stern-Volmer equation [50]:

where ${I}_{0}$ is the intensity without a quencher, $I$ is the intensity with a quencher, $k$ is the quencher rate coefficient, ${\tau }_{0}$ is the lifetime of the emission line without a quencher and $\left[Q\right]$ is the concentration of the quencher. The corrected lines used to obtain the estimates are shown in figure 7. The relative intensity of the bands varies for pure CH₄ and BG_50/50. The differences can be explained by (i) a difference in the quenching of the levels by CO₂ or (ii) a difference in the reduced electric field due to a different plasma kinetic. As the quenching is corrected, the different intensities are due to changes in the reduced electric field.

Figure  7.  Spectra after quenching correction and equation used to estimate the reduced electric field, E/N, following the method described in reference [49].

DownLoad: Full-Size Img PowerPoint

The estimated reduced electric field for pure CH₄ was 450±90 Td and 730±140 Td in the case of BG_50/50. These estimated reduced electric fields constitute a reasonable explanation for the difference in relative intensity of the lines (quenching corrected). To further understand the results, these estimates can be correlated with the calculated energy loss fractions for both mixtures, as presented in figure 8.

Figure  8.  Electron energy loss fraction as a function of the reduced electric field for (a) CH₄-air mixture, and (b) BG_50/50-air mixture. The estimated E/N values are indicated in black dashed lines.

DownLoad: Full-Size Img PowerPoint

Figure 8(a) shows the electron energy loss fraction for the CH₄-air mixture. At 450 Td, the dominant processes are the electronic excitations of the CH₄, N₂, and O₂ molecules by direct electron impact (equally 25% of the energy), and the ionization of CH₄ by direct electron impact (about 10%). In the case of the BG_50/50-air mixture (figure 8(b)), the estimated E/N was at 730 Td. The dominant processes are the excitations of the CO₂, CH₄, N₂ (equally about 18% of the energy) and O₂ (about 10%) molecules by direct electron impact, and the ionization of the same molecules by direct electron impact (9%, 10%, 7%, and 8%, respectively). In this case, the energy branching is more evenly distributed between electronic excitation and ionization. These results agree with the spectra presented in figures 6 and 7, as the ion emission is stronger in the case of BG_50/50, even when the quenching is considered.

From the results obtained in figures 7 and 8, it seems that the difference in efficiency of the CH₄-air versus BG_50/50-air plasma is due to different strengths of the reduced electric field, which leads to more generation of relevant species for the combustion process.

3.4   CH₄/CO₂ plasma kinetic

In this section, we are interested in the kinetic processes that lead to combustion enhancement. That implies shortening the time of the limiting reactions. In the case of methane combustion, the main effect of the plasma is on the initiation/oxidation of the molecule [15–17] with reactions such as dissociation by electron impact:

Recent studies have demonstrated that more reactions participate in the CH₄ dissociation/oxidation in a non-negligible manner [51]:

Note that the same reactions exist for the ground state O(³P), but the rate coefficients of such reactions are several orders of magnitude lower (negligible). Therefore, to enhance the combustion properties, it seems very efficient to produce more O(¹D) than O(³P). In our plasma, there are two main ways of producing O(¹D) via O₂ and CO₂:

with reactions (7) and (8) requiring 7 and 8 eV, respectively [52]. Figure 8(b) shows that for high values of electric fields, the electronic excitation of the CO₂ molecule becomes a dominant process. The main reaction being (7) it is more likely to produce an important amount of O(¹D). In the case of pure CH₄ (figure 8(a)), the electric field is lower. Two processes compete: reaction (8) and a similar process where two ground-state O(³P) are formed.

To summarize, the CO₂ molecule allows to produce the O(¹D) state more effectively (lower energy threshold reaction (7) versus reaction (8)) and at a higher electric field, pushing the competition between O(³P) and O(¹D) towards O(¹D) for the oxygen dissociation process [52]. The O(¹D) state greatly increases the efficiency of the CH₄ dissociation, which enhances the combustion properties.

4.   Conclusions

The effects of gliding discharges on synthetic biogas combustion were investigated. Experiments were performed for three blends of CH₄-CO₂ (100%/0%, 70%/30%, and 50%/50%) and air, at atmospheric pressure. The plasma’s effects on the flame’s lean blow-off limits are reported for two electrical powers (22 and 59 W). Main findings are:

(1) Similar to what has been reported in the literature for many fuels, nonthermal plasma discharges produced by rotating gliding discharges could significantly enhance the lean blow-off limit of synthetic biogas premixed flames stabilized by a low swirling flow.

(2) This effect of plasma was obtained for a ratio of plasma power to flame thermal power as low as 1%. The enhancement of the lean blow-off limit was stronger, the higher the dilution by carbon dioxide, over the entire range of bulk flow velocities studied.

In addition, a qualitative optical spectroscopy study was performed. Main conclusions are:

(1) The vibrational and rotational temperatures of the N₂(C) state were similar for both the methane-air and the biogas-air mixtures. The gas temperature in the plasma was independent of the dilution and could not explain the difference in the stabilization effect.

(2) The addition of CO₂ to the blend seems to increase the production of the O(¹D) state, greatly improving the CH₄ dissociation in the plasma and leading to an enhanced combustion process.

These results can be used by engineers to develop new biogas combustion systems. Further investigation is necessary to determine whether the nonequilibrium plasma produced by rotating gliding discharges does not have a detrimental impact on the system’s pollutant emissions.