- Email: [email protected]
Hysteresis in stabilization of methane diffusion flames with plasmas
Fuel 179 (2016) 362–367
Contents lists available at ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
Hysteresis in stabilization of methane diffusion flames with plasmas Jingfeng Tang a,⇑, Liqiu Wei a, Jian Song b, Daren Yu b a b
Academy of Fundamental and Interdisciplinary Sciences, Harbin Institute of Technology, Harbin, PR China School of Energy Science and Engineering, Harbin Institute of Technology, Harbin, PR China
h i g h l i g h t s The liftoff and reattachment velocities are strongly affected by plasmas. The hysteresis regime is reduced and eventually disappeared with the addition of plasmas. A topological approach is used to predict the change of flame structures in hysteresis region.
a r t i c l e
i n f o
Article history: Received 13 November 2015 Received in revised form 18 February 2016 Accepted 29 March 2016 Available online 2 April 2016 Keywords: Methane diffusion flame Stabilization Plasma Hysteresis
a b s t r a c t The stabilization characteristics of liftoff and reattachment in methane diffusion flames have been investigated experimentally with addition of plasmas. The flame liftoff and reattachment velocities have been measured by varying the applied voltages on plasmas. The results show that the liftoff and reattachment velocities are strongly affected by the plasmas. The difference between these two velocities as well as the flame hysteresis region is reduced and eventually disappeared with the addition of plasmas. A topological approach is put forward to explore the mathematic laws of hysteresis phenomena, and the existent hysteresis and catastrophe laws of flame structure transition are interpreted under the typical operation routes. It is concluded that a change of flame stabilization will occur when the fuel velocity or the applied voltage passes through stabilization boundaries in a special direction. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction Diffusion flames are used in many industry applications due to its safe advantages. The stabilization of diffusion flames always exhibit two states: liftoff and attachment state. An attached diffusion flame lifts off the nozzle when the liftoff velocity is reached and then, if the fuel supply is decreased, the flame propagates upstream until the reattachment velocity is attained, at which point the flame quickly reattaches to the nozzle. A comprehensive series of papers correlating analytical solutions with experimental data have been published to predict the stabilization boundaries for a diffusion flame [1–6]. Since the velocity required to lift an attached flame is higher than that needed for reattachment, a stable diffusion flame can exist as liftoff or attachment state at velocities between these values. The hysteresis in flame stabilization refers to the situation that the flame has dual stabilization positions, and is comprised of a set of velocities which include the liftoff velocity and the reattachment velocity. ⇑ Corresponding author. E-mail addresses: [email protected] (J. Tang), [email protected] (L. Wei). http://dx.doi.org/10.1016/j.fuel.2016.03.102 0016-2361/Ó 2016 Elsevier Ltd. All rights reserved.
This hysteresis of diffusion flames was first investigated by Scholefield and Garside, who concluded that the hysteresis could occur with a sudden decrease in height [7]. Savas demonstrated a big difference of flame structures between attachment and liftoff flames, and reported that such difference accounted for the hysteresis phenomena [8,9]. Shin taken a detailed investigation on flame structures in the hysteresis regime, and concluded the ignition position determined whether flames were attached or lifted [10]. Lin present that the flame base was mostly located at the vortex roll-up position, the roll-up of air into vortex inhibited the flame propagation and eventually caused the hysteresis phenomenon [11]. Lyons parametrically studied hysteresis behaviors by using methane diffusion flames under air-coflow conditions, and present that the co-flows could affect the nature of flame stability as well as the flame hysteresis [12–14]. Combustion enhancement by plasmas or electric fields has been investigated extensively for the purpose of improving flame stabilization. The earliest observation of electric field (plasma) interaction with flames was the work of Brande [15]. They found that flames were attached to electrodes due to the existence of charged particles in the flames. Chung [16–18], Lacoste [19], Bak [20], Belhi [21] and Lyons [22] applied AC or DC electric fields to various
363
J. Tang et al. / Fuel 179 (2016) 362–367
flames and demonstrated that the plasmas could impose an important effect on the flame structure. Chung summarized the effect of AC and DC electric fields or plasmas on diffusion flames, and demonstrated a reattachment process from a lifted flame to an attached flame with the existence of electric fields [23]. They concluded that the AC voltage amplitude always posed positive effects on flame enhancements, while the AC alternating frequency always posed positive effect on the laminar liftoff velocity as well as negative effect on the reattachment velocity. The DC voltage amplitude had positive effects for the unsteady flames, while had small effects for stationary flames. The combustion enhancement mechanisms by plasmas have mainly been explained as three effects. First is heating effect caused by plasmas [24], second is the enhancement of reaction kinetics by active and excited species produced by plasmas [25], and the third is mainly associated with ionic wind effects [26]. Recently, Vincent-Randonnier demonstrated that when a DBD plasma was activated on a diffusion methane flame, the liftoff height of the lifted flame was significantly decreased [27]. They reported that the reduction in the flame liftoff/reattachment hysteresis could be caused by the DBD discharge, but that the hysteresis couldn’t be eliminated at their works. Ju provided a comprehensive review about plasma assisted ignition and combustion, and highlighted that since plasmas has greater kinetic effects on ignition, plasma addition could change the flame ignition/extinction S-curve to a monotonic curve in the hysteresis regime [28]. With the addition of plasmas, the stabilization of diffusion flame is needed to understand the possible change of hysteresis characteristics. This present study experimentally focuses on effects of plasmas on diffusion flame stabilization in hysteresis regimes. With an argon plasma injecting into a methane diffusion flame, the diffusion flame in the hysteresis region is examined. Liftoff and reattachment velocities are collected and analyzed for various plasmas. The methane flame hysteresis features are studied based on topological geometry methods, and the topological rules of diffusion flames in the hysteresis regime are provided and also interpreted with the addition of plasmas.
Fig. 1. Schematic diagram of experimental setup.
The setup used for generating and controlling of the discharge is composed of an AC signal generator and an amplifier. The AC generator produces a sinusoidal signal with a low amplitude and with a high frequency. This signal is transmitted to the amplifier for amplitude enhancement. The amplifier output signal is applied to the needle electrode with adjustable voltage amplitudes and repetition frequencies. The current and voltage signals are measured by a Rogowski coil (Pearson 6585) and a high-voltage probe (Tektronix P6015A), and the signals are recorded by a digital oscilloscope (Tektronix DPO4104). Flame images under different experiment conditions are taken by a digital camera (Nikon D7000) with an exposure time up to1/1250 s and a fixed ISO 6400.
2. Experimental setup The schematic diagram of experiments is shown in Fig. 1. A simple floating electrode to generate an atmospheric plasma jet is adopted, which is made of a stainless steel needle with a diameter of 1 mm. When argon is injected into the needle and the alternative voltage (amplitudes up to 20 kV peak-to-peak, repetition rate ranged from 45 kHz to 50 kHz) is applied, a homogeneous plasma jet is generated in front of the end of the needle and into the surrounding air. Such discharge configuration has be shown to concentrate electrical energy into the plasma simply and efficiently [29]. The needle electrode is inserted into a T-shaped quartz tube through a rubble plug, and its tip reaches a same horizontal position with the tube top. The inner and outer diameters of the tube exit are 2 and 4 mm, respectively. The tube top inserted with the needle electrode is acted as fuel jet nozzle, and the mixing between methane and argon occurs at the exit of nozzle. A picture of the fuel jet nozzle is added into Fig. 1 to describe the possible interaction region between diffusion flames and plasmas. When methane is injected from the tube bypass inlet, and a special excitation voltage is applied to the floating electrode, a diffusion methane flame assisted by an argon plasma jet is established in front of the tube exit. With this configuration, the effects of the plasma jet on the flame can be easily observed, and the changes in flame properties (detachment height, flame height, liftoff velocity, etc.) can be quantified.
3. Structure transition of methane diffusion flames with argon plasma jets 3.1. Effect of plasma jets on the flames’ structure The classical characteristics of a diffusion methane flame are observed. The flame is always attached to the tube until the methane jet velocity VJ is up to 23 m/s, and the flame is mostly of yellow emission. At VJ = 10 m/s, the flame base partially detach from the tube. Due to the partially premixing between air and methane at this detached position, the lifted flame is less emissive. HD increases with VJ until the flame is blown off. With the addition of plasma jet, the flame structures for different velocities VJ are photographed in Fig. 2. For an attached flame with an excitation voltage amplitude U around 0–2 kV, the flame height remains unchanged and no apparent phenomenon is observed. With the voltage amplitudes increased up to 2 kV, a luminous plasma jet coming is clearly apparent in the flame bottom, and the flame keep attached structure and shook crazily with less emission. For a lifted flame with an excitation voltage U up to 2 kV, the flame structure is strongly affected by the plasma jet. At VJ = 20 m/s and U = 4 kV, the lifted flame moves upstream and the flame base becomes slanted, as exhibited in Fig. 2e. Even though the flame is partially-attached, the flame edge is relatively smooth without a cusp-like behavior. With the voltage U increased
364
J. Tang et al. / Fuel 179 (2016) 362–367
Fig. 2. Pictures of methane flames with 0.25 s exposure time (a) (VJ [m/s], U [kV]) = (10, 0); (b) (10, 40); (c) (10, 80); (d) (20, 0); (e) (20, 40); (f) (20, 80) (white dotted line: position of tube exit).
up to 8 kV, a luminous plasma jet becomes clearly apparent at the flame bottom, and an electrical contact between the flame and the plasma jet is established, as exhibited in Fig. 2f. The plasma jet goes from the needle to the flame front and expands through it, and the flame serves as an extended electrode. The similar phenomenon is also mentioned by Vincent-Randonnier’s work, and explained as that the high temperature and low density flames facilitate the progression of discharges [27]. With this electrical contact establishment, the flame is anchored to the plasma jet as well as to the tube. The attachment of the flame or the flame anchoring is mainly related with the complicated physical and chemical effects of plasma jets. 3.2. Hysteresis in structure transition of methane diffusion flames with argon plasma jets For a diffusion methane flame without applications of plasma jets, the relationship between the flame detachment distance HD and the jet velocity VJ is plotted in Fig. 3a. There are two data curves with different operation directions, such as, with increasing VJ for the attached flame situation, or with decreasing VJ for the
lifted flame situation. The flame lifts off with VJ increasing up to 23.3 m/s, and the flame reattaches at the VJ of 10.6 m/s. Since the jet velocity VJ required for liftoff an attached flame is higher than that needed for reattachment, a stable flame exists as lifted or attached state at a jet velocity from 10.6 m/s to 23.3 m/s. The hysteresis in flame stabilization refers to the situation that the flame has dual stabilization positions, and is comprised of a set of velocities which include the liftoff velocity and the reattachment velocity. Thus, the structure of the methane flame in this hysteretic region depends more on its prior state than solely on the flow velocity. The similar experiment is repeated at the voltage U of 8 kV. The difference between the liftoff velocity and the reattachment velocity is 3.3 m/s instead of 12.7 m/s for no plasma, as exhibited in Fig. 3b. An attached flame lifts off the tube when the jet velocity is reached at VJ = 20.2 m/s and then, if the jet velocity is decreased, the flame propagated upstream until the jet velocity is attained at VJ = 16.9 m/s, at which point the flame quickly reattaches to the tube. In the existence of plasmas, the flame liftoff occurs at lower VJ than that for U = 0 kV due to plasma effects. With the excitation voltage amplitudes U increased up to 13 kV, an attached flame is kept at the tube exit, especially at high jet velocity with the electrical contact between the flame and the plasma jet. As stated above, the flame is anchored to the plasma jet as well as to the tube, and kept in an attachment state. At this excitation voltage, the flame shook crazily and the blow-off occurs easily. The experiments are repeated and the attachment state is kept beyond the flame blow-off. 4. Mathematical rule: topological property in structure transition of methane diffusion flames with argon plasma jets In this section, the topic is focused on the general rules for structure transition in methane diffusion flames with argon plasma jets. 4.1. Topological invariant rules for the boundaries of structure transition The system outputs can be predicted from system inputs easily for a determined system without hysteresis. But for a system with hysteresis, it is difficult to give the system outputs only with inputs [30]. It means that the operation routes are as important as inputs, to predict the final system outputs. According to the theory of dynamic systems, the critical conditions for mode transition correspond to system singular points, which can be expressed with potential function as [30],
Fig. 3. Hysteresis effect in HD (a) without plasma; (b) with plasma at U = 8 kV.
J. Tang et al. / Fuel 179 (2016) 362–367
) @VðX; UÞ=@X ¼ 0 ðX; UÞ 2 @V ðX; UÞ=@X 2 ¼ 0
(
ð1Þ
where X is the n-dimension state variable, U is the m-dimension input variable, and V is the C1-scalar potential function of the system. Near the singular point, a small change of the system input always result in a sudden jump of the state variable, as well as a output jump containing the significant behaviors of mode transition. According to Thom’s classification theorem, the potential function for a cusp singularity can be provided by [31],
VðX; UÞ ¼ x4 þ u1 x2 þ u2 x:
ð2Þ
The cusp singularity is illustrated in Fig. 4a, which is described by the equilibrium surface under a three-dimensional coordinate space of [u1 u2 x]. The fold edges in the equilibrium surface are called as the singularity set, which can be obtained by the projection of equilibrium surface. If an operation route passes through the singularity set, a catastrophic jump will take place. If the system remains far away the singularity set, the system behavior would vary continuously with inputs. Therefore, the singularity set defines the conditions where a sudden change of the system can occur. The evolution of the flame detachment distance HD versus the jet velocity VJ is plotted in Fig. 4b. It can be seen that the flame stabilization boundary is a cusp-shaped curve, which corresponds to the singularity set for singular systems. The characteristics of singular systems provide a topological rule to understand the flame behaviors. Based on this topological property, it should be possible to predict whether the hysteresis exists, and easy to understand the flame behavior along different operation routes.
365
catastrophe, which correspond to the two stable structures: liftoff and attachment state. In fact, the catastrophe lines correspond to boundaries at which the methane flame will transit between liftoff and attachment state. The catastrophe boundaries can be used to determine when the transitions occur between these two stable structures. As remaining outside the singularity set, the flame behavior varies continuously with (U and VJ). However, the catastrophe and hysteresis are inevitable. 4.2.2. Interpretation of hysteresis and catastrophe behaviors Fig. 5 provides the conditions of mode transition by velocity VJ and voltage U. To investigate flame behaviors, operation routes around the singularity set are described in this figure, and each operation route has two opposite directions. The hysteresis behaviors along routes I have been analyzed and demonstrated in Fig. 3. Fig. 6 shows the changes of HD with U along routes II and III, respectively. The methane flame has exhibited catastrophe behaviors different from hysteresis behaviors. Once the flame structure transits from an attachment state to a liftoff state along route II, it is impossible to transit back to an attachment state. Similarly, it is also impossible to transit back to a liftoff state from an attachment state along route III, if it has previously been reattached to the tube. The corresponding experiments have been investigated to validate the catastrophe characteristic as discussed above. An attached flame with a constant VJ = 20 m/s will transits to a liftoff state while U is applied up to 10 kV, and then it is not possible to transit back to an attachment state even U is deceased at any values. Similarly, an lifted flame with a constant VJ = 15 m/s will transits to an attachment state while U is applied up to 10 kV, and then it is not
4.2. Topological interpretation of structure transition of methane diffusion flames with argon plasma jets 4.2.1. Interpretation of catastrophe boundary As shown in Fig. 4b, the occurrence of flame structure transition corresponds to the catastrophe phenomena. That is, when applied voltage U is low (U = 0 kV, 4.0 kV, 8.0 kV and 10.0 kV), there exists a two-state stable structure (liftoff and attachment state) with the change of jet velocity VJ. However, when U is relatively high (U P 12.0 kV), there exists only one stable structure (attachment state) with the change of VJ, and there is no catastrophe or jump in the process. This is the characteristic of catastrophe. As shown in Fig. 4a of the topological geometry space of catastrophe, u1 and u2 is regarded as the topological transformation of U and VJ, respectively. It can be seen that both the upper and lower branch of the pitchfork have stable equilibrium points in this
Fig. 5. Type of control routes and singularity sets descript by (U, VJ).
Fig. 4. Topological property of structure transition in methane flames with argon plasma jets (brown dotted line: singularity set; green dotted line: stability boundaries). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
366
J. Tang et al. / Fuel 179 (2016) 362–367
Fig. 6. HD versus (U, VJ) along route II and route III.
Fig. 7. Pictures (0.25 s exposure time) of the flame behaviors passing through different routes (a) and (c) VJ = 20 m/s; (b) and (d) VJ = 15 m/s.
possible to transit back from an attachment state to a liftoff state. The flame behaviors are exhibited in Fig. 7 for different control routes. Obviously, the topological property in structure transition is useful to understand and predict the methane flame behavior along different operation routes. 5. Conclusion The stabilization of the methane diffusion flames has been investigated experimentally with the addition of plasmas under a single electrode discharge configuration. The results show that the flame structure is strongly affected by the plasma jet. The difference between liftoff and reattachment velocities is also affected, as well as the flame hysteresis region is reduced and eventually disappeared with applications of plasmas. Once the flame transits from an attachment state to a liftoff state with varying the applied voltage amplitude, it is impossible to transit back from a liftoff state to an attachment state even the plasma jet application is canceled. Similarly, it is impossible for the flame state to transit back, if it has previously been an attachment state from a liftoff state by the applied plasma jet. To understand the hysteresis phenomenon
of the methane flame structure transition, a topological rule is also proposed based on Thom’s classification theorem, which provides the singularity set of structure transition. With such topological rules, the flame stabilization boundaries are obtained, and the existent hysteresis phenomenon are interpreted under typical operation routes passing through the stabilization boundaries. Such general rules in methane flames may be useful for further applications, such as improving or controlling flame stabilization. Acknowledgment This work was supported in part by National Natural Science Foundation of China (51006027, 51437002, and 51477035). References [1] Lyons KM. Toward an understanding of the stabilization mechanisms of lifted turbulent jet flames: experiments. Prog Energy Combust Sci 2007;33:211–31. [2] Chung SH, Lee BJ. On the characteristics of laminar lifted flames in a nonpremixed jet. Combust Flame 1991;86(1–2):62–72. [3] Lee BJ, Chung SH. Determination of schmidt number of mixed fuels by the characteristics of laminar lifted jet flames. Fuel 2006;85:68–74. [4] Han D, Mungal MG. Direct measurement of entrainment in reacting/ nonreacting turbulent jets. Combust Flame 2001;124(3):370–86.
J. Tang et al. / Fuel 179 (2016) 362–367 [5] Watson KA, Lyons KM, Donbar JM, Carter CD. On scalar dissipation and partially premixed flame propagation. Combust Sci Technol 2003;175 (4):649–64. [6] Upatnieks A, Driscoll JF, Rasmussen CC, Ceccio SL. Liftoff of turbulent jet flames-assessment of edge flame and other concepts using cinema-PIV. Combust Flame 2004;138(3):259–72. [7] Scholefield DA, Garside JE. The structure and stability of diffusion flames. Sympos Combust Flame Explos Phenom 1949;3(1):102–10. [8] Savas Ö, Gollahalli SR. Stability of lifted laminar round gas-jet flame. J Fluid Mech 1986;165:297–318. [9] Savas Ö, Gollahalli SR. Flow structure in near-nozzle region of gas jet flames. AIAA J 1986;24(7):1137–40. [10] Kim HJ, Choe J, Shin HD. Investigation of the structures of attached and lifted flames with various ignition positions over a triple concentric burner in the hysteresis regime. Energy Fuels 2010;24:324–32. [11] Lin CK, Jeng MS, Chao YC. The stabilization mechanism of the lifted jet diffusion flame in the hysteresis region. Exp Fluids 1993;14:353–65. [12] Terry SD, Lyons KM. Low Reynolds number turbulent lifted flames in high coflow. Combust Sci Technol 2005;177(11):2091–112. [13] Moore NJ, Terry SD, Lyons KM. Flame hysteresis effects in methane jet flames in air-coflow. J Energy Resour – ASME 2011;133(2):022202/1–2/5. [14] Terry SD, Lyons KM. Turbulent lifted flames in the hysteresis regime and the effects of co-flow. J Energy Resour – ASME 2006;128(4):319–24. [15] Brande WT. The bakerian lecture: on some new electro-chemical phenomena. Philos Trans Roy Soc 1814;104:51–61. [16] Won SH, Cha MS, Park CS, Chung SH. Effect of electric fields on reattachment and propagation speed of tribrachial flames in laminar coflow jets. Proc Combust Inst 2007;31(1):963–70. [17] Chung SH. Stabilization, propagation and instability of tribrachial triple flames. Proc Combust Inst 2007;31(1):877–92. [18] Lee BJ, Kim JS, Chung SH. Effect of dilution on the liftoff of non-premixed jet flames. Symp (Int) Combust 1994;25(1):1175–81. [19] Lacoste DA, Xu DA, Moeck JP, Laux CO. Dynamic response of a weakly turbulent lean-premixed flame to nanosecond repetitively pulsed discharges. Proc Combust Inst 2013;34(2):3259–66.
367
[20] Bak MS, Do H, Mungal MG, Cappelli MA. Plasma-assisted stabilization of laminar premixed methane/air flames around the lean flammability limit. Combust Flame 2012;159(10):3128–37. [21] Belhi M, Domingo P, Vervisch P. Direct numerical simulation of the effect of an electric field on flame stability. Combust Flame 2010;157(12):2286–97. [22] Hutchins AR, Reach WA, Kribs JD, Lyons KM. Effects of electric fields on stabilized lifted propane flames. J Energy Resour – ASME 2014;136(2):022203/ 1–022203/10. [23] Kim MK, Ryu SK, Won SH, Chung SH. Electric fields effect on liftoff and blowoff of nonpremixed laminar jet flames in a coflow. Combust Flame 2010;157:17–24. [24] Ombrello TM. Plasma-assisted combustion: systematic decoupling of the kinetic enhancement mechanisms of ignition, flame propagation, and flame stabilization by long-lifetime species [Ph.D. Dissertation]. Princeton University; 2009. [25] Starik AM, Kuleshov PS, Titova NS. Initiation of combustion of a hydrogen-air mixture with ozone impurity by UV laser radiation. Tech Phys 2008;53 (2):235–43. [26] Calcote HF, Pease RN. Electrical properties of flames. Burner flames in longitudinal electric fields. Ind Eng Chem 1951;43(12):2726–31. [27] Vincent-Randonnier A, Larigaldie S, Magre P, Sabel’nikov V. Plasma assisted combustion: effect of a coaxial DBD on a methane diffusion flame. Plasma Sources Sci Technol 2007;16:149–60. [28] Ju Y, Sun W. Plasma assisted combustion: dynamics and chemistry. Prog Energy Combust Sci 2015;48:21–83. [29] Lu XP, Jiang ZH, Xiong Q, Tang ZY, Hu XW, Pan Y. An 11cm long atmospheric pressure cold plasma plume for applications of plasma medicine. Appl Phys Lett 2008;92:081502/1–2/2. [30] Cui T, Tang SL, Zhang C, Yu DR. Hysteresis phenomenon of mode transition in ramjet engines and its topological rule. J Propul Power 2012;28(6):1277–84. [31] Thom R, Brookes WM. Mathematical models of morphogenesis. Chichester, UK: Ellis Horwood Ltd; 1984. p. 105.
Contents lists available at ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
Hysteresis in stabilization of methane diffusion flames with plasmas Jingfeng Tang a,⇑, Liqiu Wei a, Jian Song b, Daren Yu b a b
Academy of Fundamental and Interdisciplinary Sciences, Harbin Institute of Technology, Harbin, PR China School of Energy Science and Engineering, Harbin Institute of Technology, Harbin, PR China
h i g h l i g h t s The liftoff and reattachment velocities are strongly affected by plasmas. The hysteresis regime is reduced and eventually disappeared with the addition of plasmas. A topological approach is used to predict the change of flame structures in hysteresis region.
a r t i c l e
i n f o
Article history: Received 13 November 2015 Received in revised form 18 February 2016 Accepted 29 March 2016 Available online 2 April 2016 Keywords: Methane diffusion flame Stabilization Plasma Hysteresis
a b s t r a c t The stabilization characteristics of liftoff and reattachment in methane diffusion flames have been investigated experimentally with addition of plasmas. The flame liftoff and reattachment velocities have been measured by varying the applied voltages on plasmas. The results show that the liftoff and reattachment velocities are strongly affected by the plasmas. The difference between these two velocities as well as the flame hysteresis region is reduced and eventually disappeared with the addition of plasmas. A topological approach is put forward to explore the mathematic laws of hysteresis phenomena, and the existent hysteresis and catastrophe laws of flame structure transition are interpreted under the typical operation routes. It is concluded that a change of flame stabilization will occur when the fuel velocity or the applied voltage passes through stabilization boundaries in a special direction. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction Diffusion flames are used in many industry applications due to its safe advantages. The stabilization of diffusion flames always exhibit two states: liftoff and attachment state. An attached diffusion flame lifts off the nozzle when the liftoff velocity is reached and then, if the fuel supply is decreased, the flame propagates upstream until the reattachment velocity is attained, at which point the flame quickly reattaches to the nozzle. A comprehensive series of papers correlating analytical solutions with experimental data have been published to predict the stabilization boundaries for a diffusion flame [1–6]. Since the velocity required to lift an attached flame is higher than that needed for reattachment, a stable diffusion flame can exist as liftoff or attachment state at velocities between these values. The hysteresis in flame stabilization refers to the situation that the flame has dual stabilization positions, and is comprised of a set of velocities which include the liftoff velocity and the reattachment velocity. ⇑ Corresponding author. E-mail addresses: [email protected] (J. Tang), [email protected] (L. Wei). http://dx.doi.org/10.1016/j.fuel.2016.03.102 0016-2361/Ó 2016 Elsevier Ltd. All rights reserved.
This hysteresis of diffusion flames was first investigated by Scholefield and Garside, who concluded that the hysteresis could occur with a sudden decrease in height [7]. Savas demonstrated a big difference of flame structures between attachment and liftoff flames, and reported that such difference accounted for the hysteresis phenomena [8,9]. Shin taken a detailed investigation on flame structures in the hysteresis regime, and concluded the ignition position determined whether flames were attached or lifted [10]. Lin present that the flame base was mostly located at the vortex roll-up position, the roll-up of air into vortex inhibited the flame propagation and eventually caused the hysteresis phenomenon [11]. Lyons parametrically studied hysteresis behaviors by using methane diffusion flames under air-coflow conditions, and present that the co-flows could affect the nature of flame stability as well as the flame hysteresis [12–14]. Combustion enhancement by plasmas or electric fields has been investigated extensively for the purpose of improving flame stabilization. The earliest observation of electric field (plasma) interaction with flames was the work of Brande [15]. They found that flames were attached to electrodes due to the existence of charged particles in the flames. Chung [16–18], Lacoste [19], Bak [20], Belhi [21] and Lyons [22] applied AC or DC electric fields to various
363
J. Tang et al. / Fuel 179 (2016) 362–367
flames and demonstrated that the plasmas could impose an important effect on the flame structure. Chung summarized the effect of AC and DC electric fields or plasmas on diffusion flames, and demonstrated a reattachment process from a lifted flame to an attached flame with the existence of electric fields [23]. They concluded that the AC voltage amplitude always posed positive effects on flame enhancements, while the AC alternating frequency always posed positive effect on the laminar liftoff velocity as well as negative effect on the reattachment velocity. The DC voltage amplitude had positive effects for the unsteady flames, while had small effects for stationary flames. The combustion enhancement mechanisms by plasmas have mainly been explained as three effects. First is heating effect caused by plasmas [24], second is the enhancement of reaction kinetics by active and excited species produced by plasmas [25], and the third is mainly associated with ionic wind effects [26]. Recently, Vincent-Randonnier demonstrated that when a DBD plasma was activated on a diffusion methane flame, the liftoff height of the lifted flame was significantly decreased [27]. They reported that the reduction in the flame liftoff/reattachment hysteresis could be caused by the DBD discharge, but that the hysteresis couldn’t be eliminated at their works. Ju provided a comprehensive review about plasma assisted ignition and combustion, and highlighted that since plasmas has greater kinetic effects on ignition, plasma addition could change the flame ignition/extinction S-curve to a monotonic curve in the hysteresis regime [28]. With the addition of plasmas, the stabilization of diffusion flame is needed to understand the possible change of hysteresis characteristics. This present study experimentally focuses on effects of plasmas on diffusion flame stabilization in hysteresis regimes. With an argon plasma injecting into a methane diffusion flame, the diffusion flame in the hysteresis region is examined. Liftoff and reattachment velocities are collected and analyzed for various plasmas. The methane flame hysteresis features are studied based on topological geometry methods, and the topological rules of diffusion flames in the hysteresis regime are provided and also interpreted with the addition of plasmas.
Fig. 1. Schematic diagram of experimental setup.
The setup used for generating and controlling of the discharge is composed of an AC signal generator and an amplifier. The AC generator produces a sinusoidal signal with a low amplitude and with a high frequency. This signal is transmitted to the amplifier for amplitude enhancement. The amplifier output signal is applied to the needle electrode with adjustable voltage amplitudes and repetition frequencies. The current and voltage signals are measured by a Rogowski coil (Pearson 6585) and a high-voltage probe (Tektronix P6015A), and the signals are recorded by a digital oscilloscope (Tektronix DPO4104). Flame images under different experiment conditions are taken by a digital camera (Nikon D7000) with an exposure time up to1/1250 s and a fixed ISO 6400.
2. Experimental setup The schematic diagram of experiments is shown in Fig. 1. A simple floating electrode to generate an atmospheric plasma jet is adopted, which is made of a stainless steel needle with a diameter of 1 mm. When argon is injected into the needle and the alternative voltage (amplitudes up to 20 kV peak-to-peak, repetition rate ranged from 45 kHz to 50 kHz) is applied, a homogeneous plasma jet is generated in front of the end of the needle and into the surrounding air. Such discharge configuration has be shown to concentrate electrical energy into the plasma simply and efficiently [29]. The needle electrode is inserted into a T-shaped quartz tube through a rubble plug, and its tip reaches a same horizontal position with the tube top. The inner and outer diameters of the tube exit are 2 and 4 mm, respectively. The tube top inserted with the needle electrode is acted as fuel jet nozzle, and the mixing between methane and argon occurs at the exit of nozzle. A picture of the fuel jet nozzle is added into Fig. 1 to describe the possible interaction region between diffusion flames and plasmas. When methane is injected from the tube bypass inlet, and a special excitation voltage is applied to the floating electrode, a diffusion methane flame assisted by an argon plasma jet is established in front of the tube exit. With this configuration, the effects of the plasma jet on the flame can be easily observed, and the changes in flame properties (detachment height, flame height, liftoff velocity, etc.) can be quantified.
3. Structure transition of methane diffusion flames with argon plasma jets 3.1. Effect of plasma jets on the flames’ structure The classical characteristics of a diffusion methane flame are observed. The flame is always attached to the tube until the methane jet velocity VJ is up to 23 m/s, and the flame is mostly of yellow emission. At VJ = 10 m/s, the flame base partially detach from the tube. Due to the partially premixing between air and methane at this detached position, the lifted flame is less emissive. HD increases with VJ until the flame is blown off. With the addition of plasma jet, the flame structures for different velocities VJ are photographed in Fig. 2. For an attached flame with an excitation voltage amplitude U around 0–2 kV, the flame height remains unchanged and no apparent phenomenon is observed. With the voltage amplitudes increased up to 2 kV, a luminous plasma jet coming is clearly apparent in the flame bottom, and the flame keep attached structure and shook crazily with less emission. For a lifted flame with an excitation voltage U up to 2 kV, the flame structure is strongly affected by the plasma jet. At VJ = 20 m/s and U = 4 kV, the lifted flame moves upstream and the flame base becomes slanted, as exhibited in Fig. 2e. Even though the flame is partially-attached, the flame edge is relatively smooth without a cusp-like behavior. With the voltage U increased
364
J. Tang et al. / Fuel 179 (2016) 362–367
Fig. 2. Pictures of methane flames with 0.25 s exposure time (a) (VJ [m/s], U [kV]) = (10, 0); (b) (10, 40); (c) (10, 80); (d) (20, 0); (e) (20, 40); (f) (20, 80) (white dotted line: position of tube exit).
up to 8 kV, a luminous plasma jet becomes clearly apparent at the flame bottom, and an electrical contact between the flame and the plasma jet is established, as exhibited in Fig. 2f. The plasma jet goes from the needle to the flame front and expands through it, and the flame serves as an extended electrode. The similar phenomenon is also mentioned by Vincent-Randonnier’s work, and explained as that the high temperature and low density flames facilitate the progression of discharges [27]. With this electrical contact establishment, the flame is anchored to the plasma jet as well as to the tube. The attachment of the flame or the flame anchoring is mainly related with the complicated physical and chemical effects of plasma jets. 3.2. Hysteresis in structure transition of methane diffusion flames with argon plasma jets For a diffusion methane flame without applications of plasma jets, the relationship between the flame detachment distance HD and the jet velocity VJ is plotted in Fig. 3a. There are two data curves with different operation directions, such as, with increasing VJ for the attached flame situation, or with decreasing VJ for the
lifted flame situation. The flame lifts off with VJ increasing up to 23.3 m/s, and the flame reattaches at the VJ of 10.6 m/s. Since the jet velocity VJ required for liftoff an attached flame is higher than that needed for reattachment, a stable flame exists as lifted or attached state at a jet velocity from 10.6 m/s to 23.3 m/s. The hysteresis in flame stabilization refers to the situation that the flame has dual stabilization positions, and is comprised of a set of velocities which include the liftoff velocity and the reattachment velocity. Thus, the structure of the methane flame in this hysteretic region depends more on its prior state than solely on the flow velocity. The similar experiment is repeated at the voltage U of 8 kV. The difference between the liftoff velocity and the reattachment velocity is 3.3 m/s instead of 12.7 m/s for no plasma, as exhibited in Fig. 3b. An attached flame lifts off the tube when the jet velocity is reached at VJ = 20.2 m/s and then, if the jet velocity is decreased, the flame propagated upstream until the jet velocity is attained at VJ = 16.9 m/s, at which point the flame quickly reattaches to the tube. In the existence of plasmas, the flame liftoff occurs at lower VJ than that for U = 0 kV due to plasma effects. With the excitation voltage amplitudes U increased up to 13 kV, an attached flame is kept at the tube exit, especially at high jet velocity with the electrical contact between the flame and the plasma jet. As stated above, the flame is anchored to the plasma jet as well as to the tube, and kept in an attachment state. At this excitation voltage, the flame shook crazily and the blow-off occurs easily. The experiments are repeated and the attachment state is kept beyond the flame blow-off. 4. Mathematical rule: topological property in structure transition of methane diffusion flames with argon plasma jets In this section, the topic is focused on the general rules for structure transition in methane diffusion flames with argon plasma jets. 4.1. Topological invariant rules for the boundaries of structure transition The system outputs can be predicted from system inputs easily for a determined system without hysteresis. But for a system with hysteresis, it is difficult to give the system outputs only with inputs [30]. It means that the operation routes are as important as inputs, to predict the final system outputs. According to the theory of dynamic systems, the critical conditions for mode transition correspond to system singular points, which can be expressed with potential function as [30],
Fig. 3. Hysteresis effect in HD (a) without plasma; (b) with plasma at U = 8 kV.
J. Tang et al. / Fuel 179 (2016) 362–367
) @VðX; UÞ=@X ¼ 0 ðX; UÞ 2 @V ðX; UÞ=@X 2 ¼ 0
(
ð1Þ
where X is the n-dimension state variable, U is the m-dimension input variable, and V is the C1-scalar potential function of the system. Near the singular point, a small change of the system input always result in a sudden jump of the state variable, as well as a output jump containing the significant behaviors of mode transition. According to Thom’s classification theorem, the potential function for a cusp singularity can be provided by [31],
VðX; UÞ ¼ x4 þ u1 x2 þ u2 x:
ð2Þ
The cusp singularity is illustrated in Fig. 4a, which is described by the equilibrium surface under a three-dimensional coordinate space of [u1 u2 x]. The fold edges in the equilibrium surface are called as the singularity set, which can be obtained by the projection of equilibrium surface. If an operation route passes through the singularity set, a catastrophic jump will take place. If the system remains far away the singularity set, the system behavior would vary continuously with inputs. Therefore, the singularity set defines the conditions where a sudden change of the system can occur. The evolution of the flame detachment distance HD versus the jet velocity VJ is plotted in Fig. 4b. It can be seen that the flame stabilization boundary is a cusp-shaped curve, which corresponds to the singularity set for singular systems. The characteristics of singular systems provide a topological rule to understand the flame behaviors. Based on this topological property, it should be possible to predict whether the hysteresis exists, and easy to understand the flame behavior along different operation routes.
365
catastrophe, which correspond to the two stable structures: liftoff and attachment state. In fact, the catastrophe lines correspond to boundaries at which the methane flame will transit between liftoff and attachment state. The catastrophe boundaries can be used to determine when the transitions occur between these two stable structures. As remaining outside the singularity set, the flame behavior varies continuously with (U and VJ). However, the catastrophe and hysteresis are inevitable. 4.2.2. Interpretation of hysteresis and catastrophe behaviors Fig. 5 provides the conditions of mode transition by velocity VJ and voltage U. To investigate flame behaviors, operation routes around the singularity set are described in this figure, and each operation route has two opposite directions. The hysteresis behaviors along routes I have been analyzed and demonstrated in Fig. 3. Fig. 6 shows the changes of HD with U along routes II and III, respectively. The methane flame has exhibited catastrophe behaviors different from hysteresis behaviors. Once the flame structure transits from an attachment state to a liftoff state along route II, it is impossible to transit back to an attachment state. Similarly, it is also impossible to transit back to a liftoff state from an attachment state along route III, if it has previously been reattached to the tube. The corresponding experiments have been investigated to validate the catastrophe characteristic as discussed above. An attached flame with a constant VJ = 20 m/s will transits to a liftoff state while U is applied up to 10 kV, and then it is not possible to transit back to an attachment state even U is deceased at any values. Similarly, an lifted flame with a constant VJ = 15 m/s will transits to an attachment state while U is applied up to 10 kV, and then it is not
4.2. Topological interpretation of structure transition of methane diffusion flames with argon plasma jets 4.2.1. Interpretation of catastrophe boundary As shown in Fig. 4b, the occurrence of flame structure transition corresponds to the catastrophe phenomena. That is, when applied voltage U is low (U = 0 kV, 4.0 kV, 8.0 kV and 10.0 kV), there exists a two-state stable structure (liftoff and attachment state) with the change of jet velocity VJ. However, when U is relatively high (U P 12.0 kV), there exists only one stable structure (attachment state) with the change of VJ, and there is no catastrophe or jump in the process. This is the characteristic of catastrophe. As shown in Fig. 4a of the topological geometry space of catastrophe, u1 and u2 is regarded as the topological transformation of U and VJ, respectively. It can be seen that both the upper and lower branch of the pitchfork have stable equilibrium points in this
Fig. 5. Type of control routes and singularity sets descript by (U, VJ).
Fig. 4. Topological property of structure transition in methane flames with argon plasma jets (brown dotted line: singularity set; green dotted line: stability boundaries). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
366
J. Tang et al. / Fuel 179 (2016) 362–367
Fig. 6. HD versus (U, VJ) along route II and route III.
Fig. 7. Pictures (0.25 s exposure time) of the flame behaviors passing through different routes (a) and (c) VJ = 20 m/s; (b) and (d) VJ = 15 m/s.
possible to transit back from an attachment state to a liftoff state. The flame behaviors are exhibited in Fig. 7 for different control routes. Obviously, the topological property in structure transition is useful to understand and predict the methane flame behavior along different operation routes. 5. Conclusion The stabilization of the methane diffusion flames has been investigated experimentally with the addition of plasmas under a single electrode discharge configuration. The results show that the flame structure is strongly affected by the plasma jet. The difference between liftoff and reattachment velocities is also affected, as well as the flame hysteresis region is reduced and eventually disappeared with applications of plasmas. Once the flame transits from an attachment state to a liftoff state with varying the applied voltage amplitude, it is impossible to transit back from a liftoff state to an attachment state even the plasma jet application is canceled. Similarly, it is impossible for the flame state to transit back, if it has previously been an attachment state from a liftoff state by the applied plasma jet. To understand the hysteresis phenomenon
of the methane flame structure transition, a topological rule is also proposed based on Thom’s classification theorem, which provides the singularity set of structure transition. With such topological rules, the flame stabilization boundaries are obtained, and the existent hysteresis phenomenon are interpreted under typical operation routes passing through the stabilization boundaries. Such general rules in methane flames may be useful for further applications, such as improving or controlling flame stabilization. Acknowledgment This work was supported in part by National Natural Science Foundation of China (51006027, 51437002, and 51477035). References [1] Lyons KM. Toward an understanding of the stabilization mechanisms of lifted turbulent jet flames: experiments. Prog Energy Combust Sci 2007;33:211–31. [2] Chung SH, Lee BJ. On the characteristics of laminar lifted flames in a nonpremixed jet. Combust Flame 1991;86(1–2):62–72. [3] Lee BJ, Chung SH. Determination of schmidt number of mixed fuels by the characteristics of laminar lifted jet flames. Fuel 2006;85:68–74. [4] Han D, Mungal MG. Direct measurement of entrainment in reacting/ nonreacting turbulent jets. Combust Flame 2001;124(3):370–86.
J. Tang et al. / Fuel 179 (2016) 362–367 [5] Watson KA, Lyons KM, Donbar JM, Carter CD. On scalar dissipation and partially premixed flame propagation. Combust Sci Technol 2003;175 (4):649–64. [6] Upatnieks A, Driscoll JF, Rasmussen CC, Ceccio SL. Liftoff of turbulent jet flames-assessment of edge flame and other concepts using cinema-PIV. Combust Flame 2004;138(3):259–72. [7] Scholefield DA, Garside JE. The structure and stability of diffusion flames. Sympos Combust Flame Explos Phenom 1949;3(1):102–10. [8] Savas Ö, Gollahalli SR. Stability of lifted laminar round gas-jet flame. J Fluid Mech 1986;165:297–318. [9] Savas Ö, Gollahalli SR. Flow structure in near-nozzle region of gas jet flames. AIAA J 1986;24(7):1137–40. [10] Kim HJ, Choe J, Shin HD. Investigation of the structures of attached and lifted flames with various ignition positions over a triple concentric burner in the hysteresis regime. Energy Fuels 2010;24:324–32. [11] Lin CK, Jeng MS, Chao YC. The stabilization mechanism of the lifted jet diffusion flame in the hysteresis region. Exp Fluids 1993;14:353–65. [12] Terry SD, Lyons KM. Low Reynolds number turbulent lifted flames in high coflow. Combust Sci Technol 2005;177(11):2091–112. [13] Moore NJ, Terry SD, Lyons KM. Flame hysteresis effects in methane jet flames in air-coflow. J Energy Resour – ASME 2011;133(2):022202/1–2/5. [14] Terry SD, Lyons KM. Turbulent lifted flames in the hysteresis regime and the effects of co-flow. J Energy Resour – ASME 2006;128(4):319–24. [15] Brande WT. The bakerian lecture: on some new electro-chemical phenomena. Philos Trans Roy Soc 1814;104:51–61. [16] Won SH, Cha MS, Park CS, Chung SH. Effect of electric fields on reattachment and propagation speed of tribrachial flames in laminar coflow jets. Proc Combust Inst 2007;31(1):963–70. [17] Chung SH. Stabilization, propagation and instability of tribrachial triple flames. Proc Combust Inst 2007;31(1):877–92. [18] Lee BJ, Kim JS, Chung SH. Effect of dilution on the liftoff of non-premixed jet flames. Symp (Int) Combust 1994;25(1):1175–81. [19] Lacoste DA, Xu DA, Moeck JP, Laux CO. Dynamic response of a weakly turbulent lean-premixed flame to nanosecond repetitively pulsed discharges. Proc Combust Inst 2013;34(2):3259–66.
367
[20] Bak MS, Do H, Mungal MG, Cappelli MA. Plasma-assisted stabilization of laminar premixed methane/air flames around the lean flammability limit. Combust Flame 2012;159(10):3128–37. [21] Belhi M, Domingo P, Vervisch P. Direct numerical simulation of the effect of an electric field on flame stability. Combust Flame 2010;157(12):2286–97. [22] Hutchins AR, Reach WA, Kribs JD, Lyons KM. Effects of electric fields on stabilized lifted propane flames. J Energy Resour – ASME 2014;136(2):022203/ 1–022203/10. [23] Kim MK, Ryu SK, Won SH, Chung SH. Electric fields effect on liftoff and blowoff of nonpremixed laminar jet flames in a coflow. Combust Flame 2010;157:17–24. [24] Ombrello TM. Plasma-assisted combustion: systematic decoupling of the kinetic enhancement mechanisms of ignition, flame propagation, and flame stabilization by long-lifetime species [Ph.D. Dissertation]. Princeton University; 2009. [25] Starik AM, Kuleshov PS, Titova NS. Initiation of combustion of a hydrogen-air mixture with ozone impurity by UV laser radiation. Tech Phys 2008;53 (2):235–43. [26] Calcote HF, Pease RN. Electrical properties of flames. Burner flames in longitudinal electric fields. Ind Eng Chem 1951;43(12):2726–31. [27] Vincent-Randonnier A, Larigaldie S, Magre P, Sabel’nikov V. Plasma assisted combustion: effect of a coaxial DBD on a methane diffusion flame. Plasma Sources Sci Technol 2007;16:149–60. [28] Ju Y, Sun W. Plasma assisted combustion: dynamics and chemistry. Prog Energy Combust Sci 2015;48:21–83. [29] Lu XP, Jiang ZH, Xiong Q, Tang ZY, Hu XW, Pan Y. An 11cm long atmospheric pressure cold plasma plume for applications of plasma medicine. Appl Phys Lett 2008;92:081502/1–2/2. [30] Cui T, Tang SL, Zhang C, Yu DR. Hysteresis phenomenon of mode transition in ramjet engines and its topological rule. J Propul Power 2012;28(6):1277–84. [31] Thom R, Brookes WM. Mathematical models of morphogenesis. Chichester, UK: Ellis Horwood Ltd; 1984. p. 105.