Autoignitions of Diesel Fuel/Air Mixtures Behind Reflected Shock Waves

O.G. Penyazkov1*, K.L. Sevrouk1, V. Tangirala2, N. Joshi2

2

Heat and Mass Transfer Institute, Minsk, Belarus General Electric Global Research Center, Niskayuna, NY, USA

Abstract The ignition times and auto-ignition modes of Diesel fuel/Air mixtures behind reflected shock waves were measured at pressures 4.7 10.4 atm, temperatures 1065 - 1838 K, and stoichiometries = 0.5 - 2. It was shown that for studied range of post-shock conditions the reaction rate of Diesel fuel oxidation exhibit a nonlinear Arrhenius dependence with global activation energies ranged from 32.6 kcal/mole at high temperatures (> 1200 K) to 20.4 kcal/mole at low temperatures (1200 K <). For stoichiometries = 0.5 1; 0.5 - 2, overall empirical approximations for ignition delays have been derived from the experimental data. Critical post-shock temperatures required for strong and transient auto-ignitions of Diesel fuel were established from measurements. Introduction Most aviation fuels are mixtures of a large number of hydrocarbons and the chemical kinetic processes occurring in combustion of such fuels at high temperatures have not been sufficiently validated. There is scarcity of data on auto-ignition of heavy fuels at high temperatures and pressure. Only limited data have been reported for ignition of Diesel fuels in jet-stirred reactors [1] and in shock tubes [2]. Thus, there is an obvious lack of experimental data for high-temperature combustion and auto-ignition of Diesel fuel and its surrogates in air. The objectives of this study are as follows: to investigate systematically combustion and ignition properties of Diesel No.2 at 4.7 10.4 atm, temperatures 1065 - 1838 K, and stoichiometries = 0.5 2; to compare auto-ignition of premixed Diesel fuel/Air mixtures with a Jet-A/air mixture at equivalent post-shock conditions; to obtain reference data on ignition of Diesel fuel/Air mixtures at high temperatures itself. Experimental setup Stainless steel heated shock tube of 76 mm in diameter was used in the experiments. The wall thickness was 9 mm. The length of the tube was 6.0 m. The tube channel was calibrated with accuracy of 0.005 mm. Four independent current circuits provided an independent heating of the four shock tube sections to ensure the uniform temperature distributions along the tube length with accuracy 50 C. For shock tube tests presented in this report the initial temperature of the tube was kept between 100 - 1100 C. The commercial No. 2 Diesel fuel was used for these studies. The density of used Diesel sample was (0.82370.01) g/ml at 26 0 C. The runs were performed in stoichiometric = 1, lean = 0.5, and rich = 2 Diesel fuel /Air mixtures within the range of post-shock pressures of 4.7 10.4 atm and temperatures of 1065 1838 K (Table 1). Figure 1 presents the drawing of the test section for 1* Corresponding author: penyaz@dnp.itmo.by Proceedings of the European Combustion Meeting 2009

studies in Diesel fuel/Air mixtures. Pressure variations at different cross-sections of the tube were measured by three high-frequency PCB pressure sensors Model 113A24 with a rise time less than 1.5 s and with a 1.5mm spatial resolution. The end wall PCB pressure transducer measured the reflection time and pressure history behind reflected shock waves in the vicinity of the reflecting surface. For controlling ignition and reaction times the additional ion current sensor was mounted into reflecting surface of the shock tube. To detect arrival times of reaction front two additional ion current sensors were installed along the shock tube channel at different distances from the reflecting wall. This setup ensured obtaining information on propagations of shock and reaction fronts over the length of 340 mm upstream the end wall. Pressure and ion current signals were recorded and processed by an automatic 10-bit data acquisition system connected with a central computer. To provide spectroscopic observations two quartz rods of 8 mm in diameter has been mounted into the plane end wall. The first rod ensured emission observations from the gas column of 5 mm in diameter along the centerline of the tube, the second from the similar gas column along the tube wall in the boundary layer. To fix the instant at which the luminosity commences, both gas volumes ( 5 mm) has been imaged onto a photomultiplier detectors installed in a focus of the collecting quartz lens f= 400 mm. The 1mm apertures mounted in front of the photomultipliers ensured the angle selection of transmitted radiation and passed only light beams propagating along the shock tube axis. The beam-splitters divided output spectrum into three optical paths to provide the simultaneous coaxial emission observations at several spectral bands. The luminosities of OH radicals (transitions A2 X2) at wavelength of 308.9 nm and CH radicals (transitions 2 2) at wavelength of 431.5 nm were implemented to measure auto-ignition of the mixture along the centerline of the shock tube. The luminosity of C2 radicals (transitions 2 2) at wavelength of 516.5 nm

Table 1. The experimental conditions for No. 2 Diesel fuel /Air mixtures Diesel - Air Mixture 1 Mixture 2 Mixture 3 Equivalence ratio, 1 0.5 2.0 Post-shock Pressure [atm] 4.68 10.4 5.6 9.8 5.7 10.0 Post-shock Temperature [K] 1078 -1665 1117 1903 916 1838 Post-shock Density [kg/m3] 1.49 2.71 1.3 2.47 1.4 3.17

Figure 1. Schematic of the test section for auto-ignition studies in Diesel fuel/Air mixtures in a 76-mm shock tube: 1 high-frequency PCB pressure transducers; 2 ion current sensors; 3 thermocouple; 4 reflecting wall with inserted quartz rods; 5 lens f = 40 cm; 6 beam splitter; 7 aperture diaphragms; 8 doubled monochromatic filters; 9- photomultipliers.

were detected to identify the auto-ignition in the boundary layer. Ignition times were controlled also by pressure and ion current measurements at the reflecting wall. The ignition or induction time of the mixture was defined as the time difference between shock arrival at the end wall and the onset of emission within measuring gas columns (Fig.2). The applied optical setup was sensitive to the onset of auto-ignition at selected gas volumes and generated induction times of studied mixtures from the beginning of normal reflection of the incident shock wave. To obtain a correct temperature dependence of ignition time on activation energy of the mixture and fuel/oxygen concentrations all comparative shock-tube series were performed at approximately constant post-shock density. It means that fuel, oxygen, and nitrogen concentrations were kept nearly constant within a studied temperature range behind reflected shock waves. Absolute velocities of reflected shock wave (RSW) in a frame of reference attached to gas flow moving behind incident shock wave and pressures at different locations identified the auto-ignition modes of the mixture (strong, transient and weak) [3-6]. The absolute RSW velocity in the end part of the tube was defined as V = V5 + u, where V5 is RSW velocity 1* Corresponding author: penyaz@dnp.itmo.by Proceedings of the European Combustion Meeting 2009

calculated by processing shock-arrival times at pressure sensors along the tube in laboratory frame of reference, u is flow velocity behind incident shock wave. For stoichiometric Diesel fuel / Air mixture, Figure 3 illustrates the dependencies of reflected shock-wave

Figure 2. Ignition time definition criteria and their positions along the history of reflecting wall pressure and gas emission in stoichiometric Diesel fuel / Air mixture: A OH and CH emissions along the centerline of the shock tuber. Stoichiometry = 1. Post shock temperature is T = 1188 K.

Velocity on post-shock temperatures at different locations along the tube. On the basis of pressure and emission observations the inflection point of velocity curve at low temperatures for distances of 140 mm was used for determining positions of the strong explosion limit. For stoichiometric mixture, this critical temperature was 1251 K. It should be mentioned that explosion behavior of Diesel fuel was observed also at temperatures higher than 1140 K due to the mechanism of deflagration to detonation transition in a shock compressed gas volume. As is seen in the Figure 3 this transient ignition mode was realized within the temperature range of 1140 1251 K.

Figure 3. Velocities of reflected shock wave at 140 and 340 mm from reflecting wall vs. post-shock temperature in stoichiometric Diesel fuel / Air mixture. Positions of the strong and transient ignition limits are indicated on the Graphs. The main measurement uncertainties were associated with several factors. The first one is a 0.5% uncertainty in incident shock-wave velocity measurements, which produces 0.75%, 1.5%, 0.7% experimental errors in temperature (T), pressure (P), and density () of the mixture behind reflected shock waves. The second one is an uncertainty connected with the definition of specific heat and enthalpy of the Diesel fuel. Usually, this can result in a 1-1.5 % error in determination of post-shock parameters of the gas. The last one is an uncertainty connected with the definition of ignition-delay time. Usually, this can result in a 2-15 % error in ignition-delay time for studied range of parameters Results For post-shock density of 1.94 0.29 kg/m3 in stoichiometric Diesel fuel/Air mixtures, the temperature dependence of induction times is plotted in the Figure 4. The experiments were performed within the temperature range of 1078 1655 K. Experimental points for No.2 Diesel fuel correspond to ignition times measured by using co-axial emission observations along the centerline of the shock tube at = 308.9 nm (OH, transitions A2 X2) and at = 431.5 nm (CH, 1* Corresponding author: penyaz@dnp.itmo.by Proceedings of the European Combustion Meeting 2009

transitions 2 2), along the tube wall in boundary layer at = 516.5 nm (C2, transitions 2 2 ), pressure and ion current measurements at the reflecting wall. It is apparent that the data obtained by different methods and at different initial temperatures correlate well in a studied range of post-shock temperatures. The induction times of aviation kerosene Jet-A [6] obtained at nearly the same post-shock conditions are drawn on the same graph. For stoichiometric mixture, No. 2 Diesel fuel demonstrates 2.5 2.6 times longer induction periods in comparison with Jet-A. Figure 4 shows that at high temperatures ( > 1210 K ) No. 2 Diesel exhibit almost the same activation energy 16449 K (32.6 0.2 kcal/mole) as aviation kerosene Jet-A. This value is less than activation energies for n-heptane (40.2 kcal/mole) and JP-10 (44.6 kcal/mole) reported in [7, 8]. Although, generally, the ignition delay times of Diesel fuel follow the Arrhenius law, the significant decreasing of activation energy up to 10292 K (20.4 0.2 kcal/mole) has been observed in our experiments at low temperatures T < 1210 K. For stoichiometric No. 2 Diesel fuel /Air mixture, the comparison of current results with existing literature data of Spadaccini & TeVelde [1] measured in continuous flow reactor at inlet air temperatures of 650 1000 K, pressures 10 30 atm, and stoichiometries 0.3 1 exhibit significant deviations with our observations. At the same time, these experiments [1] show a noticeable decreasing of mean activation energy of Diesel fuel at high temperatures 16437 K (32.6 kcal/mole), which is very close to our results for Jet-A and No. 2 Diesel. In accordance with chosen criterion for strong ignition limits experiments demonstrate that the critical post-shock temperature required for strong initiations was equal to T = 1251 K. The transient ignition mode were detected within the temperature range of T = 1140 - 1251 K and M = 2.6 - 2.8, respectively. The measurements of the steady-state CJ detonation velocities in stoichiometric Diesel fuel /Air mixtures give VCJ 1580 15 m/s. The appropriate value in stoichiometric Jet-A mixture is equal to VCJ 1670 m/s.

Figure 4. Mean activation energies for stoichiometric Diesel fuel/ and Jet-A/Air ( = 1.0) mixtures at equivalent post-shock conditions.

For the post-shock density of 1.97 0.28 kg/m3 in lean Diesel fuel/Air ( = 0.5) mixtures, the temperature dependence of induction times is plotted in the Figures 5. Experiments were performed within the temperature range of 1117 1903 K and pressures 5.6 9.8 atm. For lean Jet-A/ Air mixture ( = 0.5), induction times at similar post-shock conditions are drawn on the same graph. Likewise in the stoichiometric mixtures, No. 2 Diesel fuel demonstrates 2.3 2.9 times longer induction periods and slightly lower activation energy 15483 K (30.7 0.2 kcal/mole) in comparison with Jet-A. For lean mixture, the temperature dependence of ignition delays follows the Arrhenius law within the studied temperature range of 1117 1903 K. Within the scatter of the experimental data both stoichiometries = 0.5 and = 1.0 exhibit the same activation energy. The lean mixture demonstrates approximately 1.5 times longer induction periods. The similar trend has been observed for Jet-A fuel in our previous studies [6]. Experiments demonstrate that the critical post-shock temperature required for strong initiations is equal to T = 1294 K. The transient ignition modes were detected within the temperature range of T = 1170 1294, respectively. For lean mixtures, measurements of the steady-state CJ detonation velocity give VCJ 1450 20 m/s. The appropriate value for lean Jet-A mixture is 1480 m/s.

Figure 6. Ignition delay time vs. reciprocal temperature for rich Diesel fuel/ and Jet-A/Air mixtures ( = 2.0) at equivalent post-shock conditions. Arrhenius law within the studied temperature range of 916 1838 K. In comparison with Jet-A Diesel fuel demonstrates 3 - 6.4 times longer induction periods at equivalent post-shock conditions. For rich and stoichiometric Diesel fuel /Air mixtures, Figure 7 shows the comparison of induction times. As is seen form the graphs, linear approximations of the experimental data for = 2.0 and = 1.0 exhibit substantially different activation energies. For temperatures higher than 1200 K, the rich mixture demonstrates longer ignition times. At low temperatures < 1200 K, induction times are approximately equal in both cases within the scatter of the experimental data. Simultaneously, at low temperatures < 1200 K the stoichiometric and rich Diesel fuel/Air blends exhibit very close activation energies equal to 10292 K (20.4 0.2 kcal/mole) and 12330 K (24.45 0.2 kcal/mole) (Fig. 7), respectively. For = 2.0, in contrast to Diesel fuel the Jet-A demonstrates absolutely different behavior [6]. Auto-ignitions of rich Jet-A/Air mixture results in shorter induction times within the temperature range of 1000 1520 K with the same activation energy as for = 0.5 and = 1.0.

Figure 5. Ignition delay time vs. reciprocal temperature for lean Diesel fuel/ and Jet-A/Air mixtures ( = 0.5) at equivalent post-shock conditions. For the post-shock density of 2.2 0.4 kg/m3 of rich Diesel fuel/Air ( = 2.0) mixture, the temperature dependence of induction times is plotted in the Figure 6. Experiments were performed within the temperature range of 916 1838 K, and pressures 5.7 10 atm. Induction times for rich Jet-A/ Air mixture ( = 2.0) are drawn on the same graph. In contrast to lean and stoichiometric Diesel fuel blends, the rich Diesel mixture demonstrates much longer ignition times and significantly lower activation energy 12330 K (24.45 0.2 kcal/mole) in comparison with Jet-A. Within the scatter of experimental data the temperature dependence of induction period for rich mixture follow the 1* Corresponding author: penyaz@dnp.itmo.by Proceedings of the European Combustion Meeting 2009

Figure 7. Mean activation energies for stoichiometric and rich Diesel fuel/ Air mixtures at equivalent post-shock conditions.

For rich Diesel fuel /Air mixtures, the critical postshock temperature required for strong auto-ignitions is equal to T = 1260 K. The transient ignition modes were detected within the temperature range of T = 1145 1260 and M = 2.83 - 3.05, respectively. Measurements give the steady-state CJ detonation velocity VCJ 1735 20 m/s. For rich Jet-A mixture, the corresponding value is 1780 m/s. For stoichiometries = 0.5 - 1, the overall empirical approximation for ignition delays have been derived from the experimental data (1) (Fig.8):
( s ) = 8.0663 10 6 exp
15473 0.28653 0.54218 [Diesel ] [O2 ] T

coincidence with experimental observations for = 2. This correlation results in 27 % standard deviation from the fitted induction times.

where, is the ignition time in (sec) , T is the postshock temperature in (K), [Diesel] is the Diesel fuel concentration in (mole/cm3), and [O2] is the Oxygen concentration in (mole/cm3). The global activation energy of Diesel fuel obtained from regression analysis is 30.7 0.26 kcal/mole. Equation (1) gives the excellent agreement with experimental data points for stoichiometries = 0.5 1. This correlation results in 13 % standard deviation from the fitted induction times.

Figure 9. Ignition delays of Diesel fue/Air mixtures correlated using equations (2) vs. reciprocal post-shock temperature. Units: ( s ); [Diesel], [O2] ( mole/cm3 ); T (K). Conclusions The ignition delay times and auto-ignition modes of Diesel fuel/Air mixtures behind reflected shock waves were measured at pressures 4.7 10.4 atm, temperatures 1065 - 1838 K, and stoichiometries = 0.5 - 2. For stoichiometries = 0.5 1 0.5 -2 the overall empirical approximations for ignition delays have been derived from the experimental data. It was shown that for studied range of post-shock conditions the reaction rate of Diesel fuel oxidation exhibit a nonlinear Arrhenius dependence with global activation energies ranged from 32.6 kcal/mole at high temperatures (> 1200 K) to 20.4 kcal/mole at low temperatures (1200 K <). Critical post-shock temperatures T = 1251 K ( = 1), T = 1294 K ( = 0.5), and T = 1260 K ( = 2.0) required for strong auto-ignitions of Diesel fuel/Air mixtures were established from experimental measurements. Within the temperature ranges of T = 1140 - 1251 K ( = 1), T = 1170 - 1294 K ( = 0.5), and T = 1145 1260 K ( = 2.0) transient auto-ignitions of Diesel fuel/Air mixtures were observed in experiments. CJ detonation velocities in Diesel fuel / Air mixtures are VCJ = 1735 20 m/s (for = 2); 1580 15 m/s (for = 1); and 1450 20 m/s (for = 0.5). Acknowledgments This work was sponsored by General Electric Global Research. References

Figure 8. Ignition delays of Diesel fuel/Air mixtures correlated using equations (1) vs. reciprocal post-shock temperature. Units: ( s ); [Diesel], [O2] ( mole/cm3 ); T (K). For a wider stoichiometry range of = 0.5 - 2, overall empirical approximation for ignition delays is (2) (Fig.9):
( s ) = 1.5563 10 5 exp
13789 0.43055 Diesel [O2 ] T

0.0404

with the global activation energy of Diesel fuel obtained from regression analysis is 27.3 0.42 kcal/mole. Approximation (2) results in the satisfactory 1* Corresponding author: penyaz@dnp.itmo.by Proceedings of the European Combustion Meeting 2009

1. Spadaccini LJ, TeVelde JA (1982) Autoignition Characteristics of Aircraft-Type Fuels. Combust. Flame 46: 283. 2. Haylett DR, Lappas PP, Davidson DF, Hanson RK (2009) Application of an aerosol shock tube to the measurement of diesel ignition delay times. Proc. of the Comb. Inst. 32 (1): 477. 3. Voevodsky VV, Soloukhin RI, Proc. Combust. Inst. 10 (1965) 279-283. 4. Meyer JW, Oppenheim AK, Proc. Combust. Inst. 13 (1971) 1153-1164. 5. Penyazkov OG, Ragotner KA, Dean AJ, Varatharajan B, Proc. Combust. Inst. 30 (2005) 1941-1947. 6. Dean AJ, Penyazkov OG, Sevruk KL, Varatharajan B. Proc. Combust. Inst. 31 (2007) 2481-2488. 7. N.B. Colket, L.J. Spadaccini, 14th Int. Symp. on Airbreathing Engines, Florence, Italy, September 510, 1999. 8. Davidson DF, Horning DC, Herbon JT, Hanson RK, Proc. Combust. Inst. 28 (2000) 1687-1692.

1* Corresponding author: penyaz@dnp.itmo.by Proceedings of the European Combustion Meeting 2009