Influence of Annular Flow Area and a 30-Degree Impingement Angle on Methane/Oxygen Diffusion Flame Stability

Abstract

This work examined the effects of secondary annular flow area on flame stability in an experimental diffusion flame burner. The burner was composed of a horizontally mounted, rectangular chamber that utilized a retractable spark plug for ignition and an inverse coaxial injector. The primary and secondary gaseous reactants were oxygen and methane, respectively. Three injectors were assessed to have a fixed primary flow area and secondary flow impingement angle of 30 degrees with the primary flow and distinct secondary annular flow areas. Resultant flames and flame standoff distances were recorded via optical windows aligned parallel to the burner axis. Flame stability regime maps were generated based on the reactant equivalence ratio, the methane Reynolds number, and the injector secondary annular flow area. Results showed that among the injectors, the greater the secondary annular flow area with an impingement angle, the better the likelihood of generating a stable, anchored, fuel-rich diffusion flame for hydrogen production over the largest range of Reynolds numbers. As the secondary flow area decreased, stable diffusion flames transitioned from existing at highly turbulent flows to experiencing near-blowoff or no ignition under the same conditions. Secondary annular flow area significantly influences the location and range of stable, anchored methane/oxygen diffusion flames.

1. Introduction

A convenient method of producing desired gaseous products and syngas in methane combustion is through partial oxidation (POx) [1]. By isolating gaseous reactants in separate injector flow passageways, the reactants may exit, mix, react, and combust (using an external ignition source), creating a product diffusion flame [2]. Specifically, in methane/oxygen (CH₄/O₂) POx burners, depending on the amount of fuel-to-oxidizer, or equivalence ratio (ϕ), diverse amounts of product gases may be generated. The greatest quantities of product gases from nonstoichiometric CH₄/O₂ combustion result as hydrogen (H₂), carbon monoxide (CO), water (H₂O), and carbon dioxide (CO₂). By varying reactant flow conditions, even larger amounts of H₂ may be produced and then gathered [3]. Furthermore, if the gaseous reactants are combusted under fuel-rich, non-stoichiometric flow conditions, additional amounts of syngas may be generated in a CH₄/O₂ reaction, thereby creating gases for collection [4]. Therefore, knowing operating regimes where highly efficient, stable CH₄/O₂ diffusion flames burn would be advantageous when producing and collecting H₂ gas.

In CH₄/O₂ fuel-rich combustion, when ϕ > 2, the primary product gases are H₂ and CO. Operating under highly fuel-rich conditions, (ϕ ~ 4) results in 60% of the combustion products being H₂. Molar fractions of products from CH₄/O₂ combustion are represented in Figure 1, where the data presented were compiled using a chemical equilibrium analysis software, version year 1999, burning CH₄/O₂ at standard temperature and pressure (STP) conditions [5]. Therefore, being able to burn fuel-rich CH₄/O₂ under stable flame conditions would be beneficial to H₂ collection.

During combustion, the resultant CH₄/O₂ diffusion flame may burn as a stable (anchored, torchlike profile) or unstable (detached, oscillating) flame. While work on diffusion flames has been extensive, the understanding of the exact stable flame regimes for fuel-rich CH₄/O₂ operation is not completely known. For example, in non-premixed CH₄/O₂ flow, there are numerous operating and geometric parameters that may contribute to a stable flame behavior. Some of these parameters include the reactant type, injector type, reactant initial conditions, reactant flow conditions, injector geometry, operating pressure, etc. In this research work, focus was placed on examining the effects of the injector type (single, coaxial flow), injector features (secondary annular flow impingement angle), and injector geometry (annular flow exit area), as well as their impact on the CH₄/O₂ diffusion flame behavior and flame condition with respect to the injector exit. These parameters, combined with reactant initial conditions and flow rates, will help to further progress in understanding CH₄/O₂ diffusion flame stability, not only leading to increases in H₂ collection but also advances in operation start-up and safety, burner combustion efficiency, minimization of burner downtime, and decreased soot formation [6,7,8].

Studies have shown that when diffusion flames are stable and attached to the injector exit, up to 95% combustion efficiency (defined as the percentage of completely burned fuel) may be attained [9]. When operating as a stable flame, the diffusion flame has an increased chance of anchoring itself to the injector exit. This location allows for the burning of reactants to occur directly at the injector exit plane, thereby minimizing any unburnt fuel entering the burner and escaping out the exit. Furthermore, by having an anchor point, the flame will have a decreased chance of experiencing flame blowoff conditions (i.e., flame extinction or disappearance). Hence, where and under what parameters stable diffusion flames operate is the challenge.

Reviewing previous works specifically with CH₄/O₂ diffusion flames using single coaxial injectors operating with a constant oxidizer-to-fuel mass ratio, (O/F)_mass, showed that any increase in the reactant Reynolds number (Re_D) resulted in the flame changing from an anchored flame to an unstable flame that was detached from the injector exit [10,11,12,13,14,15]. Outcomes from these studies were that as the reactant flow inertia rose above a critical exit velocity, the diffusion flame mixing regime required additional axial distance away from the injector exit to react and combust. Other works showed that as the gap between diffusion flame velocity and flow velocity increased, or when the diffusion flame base reached a maximum radius, the flame may result in a near-blowoff flame condition due to the imbalance of flow mixing and chemical reaction time [11,16,17,18,19,20,21].

Due to the desire to achieve stable, fuel-rich burning with CH₄/O₂ for H₂ production and collection, this experimental investigation was conceived to observe the reactant flow and injector geometric parameters that influence diffusion flame stability to determine stable CH₄/O₂ diffusion flame operating regimes. The objective of this work was to establish diffusion flame stability regimes (or maps) to narrow down the influence of reactant and injector parameters that lead to stable CH₄/O₂ diffusion flames specifically from inverse shear coaxial injectors (i.e., O₂ flowed through the center, primary injector passageway and CH₄ flowed through the secondary annular flow channel). This work looked at the variation of injector flow geometry, particularly the injector secondary annular flow exit area in an atmospheric chamber environment (i.e., 1 atm) with room-temperature reactants. The secondary annular flow variation was of interest; in other studies using hydrocarbons and shear flow injectors, researchers that examined secondary flow impingement angle effects in CH₄/O₂ coaxial injectors [1,22] or CH₄/O₂/Ar coaxial injectors [23] showed that injector geometry and flow velocities affect flame behavior through either additional mixing or flow breakup. These works used constant total mass flow rate and constant Re_D,O2 [22] or constant CH₄ and Ar mass flow rates with three channel impingement angles of 45°, 60°, and 90° with respect to the injector exit [1,23]. The difference in this work compared to the previous literature was that a constant secondary flow impingement angle of 30° was evaluated with three different secondary annular flow exit areas in a single CH₄/O₂ inverse coaxial injector with a fixed center flow area and varying reactant mass flow rates.

Outcomes from this experimental work will provide new flame behavior results for single inverse CH₄/O₂ coaxial injector diffusion flames under laminar and turbulent reactant flow conditions for different secondary annular flow areas in an atmospheric pressure environment. This work contributes to the field with new diffusion flame results as a function of injector geometry and annular flow. Specific contributions from this study will be the development of flame stability region maps for CH₄/O₂ diffusion flames. These results may be used to identify appropriate fuel-rich burning regimes where stable, efficient flames reside. The goal of this new investigation was to aid POx burner and syngas designers in achieving stable operating flame conditions with gaseous CH₄/O₂ to further increase H₂ production and efficient methane burning.

2. Materials and Methods

2.1. Experimental Burner

To house the combustion and diffusion flame of CH₄/O₂, an existing horizontal burner was utilized [10,11,24]. The 316L stainless-steel rectangular-shaped burner employed 25.40 mm-thick walls and measured 152 mm × 203 mm × 660 mm in length. The burner was fabricated as individual plates and welded together. The burner interior walls were protected using an AR-06 ISO molded EDM grade fine graphite liner. The graphite liners measured 44 mm × 95 mm × 524 mm in length and included a 55.9 mm radius on the surface that was exposed to the diffusion flame. This radius gave an axisymmetric, near-cylindrical internal chamber with a constant burner hydraulic diameter (D_H) of 105 mm over a distance of 524 mm. The choice of graphite was also to help insulate the burner.

On opposite sides of the burner length, large viewing ports were cut in the burner to access the burner interior. These cutouts allowed for a 76 mm × 508 mm viewing area that ran parallel to the diffusion flame axis once the windows were installed. The viewing windows were fabricated out of transparent heat-resistant glass ceramic. The windows were mounted on opposite sides of the combustor. Each 4.8 mm-thick window was held in place using a 12.70 mm-thick rectangular 316L stainless-steel window holder, a Buna-N O-ring seal, and 20, 7.9 mm-diameter grade 5 black oxide bolts evenly torqued to 0.9 Nm [10,11,24]. Viewing windows were selected as a non-intrusive measurement method of flame observation, which was successfully employed in previous works analyzing diffusion flames [10,11,12,13,14,24,25,26,27,28,29,30,31]. Diffusion flame conditions could be identified through the windows using video cinematography (measurement methods explained in detail in Section 2.5). An exploded view schematic and photo of the experimental diffusion flame burner are shown in Figure 2.

2.2. Inverse Coaxial Injector

At the front end of the diffusion flame burner, a single shear coaxial injector was installed. The injector was fabricated out of 316L stainless steel and was designed as an inverse non-premixed coaxial injector, with a secondary annular CH₄ reactant stream having a fixed impinging angle of 30° with the axis of the primary O₂ reactant stream. This injector design was inspired by other literature injector sizes [10,11,12,13,14,24]. The gaseous reactant flows converged at the injector exit through shear mixing. AR-06 graphite was also utilized at the front end of the burner, where the injector was installed to prevent any product gas recirculation behind the injector exit plane. A schematic and photo of the CH₄/O₂ inverse coaxial injector are shown in Figure 3.

In this work, O₂ passed through the central, primary tube of the coaxial injector, and CH₄ flowed through the secondary annular area, echoing other inverse coaxial studies [10,11,14,22,24,27,29,32]. CH₄ entered the injector radially through two opposite 6.35 mm ports, and the flow was straightened before entering the burner by way of 12 equally spaced small orifices housed in a flow straightener ring installed in the secondary annular flow region (see Figure 3). The use of a flow straightener ring was based upon other coaxial flow works [10,11,12,13,14,24].

In this study, three different secondary annular flow areas were evaluated. Each injector was designed to examine variances in the CH₄ D_H and secondary annular flow exit area on diffusion flame stability. While the secondary flow exit area varied among the injectors, the injector internal diameters, O₂ D_O2, and secondary annular flow impingement angle were fixed for all injector cases. D_O2 size was selected, designed, and fabricated as 10.54 mm ± 0.025 mm to compare with other inverse coaxial flow experimental works [10,11,24], and secondary annular flow impingement angle was designed and fabricated as 30° ± 0.1°. This allowed for new information on CH₄/O₂ diffusion flame stability to be collected as a function of reactant flow parameters, secondary annular flow area, and injector impingement angle, whereas previous literature primarily focused on flow parameters for a single-element coaxial injector design with a constant secondary flow area and impingement angle [12,13,14] or non-impinging (0°), constant flow areas [6,7,8,25,26,28,29,30,31,33,34]. Schematics of the cross-sections of the single-element shear inverse coaxial injector design cases are shown in Figure 4.

2.3. Spark Plug Igniter

To ignite the reactant gases, as well as not inhibit the diffusion flame/flow or offer another anchoring location in the burner, a single 1.0 mm gap automotive spark plug (120 VAC input; 20,000 VAC output) was utilized. This spark plug igniter design allowed for a reliable ignition source and was demonstrated successfully in other research works [10,11,24]. In addition, the spark plug igniter had the ability to retract, being mounted to the end of a piston and actuated by way of a gaseous nitrogen (N₂) double-acting pneumatic cylinder. To protect the spark plug, a cylindrical 316L stainless-steel spark plug igniter mount (housing) was used. For consistency over the entire experimental study, the ignition location in the burner was set 127 mm downstream of the injector exit. Once ignited, the spark plug was retracted flush with the chamber walls away from the centerline of the burner. The spark plug igniter mount was held in place with the burner walls using a Buna O-ring piston-type seal. Photographs of the retractable spark plug igniter are shown in Figure 5.

2.4. Test Facility and Instrumentation

All experiments were performed at the Advanced Combustion and Energetics Laboratory at Penn State Altoona. The experimental diffusion burner was placed in an existing test cell that was equipped to handle energetic experiments. The test cell was constructed of 19 mm-thick outer welded steel plates and a pair of 19 mm-thick inner plywood panels on the surrounding inner walls and ceiling that measured 2.90 m × 2.95 m × 2.43 m with a retractable bay door that opened into a secure exhaust area with 457 mm-thick reinforced concrete walls.

Experiments were conducted at temperatures of 20 °C ± 1 °C (facility controlled) and atmospheric pressure conditions (i.e., the burner exit was open and no chamber pressure was built up inside the burner). After each experiment, the optical viewing windows were removed from the burner and cleaned, and external cooling fans were directed at the experimental apparatus. The next experiment was then conducted once equilibrium temperatures were reestablished. Teledyne Hastings (Hampton, VA, USA) 300 Vue Series, Model HFM-D-301B(L) mass flow meters with a calibrated range of 0 to 2.744 g/s for CH₄ and 0 to 4.997 g/s for O₂ were installed upstream of the injector to measure and record reactant flow rates during operation. Each reactant flow circuit used a 12.70 mm pneumatic ball valve that was controlled by a custom LabVIEW sequence program to introduce reactants into burner. Precision needle valves were installed downstream of the pneumatic ball valves to manually adjust the reactant mass flow rates. Prior to each experiment, the test operator would set the mass flow rates according to the test matrix flow parameters. Reactant mass flow rate and pressure data were recorded at 1000 Hz using a custom LabVIEW 2020 data acquisition program. To assist valve operation, as well as evacuate the burner post-test, N₂ was used as the valve pressurant and burner purge gas.

Using the injector exit geometries (diameter), reactant characteristics (density, ρ, and dynamic viscosity, μ), and flow conditions (mass flow rate, velocity), reactant Re_D could be determined from Equation (1).

$${\mathrm{R}\mathrm{e}}_{\mathrm{D}}={\displaystyle \frac{\mathsf{\rho }\mathrm{V}\mathrm{D}}{\mathsf{\mu }}}$$

(1)

In Equation (1), for primary flow, D_O2 was used. For secondary annular flow, Re_D,CH4, the annular flow, D_H, was used. D_H varies depending on the injector evaluated, as seen in Figure 4. Specific secondary exit flow areas are provided for each injector in the Section 3. For determining fuel-rich and fuel-lean operating regimes, ϕ was determined using (O/F)_mass as seen in Equation (2).

$$\mathsf{\varphi }={\displaystyle \frac{\left[\frac{{\dot{\mathrm{m}}}_{\mathrm{f}\mathrm{u}\mathrm{e}\mathrm{l}}}{{\dot{\mathrm{m}}}_{\mathrm{o}\mathrm{x}\mathrm{i}\mathrm{d}}}\right]}{{\left[\frac{{\dot{\mathrm{m}}}_{\mathrm{f}\mathrm{u}\mathrm{e}\mathrm{l}}}{{\dot{\mathrm{m}}}_{\mathrm{o}\mathrm{x}\mathrm{i}\mathrm{d}}}\right]}_{\mathrm{s}\mathrm{t}\mathrm{o}\mathrm{i}\mathrm{c}\mathrm{h}\mathrm{i}\mathrm{o}\mathrm{m}\mathrm{e}\mathrm{t}\mathrm{r}\mathrm{i}\mathrm{c}}}}={\displaystyle \frac{\left[\frac{1}{{\left(\frac{\mathrm{O}}{\mathrm{F}}\right)}_{\mathrm{m}\mathrm{a}\mathrm{s}\mathrm{s}}}\right]}{\left(\frac{1}{4}\right)}}=\left[{\displaystyle \frac{4}{{\left(\frac{\mathrm{O}}{\mathrm{F}}\right)}_{\mathrm{m}\mathrm{a}\mathrm{s}\mathrm{s}}}}\right]$$

(2)

2.5. Non-Dimensional Flame Standoff Distance

In the case of a detached diffusion flame, average flame standoff (or liftoff) distance (SD) was determined from the instantaneous flame standoff distance measured from the injector exit plane. This was achieved from recorded video of the experiment through conversion to individual frames. Recording of the experiments was done with a Sony (Beijing, China) HDR-PJ10 Handycam (29.97 frames per second), directed and focused on the axis of the burner. A group of frames recorded over the final 2 s of operation (i.e., steady flow) was analyzed, providing a minimum of 60 detached diffusion flame images. An open-source imageJ version 1.53e software [35] was used to determine the detached diffusion flame standoff distance. A scale of video pixels-to-measured width of a known burner object in the direction of flow was first determined, and then, using a manual tracking plugin, the detached diffusion flame front of each individual frame was selected, and a pixel value was recorded. The position of the flame was determined by selection of the flame front at the central axis of the injector primary flow. Once all the frames were analyzed, pixels were translated to engineering units, giving flame standoff distances. SD was determined over the recorded frames and divided by D_O2 to yield a non-dimensional average flame standoff distance, X_STO, which could compare flame standoff distances among the different injectors. Figure 6 shows photos of the imageJ software process for identifying the detached diffusion flame standoff distances.

3. Results and Discussion

Combined, over 640 CH₄/O₂ experiments were conducted at the Advanced Combustion and Energetics Laboratory (ACEL) at Penn State Altoona. Diffusion flame stability regime maps were generated based upon reactant flow parameters and visual observation of the flame behavior to define the regions where each flame behavior existed for the three different injectors.

Specific flow parameters of interest for each injector examined were:

• (O/F)_mass;
• Oxidizer-to-fuel momentum ratio [(O/F)_mom];
• ϕ;
• SD;
• Secondary CH₄ annular flow Reynolds number (Re_D,CH4)
• The experimental test matrix flow parameter ranges that were used for each injector case may be seen in

  Table 1. Experimental test matrix flow parameter ranges.

  Flow Parameter	Minimum Value	Maximum Value
  $\mathrm{Fuel} \mathrm{mass} \mathrm{flow} \mathrm{rate}, {\dot{\mathrm{m}}}_{\mathrm{f}\mathrm{u}\mathrm{e}\mathrm{l}}$ (g/s)	0.15	2.40
  $\mathrm{Oxygen} \mathrm{mass} \mathrm{flow} \mathrm{rate}, {\dot{\mathrm{m}}}_{\mathrm{o}\mathrm{x}\mathrm{i}\mathrm{d}}$ (g/s)	0.15	5.10
  ϕ	0.15	5.50
  ReD,CH4	750	12,250

  .

SD, for this work, was the average flame distance, which was determined from the measured distance that the base of the diffusion flame resided from the injector exit through cinematography [36]. To capture the experimental error, the general form of uncertainty was a function of independent variables. The uncertainty error of the independent variable combined the exactness error of each instrument and measurement device, as well as the experimenter’s overall consistency, through an addition of the error, and was then used to determine the resultant (or dependent parameter) uncertainty interval using a second-power equation [37]. In total, nine uncertainty parameters were estimated for flow diameters, temperatures, pressures, mass flow rates, and time. The resultant uncertainty interval for the flow parameters (i.e., velocity, mass flow rate, ReD, etc.) determined from this analysis was represented by the data point error bars in each of the result figures, as well as data comparison tables at a 95% level of confidence. Experimental results are provided in the subsequent sections.

3.1. Visual Diffusion Flame Observations

Throughout testing, the diffusion flame operating conditions were observed in real time using video cinematography and optical viewing windows on the experimental burner. Three clear diffusion flame conditions were seen and were identified as an anchored (stable) flame, an oscillating, detached (unstable) flame, or a near-blowoff (unstable) flame. These flame behavior classifications were also reported through visual observations using video cinematography in other diffusion flame works [10,11,12,13,14,23,24,25,26,27,33,38]. Individual time stamps during steady flow operation for each of the diffusion flame behaviors are shown in Figure 7, Figure 8 and Figure 9. It should be noted that the reactant flow was from right to left in the images, and that the shadow seen near the injector exit was the camera reflection off the optical viewing windowpane.

In the flame image provided in Figure 7, the stable, anchored diffusion flame was observed to be a blue, non-pulsing, bright diffusion flame, physically anchored to the injector exit. The flame remained attached to the injector exit during the entire portion of the experiment. The diffusion flame did not oscillate or pulse and decreased in overall length as CH₄ velocity increased until the flame was no longer anchored.

In Figure 8, the detached diffusion flame displayed a distinct standoff distance (or liftoff height) between the flame and injector exit (seen and indicated by the portion of the image where there is no flame). During steady flow operation (i.e., reactant steady mass flow rate), the detached flame oscillated axially in the viewing plane, with the flame front occurring at numerous locations. These oscillations were either small or large axial deviations from the previous recorded frame as a function of reactant flow conditions (ϕ and secondary CH₄ annular flow velocity). The flame liftoff height was recorded and used to calculate SD (process covered in depth in the Section 2). Overall observations indicated that as CH₄ flow velocity exceeded the local diffusion flame speed, this, in conjunction with the 30° secondary annular flow impingement angle, resulted in the diffusion flame becoming more unstable, with SD achieving greater values for greater CH₄ flows. While the three injectors evaluated in this study had different secondary annular flow exit areas, this detached flame observation with ϕ and secondary CH₄ annular flow velocity was observed in all injector cases, as well as in other single coaxial work [10,11,12,13,14,24]. Also, unlike the symmetric, stable, anchored flame condition seen in Figure 7, the detached flame was asymmetric and displayed a large, blunt-shaped flame core.

In Figure 9, a near-blowoff diffusion flame is presented. While the flame existed during the runtime, the flame itself disappeared in and out of the optical window view. This was due to the diffusion flame experiencing even greater amounts of axial fluctuations than the detached diffusion flame. The near-blowoff diffusion flame primary burning zone was housed far from the injector exit, typically in the burner exhaust duct. As seen in Figure 9, the near-blowoff diffusion flame exhibited a yellowish/orange color, with no clear shape, indicative of less oxygen and incomplete combustion. It should be noted that despite the near-blowoff flame behavior burning downstream of the injector face, all these flame classifications were considered diffusion flames since CH₄ and O₂ were separated in the coaxial injector, reacted at a flame front [16,39,40], and covered a large area over which the chemical reactions transpired [41].

Overall, from a visual observation, anchored diffusion flames were the longest, whereas the flames became shorter when transitioning to detached and near-blowoff flames. While this work used a 30° secondary annular impingement angle, this flame length phenomenon was also seen in other studies that used no impingement angle (0°), non-premixed coflow injectors, and it may be explained by elevated (O/F)_mom, with rises in primary flow Re_D,O2 [12,28].

3.2. Diffusion Flame Results—Test Cases and Experimental Data Ranges

As discussed in earlier sections, three different injectors were evaluated in this study. The injectors were defined as either baseline, larger annular flow area, or smaller annular flow area. Specific geometric values are provided in subsequent sections for each injector case. As an overall evaluation and comparison, actual ranges from each test series are provided in

Table 2. Experimental ranges for specific flow parameters of interest for each injector.

Flow Parameter	Baseline Injector	Larger Area Injector	Smaller Area Injector
(O/F)mass	0.97 ± 0.03 to 16.41 ± 0.53	0.97 ± 0.02 to 17.23 ± 0.71	0.91 ± 0.02 to 15.74 ± 0.37
(O/F)mom	0.26 ± 0.06 to 74.09 ± 18.73	0.43 ± 0.05 to 137.92 ± 6.90	0.13 ± 0.04 to 39.19 ± 1.96
ϕ	0.24 ± 0.01 to 4.11 ± 0.11	0.23 ± 0.01 to 4.14 ± 0.48	0.25 ± 0.01 to 4.39 ± 0.16
ReD,CH4	827 ± 41 to 11,037 ± 552	1403 ± 70 to 10,655 ± 533	1519 ± 76 to 11,579 ± 579

. It should be noted that the ranges displayed from this series of experiments listed in Table 2 corresponded to outcomes of either anchored, detached, or near-blowoff flame conditions. Test outcomes in which no visual combustion was demonstrated (i.e., no ignition) may be outside of these reported results. In addition, a comparison table summarizing flame behavior trends across injectors is provided in

Table 3. Flame behavior trends across injectors.

Flame Behavior	Baseline Injector	Larger Area Injector	Smaller Area Injector
Anchored	ϕ < 1.1 with ReD,CH4 < 1600	ϕ > 1.4 with ReD,CH4 < 4000; ReD,CH4 < 2200	ϕ < 0.75 with ReD,CH4 < 3800; ϕ < 1.7 with ReD,CH4 < 1700; ϕ > 1.2 with 1900 < ReD,CH4 < 2300; 1.6 < ϕ > 2.5 with 2400 < ReD,CH4 < 2900
Detached	0.6 < ϕ < 1.9 with ReD,CH4 > 2300	ϕ < 1.94 with ReD,CH4 > 3400; 0.8 < ϕ < 1.4 with 2200 < ReD,CH4 < 4000	ϕ < 1.8 with ReD,CH4 > 3400; 0.75 < ϕ < 1.2 with 1900 < ReD,CH4 < 2900
Near-Blowoff	ϕ > 2.0 with ReD,CH4 > 3400	ϕ > 1.94 with ReD,CH4 > 4000	1.8 < ϕ < 2.6 with ReD,CH4 > 4400; ϕ > 1.8 with ReD,CH4 > 760

.

3.3. Diffusion Flame Results—Baseline Secondary Annular Flow Area Case

The first series of experiments examined the diffusion flame behavior using an injector designed with a secondary annular flow area of 4.79 × 10⁻⁵ m². This secondary annular flow area was used as a baseline case for this investigation to provide a broad range of flow parameters and a large stability map to compare against other injectors with increasing and decreasing secondary annular flow area. In addition, this size secondary annular flow area and fixed primary flow area were used in other inverse CH₄/O₂ coaxial injector studies using the same experimental burner, but with different secondary impingement angles and flow conditions [10]. In total, over 280 experiments were conducted with this injector.

Figure 10 represents a non-dimensional diffusion flame stability regime map, comparing reactant ϕ to Re_D,CH4. In the stability regime map, the three observed diffusion flame conditions of anchored, detached, or near-blowoff are presented. No ignition experiments are identified by blank regions of data in the flame stability regime maps.

In reviewing Figure 10, it was observed that there were identifiable regions among the different diffusion flame classifications, with the baseline injector having a small number of experiments (<3.9%) where there was no ignition (blank regions) over the test ranges covered. The detached and near-blowoff flame types mainly resided at fuel-rich mixtures above the laminar Re_D,CH4 (>2300) condition. It was observed that for a Re_D,CH4 between 2600–4600 and ϕ > 2, the diffusion flame behavior was a combination of anchored and near-blowoff results, indicating a sensitivity to minor variances in reactant flow rates, governed by Re_D,CH4, where increases in CH₄ velocity with decreases in O₂ flow resulting in flame liftoff from the injector and a change to a near-blowoff state [36,42]. This overlap or scatter implies for this combination of secondary annular flow area and 30° impingement angle, the ignition of CH₄/O₂ and subsequent flame speed was able to initially anchor the flame to the retractable spark plug and then travel from the ignition location to the injector exit without being overcome by the reactant flow speed to push the flame further downstream from the ignition zone. Small deviations in either (O/F)_mass or (O/F)_mom, as shown in Figure 10, would result in the flame igniting, but not be able to anchor to the retractable park plug and travel back to the injector exit. Any further increase in Re_D,CH4 would result in the flame being unstable. A similar trend was seen for ϕ = 1.9, where a boundary was formed between detached and near-blowoff flames, depending primarily on ϕ, not Re_D,CH4.

As CH₄ flow entered more turbulent conditions (Re_D,CH4 > 3000) and ϕ > 0.75, primarily unstable flame behaviors were observed. This led to the hypothesis that the flame velocity was not greater or equal to the CH₄ flow velocity, forcing the mixing zone (or diffusion zone of fuel and oxidizer) for the diffusion flame away from the injector exit, where flame velocity was the speed of the diffusion flame during burning and flow velocity was the velocity of the reactant at the injector exit plane. Any increase in ϕ at high Re_D,CH4 further pushed the flame downstream in the burner until the flame was a near-blowoff or eventual blowout condition. This observed flame behavior at high Re_D complements flame blowout behavior from other works [11,15,19,21].

From an injector geometry standpoint, the 30° impingement angle tended to inject the secondary annular fuel flow into the primary oxygen flow at the injector exit. This, combined with the large velocity differences under fuel-rich flow conditions, results in the shear mixing at the injector exit being disturbed, which is a different outcome from 0° shear coaxial injectors, where velocity variances enhanced reactant mixing by introducing fresh reactants at the injector exit [2,32]. In the lower regions of the diffusion flame stability regime map, when the flow conditions were fuel-lean (ϕ < 0.5), it was observed that the diffusion flame was an anchored flame. The combination of a lower secondary annular flow velocity and 30° impingement angle for ϕ < 0.5 flows was able to maintain a sufficient reactant mixing zone for combustion at the injector exit.

Figure 11 shows the X_STO with values ranging from 1.62 to 19.81. The maximum X_STO value occurred at a Re_D,CH4 of 7358 for ϕ = 1.86 for the baseline secondary flow area 30° impingement angle injector experiments.

In Figure 11, ϕ data were presented in ranges to accommodate not only uncertainty error but also allow for increased data points where individual ϕ may have had just one data point at the exact ϕ value. From Figure 11, it was observed that the baseline injector detached flame regime covered a wide range of ϕ and Re_D,CH4. The more fuel-rich the flame, the farther away from the injector exit the diffusion flame operated. This was due to the breakup of the reactant streams, forcing the mixing of the reactants and combustion zone further away from the injector exit. This disturbance of the reactant streams was due to the increase in the amount of high-speed turbulent fuel flow in the annular stream penetrating the primary oxygen flow. X_STO increased with Re_D,CH4, with the greatest X_STO occurring for the largest recorded ϕ value. For the particular Re_D,CH4 tested range, any increase in ϕ beyond 1.86 or X_STO beyond 19.81 resulted in a near-blowoff flame due to the diffusion flame becoming more unstable from the increase in CH₄ velocity, a phenomenon also witnessed in 0° coflow work [23].

3.4. Diffusion Flame Results—Larger Secondary Annular Flow Area Case

The next series of experiments examined the diffusion flame behavior using an injector designed with a larger secondary flow area of 8.09 × 10⁻⁵ m². This specific secondary flow area size represented an increase in the secondary flow area of 68.8% compared to the baseline case while maintaining a 30° impingement angle and the same constant primary flow area.

The three flame cases for the 210 experiments conducted are presented in a non-dimensional flame stability remine map comparing ϕ to Re_D,CH4, as shown in Figure 12.

Based upon Figure 12, compared to the baseline injector, the 68.8% increase in secondary annular flow area resulted in a decrease in the number of no-ignition tests (<2.3%) and a larger percentage of observed cases of stable, anchored flames. Unlike the baseline case, anchored flames were observed in the region of Re_D,CH4 < 4000 for high fuel-rich combustion (ϕ > 2), as well as all tested laminar flow (Re_D,CH4 < 2300) for any ϕ value. The combination of the 30° impingement angle and decreased CH₄ exit velocities from the increased secondary annular flow area further promoted mixing near the injector exit compared to the baseline case. It is hypothesized that this combination allowed mixing to occur closer to the injector exit plane in addition to providing a greater annular column of CH₄ to react with the same O₂ jet amount.

Like the baseline injector, the detached and near-blowoff flame conditions seen in Figure 12 tended to gather and burn at higher ϕ and Re_D,CH4. A distinct boundary was observed between a detached and near-blowoff flame at ϕ = 1.93; dependent on ϕ, but independent of Re_D,CH4, even as Re_D,CH4 increased from 4200.

It was observed that at a ϕ > 1.94 and Re_D,CH4 of 4200, a region of mixed anchored and near-blowoff flame behaviors existed. Any minute changes to Re_D,CH4 resulted in a transition from one flame behavior to another, skipping over a detached flame case. This implied a sensitive boundary for this injector, with any increase in fuel velocity resulting in large SD oscillations and a quick transition to a near-blowoff condition [32,36]. Like the baseline injector, all three flame types were observed at Re_D,CH4 between 4000 and 4400, with near-blowoff flames appearing in the majority at ϕ > 2. Compared to Figure 10 for the baseline injector, clear flame regimes were once again identified between each flame type.

Figure 13 shows X_STO for the larger secondary annular flow injector case ranging from 3.01 to 19.47, with the maximum X_STO occurring at a Re_D,CH4 of 6859 for ϕ = 1.88. Comparing Figure 11 and Figure 13, detached flames from the larger secondary annular area injector initially resided farther away from the injector, with the closest being an X_STO of 3.01 (compared to 1.62 of baseline data). Any ϕ beyond 1.88 in this range of Re_D,CH4 resulted in a near-blowoff flame. Among the ϕ ranges, X_STO increased slightly with Re_D,CH4, with the greatest X_STO also occurring for the detached diffusion flame with the largest recorded ϕ value. Compared to the baseline injector design, the larger secondary annular flow area injector experienced the lowest X_STO maximum value, as well as having fewer detached flames below Re_D,CH4 < 4000.

3.5. Diffusion Flame Results—Smaller Secondary Annular Flow Area Case

The last series of experiments examined the diffusion flame conditions for an injector designed with a smaller secondary annular flow area of 2.75 × 10⁻⁵ m². This specific secondary flow annular area size represented a decrease in the secondary flow area of 42.5% compared to the baseline case while maintaining a 30° impingement angle and the same fixed primary flow area.

The three flame cases for the 150 experiments conducted are presented in a non-dimensional flame stability map comparing ϕ to Re_D,CH4 shown in Figure 14.

Based upon Figure 14, it was observed that there was a large portion of the flame stability regime map that had blank space or no reported combustion data (2500 < Re_D,CH4 < 6400 and ϕ > 2.6). This space represented fuel-rich operating conditions where there was no discernible flame or any ignition after the spark igniter was initiated. With a smaller secondary annular flow area, CH₄ velocity was considerably faster compared to the other injector cases studied, which may explain not only why the flame did not ignite in the indicated region, but also the earlier transition (when ϕ > 1.8) to a near-blowoff flame as the reactant mixture became more fuel-rich at Re_D,CH4 above 4400. Overall, there were 16 experiments in this region that exhibited no ignition conditions. Like the other injector cases, the detached and near-blowoff flame conditions tended to occur at higher ϕ and Re_D,CH4, with the smaller secondary annular flow area injector having the fewest stable, anchored flame cases for highly fuel-rich operating conditions (ϕ > 2). Also, when compared against the other injector cases for Re_D,CH4 < 4000, there were significantly fewer stable flames and more no-ignition outcomes for the smaller secondary annular flow area injector, signifying that insufficient mixing between CH₄ and O₂ at the same O₂ jet amount occurred.

Although the 30° impingement angle tended to drive the secondary CH₄ annular flow into the primary O₂ flow compared to a shallower angle or 0° angle, the increased CH₄ exit velocity from a smaller D_H at the same CH₄ mass flow rates and ϕ as the larger D_H injectors led to breaking up the primary O₂ core. This flow disturbance at the injector exit led to the reactants meeting further downstream, or even not at all. As seen in Figure 14, the unstable diffusion flame regions dominated operations above stoichiometric conditions, where increased amounts of CH₄ flow resided. As ϕ and Re_D,CH4 increased, the diffusion flame became even more unstable, leading to a near-blowoff state or no ignition.

When comparing the stability maps among the three different injectors, it was observed that as the reactant mixture became more fuel-rich (ϕ > 1), the greater the secondary annular flow area, the greater the chance for an anchored, stable flame. Overall, among the three different injector secondary annular flow areas, it was observed that an increase in secondary CH₄ annular flow velocity resulted in a decrease in stable, anchored diffusion flames due to the mixing and reaction zone being disrupted at the injector face and reactant flow velocities being greater than the flame velocity. This annular flow velocity trend was also observed in CH₄/O₂ studies with straight (0° impingement angle) coaxial injector flow, further emphasizing that the diffusion flame is influenced by high-speed annular fuel flow, leading to unstable flame conditions [29]. For a larger secondary annular flow area, the CH₄ velocities were smaller at the same ϕ compared to the other injector cases studied, thereby allowing the flame speed to keep up with the reactant flow velocities and stay attached to the injector face. This trend was also observed in a diffusion flame stability study with a larger secondary annular flow impingement angle [10]. For POx reactors, when looking to produce substantial amounts of H₂ in a shorter amount of time, the larger secondary annular flow area with a 30° impingement angle coaxial injector would, therefore, be best at higher flow rates. Compared to coaxial injectors that used greater secondary flow impingement angles [1,23,24], the 30° secondary annular impingement angle showed agreement in that the larger the annular exit area and the lower the flow velocities, the less chance of flow breaking up and the more stable the diffusion flame. In addition, it was observed that the location of the flame core for this work also increased with the increase of O₂ flow rate, as seen in 45◦ [1] and 60◦ impingement angle cases, albeit surrounded by an Ar sheet [23]. However, the sharper the secondary flow impingement angle with the primary flow [i.e., 45°] for the same injector geometry, the greater the chance for a stable operating flame under the same initial flow conditions [24].

Figure 15 shows X_STO for the smaller secondary annular flow injector case, ranging from 1.95 to 24.21, with the maximum X_STO occurring at a Re_D,CH4 of 7795 for ϕ = 1.73. Any ϕ beyond 1.73 in this range of Re_D,CH4 was not a detached flame. It should be noted that for the smaller secondary annular flow area injector, there were no observed detached flames operating for ϕ beyond 1.73, whereas the baseline injector and larger secondary annular flow injector experienced detached flames for this operating condition.

As can be seen from Figure 11, Figure 13, and Figure 15, the smaller secondary annular flow injector case experienced the largest maximum X_STO. In addition, this was accomplished with the lowest ϕ among the three injector cases. Comparing average X_STO values at a set Re_D,CH4 and range of ϕ, the smaller secondary annular flow area injector had higher X_STO magnitudes than the other injector cases. This result may be from the combination of the 30° impingement angle and the greater CH₄ exit velocity mixing with the primary flow further downstream from the injector face. Also, as seen in Figure 14, there were no detached diffusion flames for this injector case at ϕ > 1.8; the impingement angle and CH₄ velocity were too great and resulted in a near-blowoff or no ignition flame.

4. Conclusions

An experimental study was performed with a CH₄/O₂ inverse coaxial injector, diffusion flame burner to examine the effects of the injector secondary annular flow area on flame stability and flame standoff distance for a constant D_O2 and 30° impingement angle. The results of the experiments demonstrated the following:

• Three diffusion flame conditions: anchored, detached, and near-blowoff flames. Distinct boundaries were observed between the different flame behaviors for all three injector cases.
• The best operating parameters for maintaining a stable, anchored fuel-rich CH₄/O₂ diffusion flame were for a larger secondary annular flow area. This suggested that for the fixed 30° impingement angle and fixed primary flow D_O2, a larger CH₄ secondary annular flow area led to better mixing at the injector exit.
• It was observed that when using an inverse coaxial injector under fuel-rich operating conditions of ϕ > 2 and Re_D,CH4 < 4000 with the particular injector geometries, either anchored or near-blowoff flames occurred. To avoid the unstable near-blowoff outcome for greater fuel-rich operating conditions, it is advised not to increase Re_D,CH4 any further.
• For all three injectors, under highly turbulent Re_D,CH4 and fuel-rich conditions, detached and near-blowoff flames were the most prevalent flames. As the secondary annular flow area decreased and the CH₄ exit velocity increased (due to a smaller flow area), flame behavior transitioned to detached, near-blowoff, and even no-ignition conditions.
• For all the secondary flow areas, X_STO increased as ϕ increased for the same Re_D,CH4 values.

Outcomes from this study showed that for an inverse non-premixed CH₄/O₂ coaxial injector with fixed primary D_O2 and secondary flow impingement angle of 30°, an increase in the secondary annular flow area of 68.8% compared to the baseline case led to the greatest chance of producing a stable, anchored diffusion flame over the largest range of flow conditions. This will provide the best operating conditions for H₂ creation and collection in POx burners.

Future Directions

The authors plan to continue to conduct CH₄/O₂ diffusion flame experiments to expand the diffusion flame stability regime maps and flame standoff distance data. The goal is to then transition to the design and fabrication of multiple coaxial injectors with varying secondary flow areas and impingement angles to be utilized in experimental testing using the existing non-premixed burner. The authors plan to conduct experiments to study the effect of varying chamber volume and burner pressure on CH₄/O₂ diffusion flame stability.