- Email: [email protected]
Comparative study on the laminar flame speeds of methylcyclohexane-methanol and toluene-methanol blends at elevated temperatures
Fuel 245 (2019) 534–543
Contents lists available at ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
Full Length Article
Comparative study on the laminar flame speeds of methylcyclohexanemethanol and toluene-methanol blends at elevated temperatures
T
Yemiao Zhang, Qianqian Li , Hu Liu, Zhiyu Yan, Zuohua Huang ⁎
State Key Laboratory of Multiphase Flow in Power Engineering, Xi'an Jiaotong University, Xi'an, People’s Republic of China
ARTICLE INFO
ABSTRACT
Keywords: Methanol Methylcyclohexane Toluene Laminar flame speed Kinetic modeling
Laminar flame speeds of methylcyclohexane (MCH)-methanol and toluene-methanol blends were experimentally determined with spherically expanding flame method in a constant volume bomb at atmospheric pressure, initial temperatures of 393 and 433 K, covering wide equivalence ratio range. The blending ratio of methanol in liquid volume varies as 0%, 20%, 40%, 60%, 80%, 100%. Nonlinear methodology was employed to remove the stretch effect in the data processing. Experimental results show that MCH-air flame propagates faster than toluene at the same condition. The addition of methanol into MCH and toluene results in acceleration of laminar flame speed especially at the rich mixtures. Since the published model suffers difficulty in reproducing experimental data, model refinements were carried out and the refined model yields better performance. Comprehensive analyses were developed regarding thermal and chemical kinetic properties. For MCH has lower adiabatic flame temperature than toluene and methanol addition into the two cyclic fuels decreases the adiabatic flame temperature, the thermal effect on laminar flame speed difference is negligible. Therefore, the effect of chemical kinetics was specifically discussed. MCH and toluene have different ring structures. The disintegration of aromatic ring plays as the limiting step in the high-temperature oxidation of toluene, resulting in low concentration of active radical pool and overall reactivity. However, the ring opening reaction of MCH occurs easily after the initial H-abstraction reaction, which favors the production of active intermediates and the enhancement of flame propagation. For the blending fuels, the analyses show that the laminar flame speed variation of the blends are primarily caused by the methanol substitution and the disturbance of reaction pathway through affecting the generation of important intermediates.
1. Introduction
the chemical and physical properties in flame propagation. It is the basis of turbulent flame study and can be employed to evaluate the accuracy of chemical kinetic model [8,9]. Until now, a variety of work have been conducted to understand the laminar flame characteristics of methanol blending fuels. Gülder et al. [10] and Beeckmann et al. [11] both measured the laminar flame speed of methanol-isooctane blends at atmospheric pressure and elevated temperatures. However, their conclusions are inconsistent. Gülder et al. [10] found the methanol-isooctane blends yield lower laminar flame speeds than both pure methanol and isooctane, while Beeckmann et al. [11] indicated that methanol-isooctane blends exhibit lower laminar flame speeds than methanol but higher values than isooctane. Li et al. [12] developed a comprehensive laminar flame study on methanol-isooctane blends, and found the blends have higher laminar flame speeds than isooctane and lower values than methanol, which supports the result of Beeckmann et al. [11]. Besides, the data of Li et al. [12] were adopted to aid the model development of methanol-isooctane. Sileghem et al. [13]
Methanol is considered as potential alternative fuels since it shows competitive performance regarding the concerns of broad sources preparation, high octane number and eco-friendly [1–3]. However, there also exists problems of high specific fuel consumption and cold start when using methanol in engines due to its low energy density and high latent heat. Consequently, methanol is extensively utilized as an additive into gasoline or diesel in nowadays. Cyclic alkanes and aromatics take high proportion in traditional fuels [4,5], among which MCH and toluene are of specific interest since they are always selected to represent cyclic alkane and aromatic components in surrogate fuels [6,7]. Therefore, a comprehensive fundamental study on the combustion characteristics of MCH-methanol and toluene-methanol blends will promote the understanding of methanol blending fuels as well as their practical use. Laminar flame speed is a representative parameter characterizing
⁎
Corresponding author at: State Key Laboratory of Multiphase Flow in Power Engineering, Xi'an Jiaotong University, Xi'an 710049, People’s Republic of China. E-mail address: [email protected] (Q. Li).
https://doi.org/10.1016/j.fuel.2019.02.091 Received 10 July 2018; Received in revised form 12 January 2019; Accepted 18 February 2019 Available online 26 February 2019 0016-2361/ © 2019 Elsevier Ltd. All rights reserved.
Fuel 245 (2019) 534–543
Y. Zhang, et al.
Fig. 1. Comparison between present data of MCH, toluene and previous data at atmospheric pressure.
determined the laminar flame speeds of methanol-isooctane and methanol-n-heptane blends using the heat flux method. Based on the experimental data, the mixing rules for the blends were validated. Mohamad et al. [14] measured the laminar flame speeds of the mixtures of methanol, isooctane and indolene at elevated temperature and pressure, and proposed empirical equations. It is found published work mostly concern the effect of methanol addition to chain alkanes, while limited data were reported for the blends of methanol and cyclic hydrocarbons. MCH differs from toluene in the unsaturation degree of ring structure. The unsaturation degree of ring structure has been suggested to contribute soot formation and auto-ignition differently [15–17], and it is speculated to cause deviation on combustion behavior. Actually, Law et al. [18] conducted the effect of unsaturation degree by investigating the laminar flames of ethane, ethylene and acetylene. They found acetylene yields the fastest laminar flame speed, then ethylene, and the flame of ethane propagates the slowest. This laminar flame speed difference is found to result from both adiabatic temperature and the chemical structure. To our knowledge, limited comparable investigation has been published on laminar flames of cyclic hydrocarbons, however the effect of different unsaturation degree was rarely concerned. Very recently, Zhong et al. [19] compared three typical C7 fuels including n-heptane, MCH and toluene in terms of laminar flame speeds and found Su0, n hep tan e > Su0, MCH > Su0, toluene at all experimental conditions. The main reason was interpreted as the different H/C ratio and hence different chemistries of H radical production and consumption through comparing the H concentrations in three fuel flames. However,
Fig. 2. Experimental (symbols) and computed (lines) laminar flame speeds of (a) methanol, (b) toluene, (c) MCH at 0.1 MPa.
the inherent kinetic effect resulted from the unsaturation degree needs to be further discussed for detailed kinetic analyses were not provided. Ji et al. [20] measured the laminar flame speeds of the blends of 80% ndodecane + 20% MCH and 80% n-dodecane + 20% toluene with the counterflow configuration. This study found MCH has faster laminar flame speed than toluene but the reason was not illustrated. Most previous research focused on the comparison of MCH, toluene with their corresponding mono-alkylated derivatives respectively. Wu et al. [21] and Ji et al. [22] determined the laminar flame speeds of 535
Fuel 245 (2019) 534–543
Y. Zhang, et al.
Table 1 The modified reactions and the corresponding constant rate. Reaction
A factor
n
EA
Reference
A1CH3 + O < = > A1CH2 + OH A1CH3 + H < = > A1CH3*+H2 A1CH3 + O < = > A1CH3*+OH A1CH3 + OH < = > A1CH3*+H2O A1CH3 + CH3 < = > A1CH3*+CH4 A1CH3 + HO2 < = > A1CH3*+H2O2 A1CH3*+O2 < = > OC6H4O + CH3
6.300e + 11 5.000e + 8 1.660e + 13 1.330e + 8 1.600e + 12 3.33e + 11 3.0e + 13
0.0 1.0 0.0 1.4 0.0 0.0 0.0
0.0 16,800 14,700 1450 15,000 28,900 9.0
46 44a 44a 44a 44a 44a 45
a
A factor reduced by 1/6 to account for one less abstraction site.
cyclohexane and mono-alkylated cyclohexane at different initial conditions, and found increasing the branching degree inhibits flame propagation. Davis et al. [23] and Wang et al. [24] compared the laminar flame speeds of benzene and toluene and found toluene propagates slower than benzene. Present study aims to provide the experimental laminar flame speeds of MCH-methanol and toluene-methanol blends at different initial conditions and methanol volume blending ratios. These measurements allow us to identity the order of laminar flame speed of MCH and toluene, as well as the laminar flame speed variation with methanol addition into different cyclic hydrocarbons. On the basis of present data, a comprehensive kinetic model published previously was refined to achieve higher accuracy. Detailed analyses regarding thermodynamics and chemistry were performed to clarify the inherent reason leading to the laminar flame speed deviation between MCH and toluene as well as the effect of methanol addition. 2. Experimental method and numerical simulation 2.1. Experimental apparatus In this study, laminar flame speeds were determined using a constant combustion bomb with the details described in previous literatures [25,26]. Here just a brief introduction is provided for clarity. The bomb is cylindrical with a heating-tap surrounded to heat the chamber and the inner mixture. The temperature was monitored using a K-type thermocouple with the accuracy of 2–3 K and the pressure was calibrated with the pressure transmitter and transducer. The combustible mixture was prepared based on partial pressure law. Before each experiment, the chamber was vacuumed and heated up to experimental initial temperature. When the liquid fuels were injected into the chamber, at least five mins was awaited to ensure the full evaporation. Then the air composing of 21% oxygen and 79% nitrogen was delivered into the chamber. Another five mins was awaited to achieve homogenous mixture prior to ignition. The mixture was ignited by tungsten electrodes located in the center of the chamber and the high speed camera was utilized to record the flame propagation. Nonlinear method developed by Frankel and Sivashinsky [27,28] was adopted to remove the stretch effect on the flame surface with the method expressed as, Sb = Sb0 Sb0 Lb ·2/ Rf , where Lb is the Markstein length of the burned mixture, Sb is the stretched laminar flame speed. Finally, the laminar flame speed could be determined according to mass conservation, Su0 = b Sb0/ u , where b and u are the density of the burned and unburned mixtures, respectively. In the laminar flame speed measurement, uncertainty exists due to the effects of ignition, chamber confinement, flame instability, radiation, etc. Burke et al. [29] suggested that the confinement effect is negligible when the flame radius is smaller than 0.3 times of the chamber wall radius. Considering the ignition effect as well, the flame radii between 8 and 22 mm was employed to process data. Present experiments were performed at atmospheric pressure, and the effect of flame instability was not taken into account. The radiation-induced uncertainty was corrected with the correlation proposed by Yu et al.
Fig. 3. Su0 of (a) MCH, (b) toluene and (c) methanol. Solid lines: simulation using the NN-model; Dash line: simulation using N-model; Symbols: experimental data.
[30]. Comprehensive uncertainty analysis was estimated using the methods proposed by Moffat et al. [31], and the uncertainty was approximate ± 2–4 cm/s according to the flame conditions. 2.2. Computational methods Laminar flame speeds were simulated using CHEMKIN PRO software [32]. The mixture-averaged model was used and the Soret effect was taken into consideration. GRAD and CURV values were set as 0.03 and 0.02, respectively, and the maximum grid points was set to be 536
Fuel 245 (2019) 534–543
Y. Zhang, et al.
developed by Narayanaswamy et al. [34] aiming at transportation fuel surrogates. It contains the low and high temperature chemistry of methylcyclohexane, n-dodecane and aromatic species. C-model indicates the diesel surrogate fuel model developed by Chang et al. [35]. This model is a skeletal model and consists of 70 species and 220 reactions. 3. Results and discussion 3.1. Validation of present experimental apparatus To test the performance of present experimental apparatus, Fig. 1 lists present data and the literature data measured over a wide temperature range from 353 to 470 K. It is seen that all data follow the rule of temperature dependence that Su0 at higher temperature is faster, though slight differences are observed among the data at the same temperature measured by different groups. In Fig. 1a, Su0 of MCH measured by Rotavera et al. [36] at 403 K is lower than the results of Ji et al. [20] at 403 K, Kumar et al. [37] and Munzar et al. [38] at 400 K. The main reason is that Rotavera et al. [36] adopted the spherical propagation method, which is different from three other groups using the counterflow configuration. From this figure, it can also be realized that present data reach the peak around equivalence ratio of 1.1, which is consistent with the data of most groups [21,36–38] except the results of Ji et al. [20,22] which give the peak value around equivalence ratio of 1.05 both at 353 and 403 K. Fig. 1b gives present data of toluene-air mixtures along with the data published in previous researches. For the data of Wang et al. [24] were obtained with the constant-volume bomb which is the same as present study, the data derived by Wang et al. [24] at 423 K were adjusted to make direct comparison with present data at 433 K by removing the 10 K temperature difference through the empirical correlation of Su0 Tu1.8 [39]. The result indicates that the adjusted data is in good agreement with present data. And as expected, Su0 measured in present work at 393 K are slightly lower than those reported by Davis et al. [23], Hui et al. [40], Kumar et al. [37] at 400 K and by Dirrenberger et al. [41] at 398 K. 3.2. Model evaluation and refinement Present data were adopted to evaluate the performance of the proposed models. Methanol is a small molecular fuel, and its model always plays as the base model of large hydrocarbons. Simulations of methanol were performed using the three models mentioned above. Fig. 2a plots the experimental and computed Su0 versus equivalence ratio at 393 K for methanol. It can be seen the simulations of N-model agree well with present data, falling within the uncertainty range of the experimental data. C-model yields satisfactory predictions at lean and rich sides, but
Fig. 4. Ignition delay times for toluene. Symbols – experiments: Vasudevan et al. [48]; solid lines – results from the numerical simulations of NN-model; dash lines – simulations result from N-model.
1500. Three chemical kinetic models were selected to simulate the blending fuels for they include the chemical information of MCH, toluene and methanol. JetSurF2.0 [33], hereafter referred to as J-model, contains 348 species and 2163 reactions and includes the pyrolysis and high-temperature oxidation kinetics of jet-fuel surrogate. The main species in J-model are C1–C12 n-alkanes, cycloalkanes (cyclohexane, methyl-, ethyl-, n-propyl-, and n-butyl-cyclohexane), and aromatics (benzene and toluene). N-model represents the comprehensive model
Fig. 5. Comparison between present Su0 of MCH and toluene. 537
Fuel 245 (2019) 534–543
Y. Zhang, et al.
Fig. 6. Reaction path analyses of (a) MCH, (b) toluene at ϕ = 1.5, Tu = 393 K, and Pu = 0.1 MPa. The numbers indicate the branching ratios.
slightly over predicts the experimental data around stoichiometric ratio. J-model yields reasonable predictions at lean mixtures, however significantly over predicts the experimental data at the equivalence ratio over 1.0. Considering the poor prediction ability of J-model, Nmodel and C-model were selected to do the further validation. Fig. 2b and c plot the comparison between the experimental and simulated laminar flame speeds of toluene and MCH, respectively. The simulations of N-model are indicated as solid lines and those of C-model are indicated as dash lines. For toluene in Fig. 2b, N-model reasonably predicts the experimental data at lean and rich mixture conditions but yields over-predictions around stoichiometric equivalence ratio. Similar result was also reported when the toluene sub-model was proposed by Narayanaswamy et al. [42] that the simulated laminar flame speeds of
toluene are noticeably higher than the experimental data measured by Johnston and Farrel [43]. C-model over predicts the experimental data at all equivalence ratio range with the deviation especially greater at the lean and rich mixtures. For MCH in Fig. 2c, it is found that the simulations of C-model reasonably agree with the experimental data. The simulations of N-model fall within the experimental uncertainty range at the lean and rich mixtures but are beyond the uncertainty range around stoichiometric equivalence ratio. Therefore, both Nmodel and C-model have difficulty in accurately simulating the flames of MCH, toluene and methanol simultaneously. Considering N-model includes more detailed chemistry information than C-model, N-model is selected to be improved for further analysis. It is obviously that the sub-model of toluene in N-model is the key 538
Fuel 245 (2019) 534–543
Y. Zhang, et al.
Fig. 8. Adiabatic flame temperature of (a) MCH-methanol blends (b) toluenemethanol blends versus equivalence ratio at 393 K and 0.1 MPa. Fig. 7. Laminar flame speeds of (a) MCH-methanol (b) toluene-methanol blends at different blending ratios (Symbols: experimental data; Lines: simulations of NN-model).
experimental ignition delay times measured by Vasudevan et al. [48], as seen in Fig. 4. It is observed that NN-model yields better predictions on the experimental data at ϕ = 0.5 than N-model, and comparable predictions at 1.0 and 1.5. Thus, NN-model was adopted to do further kinetic analysis.
point to be refined. After examining N-model in details, problems are found as follows. Wang et al. [24] indicated that reactions related to fuel-specific species play important roles in laminar flame propagation except small molecule species. Methylphenyl radical(C6H4CH3) was initially generated from the initial H abstraction reactions and the rate constants of those reactions in N-model were obtained by analogy with those of benzene forming phenyl radicals. Considering the structure difference between toluene and benzene, it is more rational to reduce the A factor of analogy reactions of benzene by 1/6 to account for five abstraction sites [44,45]. Moreover, the rate constant of a dominant reaction, which is toluene (C6H5CH3) converting to benzyl(C6H5CH2), C6H5CH3 + O=C6H5CH2 + OH shows big uncertainty for its exact origin is unclear. Thus, the value was updated with the experimental data given by Hoffmann et al. [46]. In addition, according to the research of Yuan et al. [47] and Bounaceur et al. [45], the reactions producing stable benzoquinonc (o-C6H4O2) radicals from toluene are lost in N-model and should be included in the refined model. All modified reactions are shown in Table 1 and the refined kinetic model is named NN-model. Fig. 3 depicts Su0 of MCH, toluene and methanol experimentally and simulated with NN-model and N-model at 0.1 MPa. It is seen that NNmodel yields much better predictions than N-model for toluene, and satisfactorily predicts the experimental data of MCH and methanol as well. The accuracy of NN-model is further verified by simulating the
3.3. Analyses on laminar flame speed variation 3.3.1. Comparison between MCH and toluene Fig. 5 gives the comparison between present experimental Su0 of MCH and toluene at 393 and 433 K. It is shown that the two fuels reach the peak value at equivalence ratio around 1.1 and MCH presents faster laminar flame speeds than toluene, especially near stoichiometric equivalence ratio, which is consistent with the conclusion of Zhong et al. [19]. Zhong et al. also indicated that the laminar flame speed deviation between MCH and toluene is dominated by kinetic effect for toluene has higher adiabatic flame temperature than MCH, but limited kinetic analyses were presented. Therefore, the reaction pathway analyses were carried out to illustrate the inherent reason, as seen in Fig. 6. Reaction pathway analyses were carried out with NN-model for MCH and toluene, as seen in Fig. 6. At high temperature, MCH primarily undergoes H abstraction reactions to generate five isomerized methylcyclohexyl radicals differentiated by the carbon sites abstracted. Then the ring opening reaction occurs subsequently in most of these radicals, producing the straight and branched chain intermediates. These intermediates finally crack into various small radicals, including large amount of ethylene (C2H4) and vinyl radical (C2H3). These two radicals have been concluded to be active and enhance the overall 539
Fuel 245 (2019) 534–543
Y. Zhang, et al.
reaction rate [18], thus their high concentrations are positive for the flame propagation of MCH. However, for toluene flame in Fig. 6b, it is seen that toluene is difficult to open the ring due to the stable chemical structure and the ring opening process is undoubtedly the limiting step in terms of toluene decomposing at high temperature. Specifically, toluene initially goes to benzene (C6H6), benzyl radical (C6H5CH2) and methylphenyl (C6H4CH3). And then almost 80% of benzene, over 10% of benzyl radical and other intermediate radicals, like benzaldehyde (C6H5CHO), will form phenyl radical(C6H5). The reaction rates of C6H5 related reactions are much slower when compared to those of intermediates in MCH flames, retarding the flame propagation of toluene. In addition, the chain termination reactions in toluene flame especially the reaction of C6H5CH2 + H < = > C6H5CH3, further decrease Su0 . Moreover, it is observed there are considerable amount of propargyl (C3H3) generated. This species participates in several chain terminating reactions such as C3H3 + H=pC3H4 and put an adverse impact on the flame propagation [20]. As a consequence, there are a variety of active species generated in MCH flame while relative stable species produced in toluene flame, causing significant difference of laminar flame speed between MCH and toluene.
3.3.2. Analyses on laminar flame speed variation of blending fuels Fig. 7 gives the comparison between the measured laminar flame speeds and simulations for the MCH-methanol and toluene-methanol blends at different blending ratios. In general, the simulation results of NN-model reasonably agree with the experimental data and accurately capture the laminar flame speed variation. With the methanol addition, the laminar flame speed increases and the increment is remarkable at the equivalence ratio over 1.0 and methanol blending ratio over 40%. When the blending ratio of methanol is less than 40%, the laminar flame speed increases slightly. At the equivalence ratio smaller than 1.0, the laminar flame speed varies little and even slight decrease is observed at high methanol blending ratio for MCH-methanol blends. To investigate the reason leading to above change on laminar flame speed, the adiabatic temperature (Tad) of the blends were calculated. As shown in Fig. 8, for both MCH-methanol and toluene-methanol blends, Tad presents decreasing trend with the addition of methanol, which is different from the laminar flame speed variation, demonstrating the thermal effect is not the dominant cause. As concluded in the previous study [12], methanol is a high-activity fuel. At high temperature, methanol primarily goes to HCO through H abstraction reactions step by step and finally release high concentration of H radical pool. With the addition of methanol, the H radical pool is
Fig. 9. Reaction path analyses of (a) MCH-methanol, (b) methanol-toluene at ϕ = 1.5, Tu = 393 K, and Pu = 0.1 MPa at M20 (normal text), M50 (italic text) and M80 (bond text). The numbers indicate the branching ratios. 540
Fuel 245 (2019) 534–543
Y. Zhang, et al.
Fig. 9. (continued)
enriched correspondingly. Moreover, as one important intermediate, methanol addition will exert an effect on the reaction process of MCH and toluene. Fig. 9a and b shows the reaction pathways of MCH-methanol and toluene-methanol blends respectively at ϕ = 1.5 and different methanol volume blending ratios (M20, M50, M80). It is observed that the cracking procedures of MCH and toluene in the blends are almost the same with their pure fuel as illustrated in details above. With the increase of methanol blending ratio, the branching ratios of the initial reactions vary little. For MCH-methanol blends, the reactions producing C2 and C3 radicals are influenced with the branching ratios changed. Vinyl (C2H3) and ethylene (C2H4), as mentioned previously, exert positive effect on the propagation of flame. On the contrary, allyl (C3H5) and propene can participate chain terminating reactions of aC3H5 + H + M=C3H6 + M and C3H6 + H=aC3H5 + H2, decreasing the concentration of H radical. As shown in Fig. 10, the concentrations of C2H3, C2H4, aC3H5 and C3H6 all decrease with the addition of methanol. The competition among those radicals related reactions finally characterizes as the concentration variation of reactive radicals like H and OH. As shown in Fig. 10c, the addition of methanol significantly increased the concentration of H and OH, demonstrating the increase of overall reactivity. As shown in Fig. 9b, for toluene-methanol blends, the methanol addition exerts inhibiting effect on the production of 5-member ring, such as cyclopentadiene(C5H6), fulvene(Cy-C5H4=CH2) and
cyclopentadiene-1-one(C5H4=O), which are regarded as the main resources of C3H3. As mentioned above, C3H3 plays negative effect on flame propagation through participating chain terminating reactions. Fig. 11 shows the mole fractions of C3H3 produced in the flames of toluene-methanol blends. It is found less C3H3 are generated with more methanol blended, demonstrating that the addition of methanol inhibits propargyl production and prompt the overall reactivity in toluenemethanol blends. Therefore, the kinetic effect of methanol addition on the laminar flame propagation primarily results from both methanol substitution and the disturbance of chemical reaction process. 4. Conclusions A comprehensive study was conducted experimentally and numerically on the laminar flames of MCH-methanol and toluene-methanol blends to investigate the effect of unsaturation degree and the addition of methanol on laminar flame speed. Laminar flame speeds were determined at atmospheric pressure and elevated temperatures, over a wide range of equivalence ratio. The experimental data were employed to validate and improve the accuracy of published models. And the refined model is demonstrated to yield satisfactory predictions on present data and ignition delay times published previously. With the refined model, detailed analyses were conducted to illustrate the inherent reason causing the laminar flame speed variation. 541
Fuel 245 (2019) 534–543
Y. Zhang, et al.
Fig. 11. Concentrations of C3H3 in the flames of toluene-methanol blends at different methanol blending ratios.
specifically discussed. Reaction path analyses revealed that the aromatic ring in toluene is much more stable than the ring in MCH, resulting in more difficulty in opening the ring for aromatic ring. Besides, the production of phenyl radicals and other lower-activity radicals in toluene flame also exert an inhibiting effect on the flame propagation. Therefore, the chemical structure difference finally causes the laminar flame speed deviation between toluene and MCH. The kinetic effect of methanol addition into MCH and toluene on flame propagation is primarily due to methanol substitution and the disturbance of chemical reaction process. Compared to MCH and toluene, methanol is much more active with high concentrations of H radical produced in methanol flame. With the addition of methanol, the H radical pool is enriched correspondingly. Besides, the methanol addition changes the branching ratios of reactions generating C2, C3 and the other active species, which ultimately influences the concentrations of H and OH and hence the flame propagation process. Acknowledgements This work is supported by the National Natural Science Foundation of China (Grant No. 51776163, 51406159 and 91441203), the Postdoctoral Science Foundation of Shaanxi Province, and State Key Laboratory of Engines, Tianjin University (K2016-02) Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.fuel.2019.02.091. References [1] Sarathy SM, Oßwald P, Hansen N, Kohse-Höinghaus K. Alcohol combustion chemistry. Prog Energy Combust Sci 2014;44:40–102. [2] Li J, Gong C, Wang E, Yu X, Wang Z, Liu X. Emissions of formaldehyde and unburned methanol from a spark-ignition methanol engine during cold start. Energy Fuels 2010;24(2):863–70. [3] Ninomiya JS, Golovoy A, Labana SS. Effect of methanol on exhaust composition of a fuel containing toluene, n-heptane, and isooctane. Air Repair 1970;20(5):314. [4] Edwards T. Liquid fuels and propellants for aerospace propulsion: 1903–2003. J Propul Power 2003;19(6):1089–107. [5] Briker Y, Ring Z, Iacchelli A, Mclean NPMR, Fairbridge C, et al. Diesel fuel analysis by gc−fims: aromatics, n-paraffins, and isoparaffins. Energy Fuels 2016;15(1):23–37. [6] AIAA. Development of an experimental database and kinetic models for surrogate jet fuels. SAE Technical Papers 2007. [7] Dooley S, Sang HW, Chaos M, Heyne J, Ju Y, Dryer FL, et al. A jet fuel surrogate formulated by real fuel properties. Combust Flame 2010;157(12):2333–9. [8] Driscoll JF. Turbulent premixed combustion: flamelet structure and its effect on turbulent burning velocities. Prog Energy Combust Sci 2008;34(1):91–134. [9] Ranzi E, Frassoldati A, Grana R, Cuoci A, Faravelli T, Kelley AP, et al. Hierarchical and comparative kinetic modeling of laminar flame speeds of hydrocarbon and
Fig. 10. Concentrations of selected intermediates in the flames of MCH-methanol blends at different methanol blending ratios.
Present experimental data and simulation results all show that the laminar flame speeds of MCH are higher than toluene. With more methanol added into MCH and toluene, laminar flame speed increases especially at rich mixture side and methanol blending ratio over 40%. At the lean mixtures, laminar flame speed varies little and even slight decrease can be observed at high methanol blending ratio for MCHmethanol blends. For toluene has higher adiabatic temperature than MCH and methanol addition into MCH and toluene will decrease the adiabatic temperature, the thermal effect is excluded as the main reason causing the laminar flame speed variation. Thus, the kinetic effect was 542
Fuel 245 (2019) 534–543
Y. Zhang, et al. oxygenated fuels. Prog Energy Combust Sci 2012;38(4):468–501. [10] Gulder ÃML. Laminar burning velocities of methanol, isooctane and isooctane/ methanol blends. Combust Sci Technol 1983;33(1–4):179–92. [11] Beeckmann J, Röhl O, Peters N. Experimental and numerical investigation of isooctane, methanol and ethanol regarding laminar burning velocity at elevated pressure and temperature. SAE Int 2009. [12] Li Q, Zhang W, Jin W, Xie Y, Huang Z. Laminar flame characteristics and kinetic modeling study of methanol-isooctane blends at elevated temperatures. Fuel 2016;184:836–45. [13] Sileghem L, Alekseev VA, Vancoillie J, Nilsson EJK, Verhelst S, Konnov AA. Laminar burning velocities of primary reference fuels and simple alcohols. Fuel 2014;115(1):32–40. [14] Metghalchi M, Keck JC. Burning velocities of mixtures of air with methanol, isooctane, and indolene at high pressure and temperature. Combust Flame 1982;48(82):191–210. [15] Wang Z, Li L, Wang J, Reitz RD. Effect of biodiesel saturation on soot formation in diesel engines. Fuel 2016;175:240–8. [16] Westbrook CK, Pitz WJ, Sarathy SM, Mehl M. Detailed chemical kinetic modeling of the effects of cc double bonds on the ignition of biodiesel fuels. Proc Combust Inst 2013;34(2):3049–56. [17] Zhang K, Togbé C, Dayma G, Dagaut P. Experimental and kinetic modeling study of trans-methyl-3-hexenoate oxidation in jsr and the role of cc double bond. Combust Flame 2014;161(3):818–25. [18] Davis SG, Law CK. Determination of and fuel structure effects on laminar flame speeds of c1 to c8 hydrocarbons. Combust Sci Technol 1998;140(1–6):427–49. [19] Zhong BJ, Peng HS, Zheng D. The effect of different class of hydrocarbons on laminar flame speeds of three c 7 fuels. Fuel 2018;225:225–9. [20] Ji C, Egolfopoulos FN. Flame propagation of mixtures of air with binary liquid fuel mixtures. Proc Combust Inst 2011;33(1):955–61. [21] Wu F, Kelley AP, Law CK. Laminar flame speeds of cyclohexane and mono-alkylated cyclohexanes at elevated pressures. Combust Flame 2012;159(4):1417–25. [22] Ji C, Dames E, Sirjean B, Wang H, Egolfopoulos FN. An experimental and modeling study of the propagation of cyclohexane and mono-alkylated cyclohexane flames. Proc Combust Inst 2011;33(1):971–8. [23] Davis SG, Wang H, Breinsky K, Law CK. Laminar flame speeds and oxidation kinetics of benene-air and toluene-air flames. Symp Combust 1996;26(1):1025–33. [24] Wang G, Li Y, Yuan W, Zhou Z, Wang Y, Wang Z. Investigation on laminar burning velocities of benzene, toluene and ethylbenzene up to 20 atm. Combust Flame 2017;184:312–23. [25] Tang C, Huang Z, Wang J, Zheng J. Effects of hydrogen addition on cellular instabilities of the spherically expanding propane flames. Int J Hydrogen Energy 2009;34(5):2483–7. [26] Hu EJ, Huang ZH, He JJ, Jin C, Miao HY, Wang XB. Measurements of laminar burning velocities and flame stability analysis for hydrogen-air-diluent mixtures. Chin Sci Bull 2009;54(5):846–57. [27] Frankel ML, Sivashinsky GI. On effects due to thermal expansion and lewis number in spherical flame propagation. Combust Sci Technol 1983;31(3–4):131–8. [28] Chen Z. On the extraction of laminar flame speed and markstein length from outwardly propagating spherical flames. Combust Flame 2011;158(2):291–300. [29] Burke MP, Chen Z, Ju Y, Dryer FL. Effect of cylindrical confinement on the determination of laminar flame speeds using outwardly propagating flames. Combust
Flame 2009;156(4):771–9. [30] Yu H, Han W, Santner J, Gou X, Sohn CH, Ju Y, et al. Radiation-induced uncertainty in laminar flame speed measured from propagating spherical flames. Combust Flame 2014;161(11):2815–24. [31] Moffat RJ. Describing the uncertainties in experimental results. Exp Therm Fluid Sci 1988;1(1):3–17. [32] Chemkin-pro, reaction design inc., san diego, ca 92121, USA, Available from: http://www.reactiondesign.com. [33] Wang H, Dames E, Sirjean B. A high-temperature chemical kinetic model of n-alkane (up to n-dodecane), cyclohexane, and methyl-, ethyl-, n-propyl and n-butylcyclohexane oxidation at high temperatures, jetsurf version 2.0, 2010. Available from: http://web.stanford.edu/group/haiwanglab/JetSurF/JetSurF2.0/index.html. [34] Narayanaswamy K, Pitsch H, Pepiot P. A chemical mechanism for low to high temperature oxidation of methylcyclohexane as a component of transportation fuel surrogates. Combust Flame 2015;162(4):1193–213. [35] Chang Y, Jia M, Li Y, Liu Y, Xie M, Wang H, et al. Development of a skeletal mechanism for diesel surrogate fuel by using a decoupling methodology. Combust Flame 2015;162(10):3785–802. [36] Rotavera B, Krejci M, Vissotski A, Petersen E. Laminar flame speed measurements of methyl octanoate, n-nonane, and methylcyclohexane, 2013. [37] Kumar K, Sung C-J. Flame propagation and extinction characteristics of neat surrogate fuel components. Energy Fuels 2010;24(7):3840–9. [38] Munzar JD, Akih-Kumgeh B, Denman BM, Zia A, Bergthorson JM. An experimental and reduced modeling study of the laminar flame speed of jet fuel surrogate components. Fuel 2013;113:586–97. [39] Wu F, Law CK. An experimental and mechanistic study on the laminar flame speed, markstein length and flame chemistry of the butanol isomers. Combust Flame 2013;160(12):2744–56. [40] Hui X, Das AK, Kumar K, Sung C-J, Dooley S, Dryer FL. Laminar flame speeds and extinction stretch rates of selected aromatic hydrocarbons. Fuel 2012;97:695–702. [41] Dirrenberger P, Glaude PA, Bounaceur R, Gall HL, Cruz APD, Konnov AA, et al. Laminar burning velocity of gasolines with addition of ethanol. Fuel 2014;115(1):162–9. [42] Narayanaswamy K, Blanquart G, Pitsch H. A consistent chemical mechanism for oxidation of substituted aromatic species. Combust Flame 2010;157(10):1879–98. [43] Johnston RJ, Farrell JT. Laminar burning velocities and markstein lengths of aromatics at elevated temperature and pressure. Proc Combust Inst 2005;30(1):217–24. [44] Costa ID, Fournet R, Billaud F, Battin-Leclerc F. Experimental and modeling study of the oxidation of benzene. Int J Chem Kinet 2003;35(10):503–24. [45] Bounaceur R, Da Costa I, Fournet R, Billaud F, Battin-Leclerc F. Experimental and modeling study of the oxidation of toluene. Int J Chem Kinet 2005;37(1):25–49. [46] Hoffmann BA, Klatt M, Wagner HG. An investigation of the faction between 0(3p) and toluene at high temperatures. Z Phys Chem 1990;168(Part_1):1–12. [47] Yuan W, Li Y, Dagaut P, Yang J, Qi F. Investigation on the pyrolysis and oxidation of toluene over a wide range conditions. Ii. A comprehensive kinetic modeling study. Combust Flame 2015;162(1):22–40. [48] Vasudevan V, Davidson DF, Hanson RK. Shock tube measurements of toluene ignition times and oh concentration time histories. Proc Combust Inst 2005;30(1):1155–63.
543
Contents lists available at ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
Full Length Article
Comparative study on the laminar flame speeds of methylcyclohexanemethanol and toluene-methanol blends at elevated temperatures
T
Yemiao Zhang, Qianqian Li , Hu Liu, Zhiyu Yan, Zuohua Huang ⁎
State Key Laboratory of Multiphase Flow in Power Engineering, Xi'an Jiaotong University, Xi'an, People’s Republic of China
ARTICLE INFO
ABSTRACT
Keywords: Methanol Methylcyclohexane Toluene Laminar flame speed Kinetic modeling
Laminar flame speeds of methylcyclohexane (MCH)-methanol and toluene-methanol blends were experimentally determined with spherically expanding flame method in a constant volume bomb at atmospheric pressure, initial temperatures of 393 and 433 K, covering wide equivalence ratio range. The blending ratio of methanol in liquid volume varies as 0%, 20%, 40%, 60%, 80%, 100%. Nonlinear methodology was employed to remove the stretch effect in the data processing. Experimental results show that MCH-air flame propagates faster than toluene at the same condition. The addition of methanol into MCH and toluene results in acceleration of laminar flame speed especially at the rich mixtures. Since the published model suffers difficulty in reproducing experimental data, model refinements were carried out and the refined model yields better performance. Comprehensive analyses were developed regarding thermal and chemical kinetic properties. For MCH has lower adiabatic flame temperature than toluene and methanol addition into the two cyclic fuels decreases the adiabatic flame temperature, the thermal effect on laminar flame speed difference is negligible. Therefore, the effect of chemical kinetics was specifically discussed. MCH and toluene have different ring structures. The disintegration of aromatic ring plays as the limiting step in the high-temperature oxidation of toluene, resulting in low concentration of active radical pool and overall reactivity. However, the ring opening reaction of MCH occurs easily after the initial H-abstraction reaction, which favors the production of active intermediates and the enhancement of flame propagation. For the blending fuels, the analyses show that the laminar flame speed variation of the blends are primarily caused by the methanol substitution and the disturbance of reaction pathway through affecting the generation of important intermediates.
1. Introduction
the chemical and physical properties in flame propagation. It is the basis of turbulent flame study and can be employed to evaluate the accuracy of chemical kinetic model [8,9]. Until now, a variety of work have been conducted to understand the laminar flame characteristics of methanol blending fuels. Gülder et al. [10] and Beeckmann et al. [11] both measured the laminar flame speed of methanol-isooctane blends at atmospheric pressure and elevated temperatures. However, their conclusions are inconsistent. Gülder et al. [10] found the methanol-isooctane blends yield lower laminar flame speeds than both pure methanol and isooctane, while Beeckmann et al. [11] indicated that methanol-isooctane blends exhibit lower laminar flame speeds than methanol but higher values than isooctane. Li et al. [12] developed a comprehensive laminar flame study on methanol-isooctane blends, and found the blends have higher laminar flame speeds than isooctane and lower values than methanol, which supports the result of Beeckmann et al. [11]. Besides, the data of Li et al. [12] were adopted to aid the model development of methanol-isooctane. Sileghem et al. [13]
Methanol is considered as potential alternative fuels since it shows competitive performance regarding the concerns of broad sources preparation, high octane number and eco-friendly [1–3]. However, there also exists problems of high specific fuel consumption and cold start when using methanol in engines due to its low energy density and high latent heat. Consequently, methanol is extensively utilized as an additive into gasoline or diesel in nowadays. Cyclic alkanes and aromatics take high proportion in traditional fuels [4,5], among which MCH and toluene are of specific interest since they are always selected to represent cyclic alkane and aromatic components in surrogate fuels [6,7]. Therefore, a comprehensive fundamental study on the combustion characteristics of MCH-methanol and toluene-methanol blends will promote the understanding of methanol blending fuels as well as their practical use. Laminar flame speed is a representative parameter characterizing
⁎
Corresponding author at: State Key Laboratory of Multiphase Flow in Power Engineering, Xi'an Jiaotong University, Xi'an 710049, People’s Republic of China. E-mail address: [email protected] (Q. Li).
https://doi.org/10.1016/j.fuel.2019.02.091 Received 10 July 2018; Received in revised form 12 January 2019; Accepted 18 February 2019 Available online 26 February 2019 0016-2361/ © 2019 Elsevier Ltd. All rights reserved.
Fuel 245 (2019) 534–543
Y. Zhang, et al.
Fig. 1. Comparison between present data of MCH, toluene and previous data at atmospheric pressure.
determined the laminar flame speeds of methanol-isooctane and methanol-n-heptane blends using the heat flux method. Based on the experimental data, the mixing rules for the blends were validated. Mohamad et al. [14] measured the laminar flame speeds of the mixtures of methanol, isooctane and indolene at elevated temperature and pressure, and proposed empirical equations. It is found published work mostly concern the effect of methanol addition to chain alkanes, while limited data were reported for the blends of methanol and cyclic hydrocarbons. MCH differs from toluene in the unsaturation degree of ring structure. The unsaturation degree of ring structure has been suggested to contribute soot formation and auto-ignition differently [15–17], and it is speculated to cause deviation on combustion behavior. Actually, Law et al. [18] conducted the effect of unsaturation degree by investigating the laminar flames of ethane, ethylene and acetylene. They found acetylene yields the fastest laminar flame speed, then ethylene, and the flame of ethane propagates the slowest. This laminar flame speed difference is found to result from both adiabatic temperature and the chemical structure. To our knowledge, limited comparable investigation has been published on laminar flames of cyclic hydrocarbons, however the effect of different unsaturation degree was rarely concerned. Very recently, Zhong et al. [19] compared three typical C7 fuels including n-heptane, MCH and toluene in terms of laminar flame speeds and found Su0, n hep tan e > Su0, MCH > Su0, toluene at all experimental conditions. The main reason was interpreted as the different H/C ratio and hence different chemistries of H radical production and consumption through comparing the H concentrations in three fuel flames. However,
Fig. 2. Experimental (symbols) and computed (lines) laminar flame speeds of (a) methanol, (b) toluene, (c) MCH at 0.1 MPa.
the inherent kinetic effect resulted from the unsaturation degree needs to be further discussed for detailed kinetic analyses were not provided. Ji et al. [20] measured the laminar flame speeds of the blends of 80% ndodecane + 20% MCH and 80% n-dodecane + 20% toluene with the counterflow configuration. This study found MCH has faster laminar flame speed than toluene but the reason was not illustrated. Most previous research focused on the comparison of MCH, toluene with their corresponding mono-alkylated derivatives respectively. Wu et al. [21] and Ji et al. [22] determined the laminar flame speeds of 535
Fuel 245 (2019) 534–543
Y. Zhang, et al.
Table 1 The modified reactions and the corresponding constant rate. Reaction
A factor
n
EA
Reference
A1CH3 + O < = > A1CH2 + OH A1CH3 + H < = > A1CH3*+H2 A1CH3 + O < = > A1CH3*+OH A1CH3 + OH < = > A1CH3*+H2O A1CH3 + CH3 < = > A1CH3*+CH4 A1CH3 + HO2 < = > A1CH3*+H2O2 A1CH3*+O2 < = > OC6H4O + CH3
6.300e + 11 5.000e + 8 1.660e + 13 1.330e + 8 1.600e + 12 3.33e + 11 3.0e + 13
0.0 1.0 0.0 1.4 0.0 0.0 0.0
0.0 16,800 14,700 1450 15,000 28,900 9.0
46 44a 44a 44a 44a 44a 45
a
A factor reduced by 1/6 to account for one less abstraction site.
cyclohexane and mono-alkylated cyclohexane at different initial conditions, and found increasing the branching degree inhibits flame propagation. Davis et al. [23] and Wang et al. [24] compared the laminar flame speeds of benzene and toluene and found toluene propagates slower than benzene. Present study aims to provide the experimental laminar flame speeds of MCH-methanol and toluene-methanol blends at different initial conditions and methanol volume blending ratios. These measurements allow us to identity the order of laminar flame speed of MCH and toluene, as well as the laminar flame speed variation with methanol addition into different cyclic hydrocarbons. On the basis of present data, a comprehensive kinetic model published previously was refined to achieve higher accuracy. Detailed analyses regarding thermodynamics and chemistry were performed to clarify the inherent reason leading to the laminar flame speed deviation between MCH and toluene as well as the effect of methanol addition. 2. Experimental method and numerical simulation 2.1. Experimental apparatus In this study, laminar flame speeds were determined using a constant combustion bomb with the details described in previous literatures [25,26]. Here just a brief introduction is provided for clarity. The bomb is cylindrical with a heating-tap surrounded to heat the chamber and the inner mixture. The temperature was monitored using a K-type thermocouple with the accuracy of 2–3 K and the pressure was calibrated with the pressure transmitter and transducer. The combustible mixture was prepared based on partial pressure law. Before each experiment, the chamber was vacuumed and heated up to experimental initial temperature. When the liquid fuels were injected into the chamber, at least five mins was awaited to ensure the full evaporation. Then the air composing of 21% oxygen and 79% nitrogen was delivered into the chamber. Another five mins was awaited to achieve homogenous mixture prior to ignition. The mixture was ignited by tungsten electrodes located in the center of the chamber and the high speed camera was utilized to record the flame propagation. Nonlinear method developed by Frankel and Sivashinsky [27,28] was adopted to remove the stretch effect on the flame surface with the method expressed as, Sb = Sb0 Sb0 Lb ·2/ Rf , where Lb is the Markstein length of the burned mixture, Sb is the stretched laminar flame speed. Finally, the laminar flame speed could be determined according to mass conservation, Su0 = b Sb0/ u , where b and u are the density of the burned and unburned mixtures, respectively. In the laminar flame speed measurement, uncertainty exists due to the effects of ignition, chamber confinement, flame instability, radiation, etc. Burke et al. [29] suggested that the confinement effect is negligible when the flame radius is smaller than 0.3 times of the chamber wall radius. Considering the ignition effect as well, the flame radii between 8 and 22 mm was employed to process data. Present experiments were performed at atmospheric pressure, and the effect of flame instability was not taken into account. The radiation-induced uncertainty was corrected with the correlation proposed by Yu et al.
Fig. 3. Su0 of (a) MCH, (b) toluene and (c) methanol. Solid lines: simulation using the NN-model; Dash line: simulation using N-model; Symbols: experimental data.
[30]. Comprehensive uncertainty analysis was estimated using the methods proposed by Moffat et al. [31], and the uncertainty was approximate ± 2–4 cm/s according to the flame conditions. 2.2. Computational methods Laminar flame speeds were simulated using CHEMKIN PRO software [32]. The mixture-averaged model was used and the Soret effect was taken into consideration. GRAD and CURV values were set as 0.03 and 0.02, respectively, and the maximum grid points was set to be 536
Fuel 245 (2019) 534–543
Y. Zhang, et al.
developed by Narayanaswamy et al. [34] aiming at transportation fuel surrogates. It contains the low and high temperature chemistry of methylcyclohexane, n-dodecane and aromatic species. C-model indicates the diesel surrogate fuel model developed by Chang et al. [35]. This model is a skeletal model and consists of 70 species and 220 reactions. 3. Results and discussion 3.1. Validation of present experimental apparatus To test the performance of present experimental apparatus, Fig. 1 lists present data and the literature data measured over a wide temperature range from 353 to 470 K. It is seen that all data follow the rule of temperature dependence that Su0 at higher temperature is faster, though slight differences are observed among the data at the same temperature measured by different groups. In Fig. 1a, Su0 of MCH measured by Rotavera et al. [36] at 403 K is lower than the results of Ji et al. [20] at 403 K, Kumar et al. [37] and Munzar et al. [38] at 400 K. The main reason is that Rotavera et al. [36] adopted the spherical propagation method, which is different from three other groups using the counterflow configuration. From this figure, it can also be realized that present data reach the peak around equivalence ratio of 1.1, which is consistent with the data of most groups [21,36–38] except the results of Ji et al. [20,22] which give the peak value around equivalence ratio of 1.05 both at 353 and 403 K. Fig. 1b gives present data of toluene-air mixtures along with the data published in previous researches. For the data of Wang et al. [24] were obtained with the constant-volume bomb which is the same as present study, the data derived by Wang et al. [24] at 423 K were adjusted to make direct comparison with present data at 433 K by removing the 10 K temperature difference through the empirical correlation of Su0 Tu1.8 [39]. The result indicates that the adjusted data is in good agreement with present data. And as expected, Su0 measured in present work at 393 K are slightly lower than those reported by Davis et al. [23], Hui et al. [40], Kumar et al. [37] at 400 K and by Dirrenberger et al. [41] at 398 K. 3.2. Model evaluation and refinement Present data were adopted to evaluate the performance of the proposed models. Methanol is a small molecular fuel, and its model always plays as the base model of large hydrocarbons. Simulations of methanol were performed using the three models mentioned above. Fig. 2a plots the experimental and computed Su0 versus equivalence ratio at 393 K for methanol. It can be seen the simulations of N-model agree well with present data, falling within the uncertainty range of the experimental data. C-model yields satisfactory predictions at lean and rich sides, but
Fig. 4. Ignition delay times for toluene. Symbols – experiments: Vasudevan et al. [48]; solid lines – results from the numerical simulations of NN-model; dash lines – simulations result from N-model.
1500. Three chemical kinetic models were selected to simulate the blending fuels for they include the chemical information of MCH, toluene and methanol. JetSurF2.0 [33], hereafter referred to as J-model, contains 348 species and 2163 reactions and includes the pyrolysis and high-temperature oxidation kinetics of jet-fuel surrogate. The main species in J-model are C1–C12 n-alkanes, cycloalkanes (cyclohexane, methyl-, ethyl-, n-propyl-, and n-butyl-cyclohexane), and aromatics (benzene and toluene). N-model represents the comprehensive model
Fig. 5. Comparison between present Su0 of MCH and toluene. 537
Fuel 245 (2019) 534–543
Y. Zhang, et al.
Fig. 6. Reaction path analyses of (a) MCH, (b) toluene at ϕ = 1.5, Tu = 393 K, and Pu = 0.1 MPa. The numbers indicate the branching ratios.
slightly over predicts the experimental data around stoichiometric ratio. J-model yields reasonable predictions at lean mixtures, however significantly over predicts the experimental data at the equivalence ratio over 1.0. Considering the poor prediction ability of J-model, Nmodel and C-model were selected to do the further validation. Fig. 2b and c plot the comparison between the experimental and simulated laminar flame speeds of toluene and MCH, respectively. The simulations of N-model are indicated as solid lines and those of C-model are indicated as dash lines. For toluene in Fig. 2b, N-model reasonably predicts the experimental data at lean and rich mixture conditions but yields over-predictions around stoichiometric equivalence ratio. Similar result was also reported when the toluene sub-model was proposed by Narayanaswamy et al. [42] that the simulated laminar flame speeds of
toluene are noticeably higher than the experimental data measured by Johnston and Farrel [43]. C-model over predicts the experimental data at all equivalence ratio range with the deviation especially greater at the lean and rich mixtures. For MCH in Fig. 2c, it is found that the simulations of C-model reasonably agree with the experimental data. The simulations of N-model fall within the experimental uncertainty range at the lean and rich mixtures but are beyond the uncertainty range around stoichiometric equivalence ratio. Therefore, both Nmodel and C-model have difficulty in accurately simulating the flames of MCH, toluene and methanol simultaneously. Considering N-model includes more detailed chemistry information than C-model, N-model is selected to be improved for further analysis. It is obviously that the sub-model of toluene in N-model is the key 538
Fuel 245 (2019) 534–543
Y. Zhang, et al.
Fig. 8. Adiabatic flame temperature of (a) MCH-methanol blends (b) toluenemethanol blends versus equivalence ratio at 393 K and 0.1 MPa. Fig. 7. Laminar flame speeds of (a) MCH-methanol (b) toluene-methanol blends at different blending ratios (Symbols: experimental data; Lines: simulations of NN-model).
experimental ignition delay times measured by Vasudevan et al. [48], as seen in Fig. 4. It is observed that NN-model yields better predictions on the experimental data at ϕ = 0.5 than N-model, and comparable predictions at 1.0 and 1.5. Thus, NN-model was adopted to do further kinetic analysis.
point to be refined. After examining N-model in details, problems are found as follows. Wang et al. [24] indicated that reactions related to fuel-specific species play important roles in laminar flame propagation except small molecule species. Methylphenyl radical(C6H4CH3) was initially generated from the initial H abstraction reactions and the rate constants of those reactions in N-model were obtained by analogy with those of benzene forming phenyl radicals. Considering the structure difference between toluene and benzene, it is more rational to reduce the A factor of analogy reactions of benzene by 1/6 to account for five abstraction sites [44,45]. Moreover, the rate constant of a dominant reaction, which is toluene (C6H5CH3) converting to benzyl(C6H5CH2), C6H5CH3 + O=C6H5CH2 + OH shows big uncertainty for its exact origin is unclear. Thus, the value was updated with the experimental data given by Hoffmann et al. [46]. In addition, according to the research of Yuan et al. [47] and Bounaceur et al. [45], the reactions producing stable benzoquinonc (o-C6H4O2) radicals from toluene are lost in N-model and should be included in the refined model. All modified reactions are shown in Table 1 and the refined kinetic model is named NN-model. Fig. 3 depicts Su0 of MCH, toluene and methanol experimentally and simulated with NN-model and N-model at 0.1 MPa. It is seen that NNmodel yields much better predictions than N-model for toluene, and satisfactorily predicts the experimental data of MCH and methanol as well. The accuracy of NN-model is further verified by simulating the
3.3. Analyses on laminar flame speed variation 3.3.1. Comparison between MCH and toluene Fig. 5 gives the comparison between present experimental Su0 of MCH and toluene at 393 and 433 K. It is shown that the two fuels reach the peak value at equivalence ratio around 1.1 and MCH presents faster laminar flame speeds than toluene, especially near stoichiometric equivalence ratio, which is consistent with the conclusion of Zhong et al. [19]. Zhong et al. also indicated that the laminar flame speed deviation between MCH and toluene is dominated by kinetic effect for toluene has higher adiabatic flame temperature than MCH, but limited kinetic analyses were presented. Therefore, the reaction pathway analyses were carried out to illustrate the inherent reason, as seen in Fig. 6. Reaction pathway analyses were carried out with NN-model for MCH and toluene, as seen in Fig. 6. At high temperature, MCH primarily undergoes H abstraction reactions to generate five isomerized methylcyclohexyl radicals differentiated by the carbon sites abstracted. Then the ring opening reaction occurs subsequently in most of these radicals, producing the straight and branched chain intermediates. These intermediates finally crack into various small radicals, including large amount of ethylene (C2H4) and vinyl radical (C2H3). These two radicals have been concluded to be active and enhance the overall 539
Fuel 245 (2019) 534–543
Y. Zhang, et al.
reaction rate [18], thus their high concentrations are positive for the flame propagation of MCH. However, for toluene flame in Fig. 6b, it is seen that toluene is difficult to open the ring due to the stable chemical structure and the ring opening process is undoubtedly the limiting step in terms of toluene decomposing at high temperature. Specifically, toluene initially goes to benzene (C6H6), benzyl radical (C6H5CH2) and methylphenyl (C6H4CH3). And then almost 80% of benzene, over 10% of benzyl radical and other intermediate radicals, like benzaldehyde (C6H5CHO), will form phenyl radical(C6H5). The reaction rates of C6H5 related reactions are much slower when compared to those of intermediates in MCH flames, retarding the flame propagation of toluene. In addition, the chain termination reactions in toluene flame especially the reaction of C6H5CH2 + H < = > C6H5CH3, further decrease Su0 . Moreover, it is observed there are considerable amount of propargyl (C3H3) generated. This species participates in several chain terminating reactions such as C3H3 + H=pC3H4 and put an adverse impact on the flame propagation [20]. As a consequence, there are a variety of active species generated in MCH flame while relative stable species produced in toluene flame, causing significant difference of laminar flame speed between MCH and toluene.
3.3.2. Analyses on laminar flame speed variation of blending fuels Fig. 7 gives the comparison between the measured laminar flame speeds and simulations for the MCH-methanol and toluene-methanol blends at different blending ratios. In general, the simulation results of NN-model reasonably agree with the experimental data and accurately capture the laminar flame speed variation. With the methanol addition, the laminar flame speed increases and the increment is remarkable at the equivalence ratio over 1.0 and methanol blending ratio over 40%. When the blending ratio of methanol is less than 40%, the laminar flame speed increases slightly. At the equivalence ratio smaller than 1.0, the laminar flame speed varies little and even slight decrease is observed at high methanol blending ratio for MCH-methanol blends. To investigate the reason leading to above change on laminar flame speed, the adiabatic temperature (Tad) of the blends were calculated. As shown in Fig. 8, for both MCH-methanol and toluene-methanol blends, Tad presents decreasing trend with the addition of methanol, which is different from the laminar flame speed variation, demonstrating the thermal effect is not the dominant cause. As concluded in the previous study [12], methanol is a high-activity fuel. At high temperature, methanol primarily goes to HCO through H abstraction reactions step by step and finally release high concentration of H radical pool. With the addition of methanol, the H radical pool is
Fig. 9. Reaction path analyses of (a) MCH-methanol, (b) methanol-toluene at ϕ = 1.5, Tu = 393 K, and Pu = 0.1 MPa at M20 (normal text), M50 (italic text) and M80 (bond text). The numbers indicate the branching ratios. 540
Fuel 245 (2019) 534–543
Y. Zhang, et al.
Fig. 9. (continued)
enriched correspondingly. Moreover, as one important intermediate, methanol addition will exert an effect on the reaction process of MCH and toluene. Fig. 9a and b shows the reaction pathways of MCH-methanol and toluene-methanol blends respectively at ϕ = 1.5 and different methanol volume blending ratios (M20, M50, M80). It is observed that the cracking procedures of MCH and toluene in the blends are almost the same with their pure fuel as illustrated in details above. With the increase of methanol blending ratio, the branching ratios of the initial reactions vary little. For MCH-methanol blends, the reactions producing C2 and C3 radicals are influenced with the branching ratios changed. Vinyl (C2H3) and ethylene (C2H4), as mentioned previously, exert positive effect on the propagation of flame. On the contrary, allyl (C3H5) and propene can participate chain terminating reactions of aC3H5 + H + M=C3H6 + M and C3H6 + H=aC3H5 + H2, decreasing the concentration of H radical. As shown in Fig. 10, the concentrations of C2H3, C2H4, aC3H5 and C3H6 all decrease with the addition of methanol. The competition among those radicals related reactions finally characterizes as the concentration variation of reactive radicals like H and OH. As shown in Fig. 10c, the addition of methanol significantly increased the concentration of H and OH, demonstrating the increase of overall reactivity. As shown in Fig. 9b, for toluene-methanol blends, the methanol addition exerts inhibiting effect on the production of 5-member ring, such as cyclopentadiene(C5H6), fulvene(Cy-C5H4=CH2) and
cyclopentadiene-1-one(C5H4=O), which are regarded as the main resources of C3H3. As mentioned above, C3H3 plays negative effect on flame propagation through participating chain terminating reactions. Fig. 11 shows the mole fractions of C3H3 produced in the flames of toluene-methanol blends. It is found less C3H3 are generated with more methanol blended, demonstrating that the addition of methanol inhibits propargyl production and prompt the overall reactivity in toluenemethanol blends. Therefore, the kinetic effect of methanol addition on the laminar flame propagation primarily results from both methanol substitution and the disturbance of chemical reaction process. 4. Conclusions A comprehensive study was conducted experimentally and numerically on the laminar flames of MCH-methanol and toluene-methanol blends to investigate the effect of unsaturation degree and the addition of methanol on laminar flame speed. Laminar flame speeds were determined at atmospheric pressure and elevated temperatures, over a wide range of equivalence ratio. The experimental data were employed to validate and improve the accuracy of published models. And the refined model is demonstrated to yield satisfactory predictions on present data and ignition delay times published previously. With the refined model, detailed analyses were conducted to illustrate the inherent reason causing the laminar flame speed variation. 541
Fuel 245 (2019) 534–543
Y. Zhang, et al.
Fig. 11. Concentrations of C3H3 in the flames of toluene-methanol blends at different methanol blending ratios.
specifically discussed. Reaction path analyses revealed that the aromatic ring in toluene is much more stable than the ring in MCH, resulting in more difficulty in opening the ring for aromatic ring. Besides, the production of phenyl radicals and other lower-activity radicals in toluene flame also exert an inhibiting effect on the flame propagation. Therefore, the chemical structure difference finally causes the laminar flame speed deviation between toluene and MCH. The kinetic effect of methanol addition into MCH and toluene on flame propagation is primarily due to methanol substitution and the disturbance of chemical reaction process. Compared to MCH and toluene, methanol is much more active with high concentrations of H radical produced in methanol flame. With the addition of methanol, the H radical pool is enriched correspondingly. Besides, the methanol addition changes the branching ratios of reactions generating C2, C3 and the other active species, which ultimately influences the concentrations of H and OH and hence the flame propagation process. Acknowledgements This work is supported by the National Natural Science Foundation of China (Grant No. 51776163, 51406159 and 91441203), the Postdoctoral Science Foundation of Shaanxi Province, and State Key Laboratory of Engines, Tianjin University (K2016-02) Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.fuel.2019.02.091. References [1] Sarathy SM, Oßwald P, Hansen N, Kohse-Höinghaus K. Alcohol combustion chemistry. Prog Energy Combust Sci 2014;44:40–102. [2] Li J, Gong C, Wang E, Yu X, Wang Z, Liu X. Emissions of formaldehyde and unburned methanol from a spark-ignition methanol engine during cold start. Energy Fuels 2010;24(2):863–70. [3] Ninomiya JS, Golovoy A, Labana SS. Effect of methanol on exhaust composition of a fuel containing toluene, n-heptane, and isooctane. Air Repair 1970;20(5):314. [4] Edwards T. Liquid fuels and propellants for aerospace propulsion: 1903–2003. J Propul Power 2003;19(6):1089–107. [5] Briker Y, Ring Z, Iacchelli A, Mclean NPMR, Fairbridge C, et al. Diesel fuel analysis by gc−fims: aromatics, n-paraffins, and isoparaffins. Energy Fuels 2016;15(1):23–37. [6] AIAA. Development of an experimental database and kinetic models for surrogate jet fuels. SAE Technical Papers 2007. [7] Dooley S, Sang HW, Chaos M, Heyne J, Ju Y, Dryer FL, et al. A jet fuel surrogate formulated by real fuel properties. Combust Flame 2010;157(12):2333–9. [8] Driscoll JF. Turbulent premixed combustion: flamelet structure and its effect on turbulent burning velocities. Prog Energy Combust Sci 2008;34(1):91–134. [9] Ranzi E, Frassoldati A, Grana R, Cuoci A, Faravelli T, Kelley AP, et al. Hierarchical and comparative kinetic modeling of laminar flame speeds of hydrocarbon and
Fig. 10. Concentrations of selected intermediates in the flames of MCH-methanol blends at different methanol blending ratios.
Present experimental data and simulation results all show that the laminar flame speeds of MCH are higher than toluene. With more methanol added into MCH and toluene, laminar flame speed increases especially at rich mixture side and methanol blending ratio over 40%. At the lean mixtures, laminar flame speed varies little and even slight decrease can be observed at high methanol blending ratio for MCHmethanol blends. For toluene has higher adiabatic temperature than MCH and methanol addition into MCH and toluene will decrease the adiabatic temperature, the thermal effect is excluded as the main reason causing the laminar flame speed variation. Thus, the kinetic effect was 542
Fuel 245 (2019) 534–543
Y. Zhang, et al. oxygenated fuels. Prog Energy Combust Sci 2012;38(4):468–501. [10] Gulder ÃML. Laminar burning velocities of methanol, isooctane and isooctane/ methanol blends. Combust Sci Technol 1983;33(1–4):179–92. [11] Beeckmann J, Röhl O, Peters N. Experimental and numerical investigation of isooctane, methanol and ethanol regarding laminar burning velocity at elevated pressure and temperature. SAE Int 2009. [12] Li Q, Zhang W, Jin W, Xie Y, Huang Z. Laminar flame characteristics and kinetic modeling study of methanol-isooctane blends at elevated temperatures. Fuel 2016;184:836–45. [13] Sileghem L, Alekseev VA, Vancoillie J, Nilsson EJK, Verhelst S, Konnov AA. Laminar burning velocities of primary reference fuels and simple alcohols. Fuel 2014;115(1):32–40. [14] Metghalchi M, Keck JC. Burning velocities of mixtures of air with methanol, isooctane, and indolene at high pressure and temperature. Combust Flame 1982;48(82):191–210. [15] Wang Z, Li L, Wang J, Reitz RD. Effect of biodiesel saturation on soot formation in diesel engines. Fuel 2016;175:240–8. [16] Westbrook CK, Pitz WJ, Sarathy SM, Mehl M. Detailed chemical kinetic modeling of the effects of cc double bonds on the ignition of biodiesel fuels. Proc Combust Inst 2013;34(2):3049–56. [17] Zhang K, Togbé C, Dayma G, Dagaut P. Experimental and kinetic modeling study of trans-methyl-3-hexenoate oxidation in jsr and the role of cc double bond. Combust Flame 2014;161(3):818–25. [18] Davis SG, Law CK. Determination of and fuel structure effects on laminar flame speeds of c1 to c8 hydrocarbons. Combust Sci Technol 1998;140(1–6):427–49. [19] Zhong BJ, Peng HS, Zheng D. The effect of different class of hydrocarbons on laminar flame speeds of three c 7 fuels. Fuel 2018;225:225–9. [20] Ji C, Egolfopoulos FN. Flame propagation of mixtures of air with binary liquid fuel mixtures. Proc Combust Inst 2011;33(1):955–61. [21] Wu F, Kelley AP, Law CK. Laminar flame speeds of cyclohexane and mono-alkylated cyclohexanes at elevated pressures. Combust Flame 2012;159(4):1417–25. [22] Ji C, Dames E, Sirjean B, Wang H, Egolfopoulos FN. An experimental and modeling study of the propagation of cyclohexane and mono-alkylated cyclohexane flames. Proc Combust Inst 2011;33(1):971–8. [23] Davis SG, Wang H, Breinsky K, Law CK. Laminar flame speeds and oxidation kinetics of benene-air and toluene-air flames. Symp Combust 1996;26(1):1025–33. [24] Wang G, Li Y, Yuan W, Zhou Z, Wang Y, Wang Z. Investigation on laminar burning velocities of benzene, toluene and ethylbenzene up to 20 atm. Combust Flame 2017;184:312–23. [25] Tang C, Huang Z, Wang J, Zheng J. Effects of hydrogen addition on cellular instabilities of the spherically expanding propane flames. Int J Hydrogen Energy 2009;34(5):2483–7. [26] Hu EJ, Huang ZH, He JJ, Jin C, Miao HY, Wang XB. Measurements of laminar burning velocities and flame stability analysis for hydrogen-air-diluent mixtures. Chin Sci Bull 2009;54(5):846–57. [27] Frankel ML, Sivashinsky GI. On effects due to thermal expansion and lewis number in spherical flame propagation. Combust Sci Technol 1983;31(3–4):131–8. [28] Chen Z. On the extraction of laminar flame speed and markstein length from outwardly propagating spherical flames. Combust Flame 2011;158(2):291–300. [29] Burke MP, Chen Z, Ju Y, Dryer FL. Effect of cylindrical confinement on the determination of laminar flame speeds using outwardly propagating flames. Combust
Flame 2009;156(4):771–9. [30] Yu H, Han W, Santner J, Gou X, Sohn CH, Ju Y, et al. Radiation-induced uncertainty in laminar flame speed measured from propagating spherical flames. Combust Flame 2014;161(11):2815–24. [31] Moffat RJ. Describing the uncertainties in experimental results. Exp Therm Fluid Sci 1988;1(1):3–17. [32] Chemkin-pro, reaction design inc., san diego, ca 92121, USA, Available from: http://www.reactiondesign.com. [33] Wang H, Dames E, Sirjean B. A high-temperature chemical kinetic model of n-alkane (up to n-dodecane), cyclohexane, and methyl-, ethyl-, n-propyl and n-butylcyclohexane oxidation at high temperatures, jetsurf version 2.0, 2010. Available from: http://web.stanford.edu/group/haiwanglab/JetSurF/JetSurF2.0/index.html. [34] Narayanaswamy K, Pitsch H, Pepiot P. A chemical mechanism for low to high temperature oxidation of methylcyclohexane as a component of transportation fuel surrogates. Combust Flame 2015;162(4):1193–213. [35] Chang Y, Jia M, Li Y, Liu Y, Xie M, Wang H, et al. Development of a skeletal mechanism for diesel surrogate fuel by using a decoupling methodology. Combust Flame 2015;162(10):3785–802. [36] Rotavera B, Krejci M, Vissotski A, Petersen E. Laminar flame speed measurements of methyl octanoate, n-nonane, and methylcyclohexane, 2013. [37] Kumar K, Sung C-J. Flame propagation and extinction characteristics of neat surrogate fuel components. Energy Fuels 2010;24(7):3840–9. [38] Munzar JD, Akih-Kumgeh B, Denman BM, Zia A, Bergthorson JM. An experimental and reduced modeling study of the laminar flame speed of jet fuel surrogate components. Fuel 2013;113:586–97. [39] Wu F, Law CK. An experimental and mechanistic study on the laminar flame speed, markstein length and flame chemistry of the butanol isomers. Combust Flame 2013;160(12):2744–56. [40] Hui X, Das AK, Kumar K, Sung C-J, Dooley S, Dryer FL. Laminar flame speeds and extinction stretch rates of selected aromatic hydrocarbons. Fuel 2012;97:695–702. [41] Dirrenberger P, Glaude PA, Bounaceur R, Gall HL, Cruz APD, Konnov AA, et al. Laminar burning velocity of gasolines with addition of ethanol. Fuel 2014;115(1):162–9. [42] Narayanaswamy K, Blanquart G, Pitsch H. A consistent chemical mechanism for oxidation of substituted aromatic species. Combust Flame 2010;157(10):1879–98. [43] Johnston RJ, Farrell JT. Laminar burning velocities and markstein lengths of aromatics at elevated temperature and pressure. Proc Combust Inst 2005;30(1):217–24. [44] Costa ID, Fournet R, Billaud F, Battin-Leclerc F. Experimental and modeling study of the oxidation of benzene. Int J Chem Kinet 2003;35(10):503–24. [45] Bounaceur R, Da Costa I, Fournet R, Billaud F, Battin-Leclerc F. Experimental and modeling study of the oxidation of toluene. Int J Chem Kinet 2005;37(1):25–49. [46] Hoffmann BA, Klatt M, Wagner HG. An investigation of the faction between 0(3p) and toluene at high temperatures. Z Phys Chem 1990;168(Part_1):1–12. [47] Yuan W, Li Y, Dagaut P, Yang J, Qi F. Investigation on the pyrolysis and oxidation of toluene over a wide range conditions. Ii. A comprehensive kinetic modeling study. Combust Flame 2015;162(1):22–40. [48] Vasudevan V, Davidson DF, Hanson RK. Shock tube measurements of toluene ignition times and oh concentration time histories. Proc Combust Inst 2005;30(1):1155–63.
543