CARACTERIZATION OF THE BIODIESEL COMBUSTION WITH LDV AND EMISSION SPECTROSCOPY IN A COUNTERFLOW BURNER

Transcription

1 CARACTERIZATION OF THE BIODIESEL COMBUSTION WITH LDV AND EMISSION SPECTROSCOPY IN A COUNTERFLOW BURNER Dario ALVISO a,b,*, Diego ZABRODIEC b, Gustavo FERREIRA b, Philippe SCOUFFLAIRE a, Jorge MOLINA b, Nasser DARABIHA a Juan Carlos ROLON b a Laboratoire EM2C, CNRS UPR 285, Ecole Centrale Paris, Châtenay-Malabry, France b Laboratorio de Mecánica y Energía, Facultad de Ingeniería - UNA, Asunción, Paraguay *dario.alviso@ecp.fr Abstract In this work we propose to study the combustion of biodiesel (rapeseed and soybean) using a diagnostic technique based on spontaneous flame emission, resolved spatially and spectrally through interference filters and a OMA spectrometer, which provides information about the spectral intensity of active species. The Laser Doppler Velocimetry was used to measure the axial velocities. An opposed flow two premixed flames configuration was chosen, with a methane flame providing, if needed, the additional required temperature to achieve the ignition of the biodiesel that is injected in the form of droplets. The objective been to gather information that can be useful for modeling this type of fuels. Mots Clés : Biodiesel / Premixed flames / Opposed flow / LDV / Emission Spectroscopy 1. Introduction In recent years, biodiesel has become interesting as an additive to diesel fuel for two main reasons. This renewable alternative fuel can reduce dependence on imported petroleum and can also contribute to environmental preservation by lowering net emissions of greenhouse gases. The use of biodiesel in diesel engines decreases emissions of pollutants such as carbon monoxide, unburned hydrocarbons, and particulate matter, although a slight increase in emissions of nitrogen oxides is observed in some cases. Biodiesel is a multiple-component mixture of monoalkyl esters of long-chain fatty acids derived from vegetable oils and animal fats. Most biodiesel fuels used in the world are made from soy oil and rapeseed oil by transesterification with an alcohol. The soy and rapeseedderived biodiesels are complex mixtures composed mainly of five saturated and unsaturated methyl esters (when methanol is used for the transesterification process): methyl palmitate (C 17 H 34 O 2 ), methyl stearate (C 19 H 36 O 2 ), methyl oleate (C 19 H 34 O 2 ), methyl linoleate (C 19 H 32 O 2 ), and methyl linolenate (C 19 H 30 O 2 ) «Herbinet et al. [2007]». The objective of this study is to characterize the combustion of biodiesel from a fundamental point of view, knowing that the characteristics of the combustion is highly dependent on the composition of the fuel, which in this case depends strongly on the soil and the local climatic conditions. So, it is not correct to extrapolate the knowing characteristics of those fuels from other latitudes, but rather to do a study of the combustion of these fuels to characterize them properly. The main objective of the experiments results presented below is to gather information that can be useful for modeling the combustion of biodiesel, especially in the construction of Copyright 2011 FLUVISU Page 1

2 reduced kinetic models. To this end, we intend to explore two areas, the velocity and the species emission profiles. 2. Experimental Setup The experimental setup is presented in Fig. 1.A. We have chosen to work in a opposed-flow premixed flames configuration, with the hot burnt gases of the methane premixed flame providing, if needed, the additional required temperature to achieve the ignition of the biodiesel droplets Counterflow burner Experimental studies were carried out in a counterflow burner, consisting of two identical burners facing each other, as one can see in Figure 1. Only a brief description is presented here, additional details are available in «Rolon et al. [1991, 1993, 1999]».. The burner consists of two axisymmetric 20 mm diameter injection nozzles. A premixed gaseous flow of methane and air is injected on the lower side of the burner. In the upper side is injected a mixture of air and biodiesel fuel, in the form of droplets of similar size, which is preheated with a heated hose to ensure a partial evaporation of the biodiesel. The nozzle separation was set to 40 mm in our experiments. Each nozzle is surrounded by a coaxial nozzle issuing nitrogen to isolate the reactive stream from outer perturbations. The upper side of the burner is preheated, with a heater resistance. And nitrogen is transported with a heated hose to ensure a uniform temperature throughout the upper burner. The number of biodiesel droplets is regulated through a regulating valve, the closer the valve is, the more droplets we have. A B Figure 1. Experimental Setup for a biodiesel premixed flame in a counterflow burner. An image of two premixed flames is shown in Figure 1.B, in which we observe in the upper burner the injection of a mix of air and biodiesel droplets, visualized by a laser sheet (produced by an Argon laser of 457 to 514 nm and a diverging cylindrical lens), the evaporation zone and the stabilized premixed biodiesel flame. Below, we can see the premixed methane flame, and between the two flames, the hot zone of combustion products. In Table 1 are presented the experimental conditions with which we carry out all experiments described below. As we can see the Methane flames are the same for Flames I to V, in order Copyright 2011 FLUVISU Page 2

3 to isolate the influence of the different biodiesel materials raw as well as the different initial conditions. Biodiesel flame Methane flame Biodiesel Air N 2 Methane Air N 2 Flame I Raw material Rapeseed m (g/h) 8 Q (m 3 n/h) T ( C) Flame II Raw material Rapeseed m (g/h) 4.3 Q (m 3 n/h) T ( C) Flame III Raw material Soybean m (g/h) 6.94 Q (m 3 n/h) T ( C) Flame IV Raw material Soybean m (g/h) 2.86 Q (m 3 n/h) T ( C) Flame V Raw material Rapeseed m (g/h) 7.27 Q (m 3 n/h) T ( C) Table 1 Experimental conditions of two premixed biodiesel and methane flames 3. Results and discussion 3.1 Laser Doppler Anemometry Laser Doppler Anemometry (LDA) is a non-intrusive technique used to measure the velocity of particles suspended in a flow. If these particles are small, in the order of microns, they can be assumed to be good flow tracers following the flow and thus their velocity corresponds to the fluid velocity «Jensen [2004]». The experimental setup illustrated in Figure 2 was chosen to carry on with the LDA measurements. The measuring volume is located in the area where a mixture of air and biodiesel droplets is injected through the upper burner. The objective is to determine the velocity profile where biodiesel droplets (tracers) are present, i.e. from the upper burner outlet to the area immediately prior to the vaporization zone (see Figure 1B). And as we are interested in the axial velocity profile the intersection point of the laser beams was placed coincident with the burner axis. As one can see we have chosen the "Forward Scatter" configuration. A Bragg cell is introduced in the path of the laser beams. Even if we didn t have a directional ambiguity, the Bragg cell allowed us to move to parts of the spectrum that had a lower density of electronic and optical noise, thus improving the Doppler signal quality. Copyright 2011 FLUVISU Page 3

4 A Helium-Neon Laser (Spectra Physics Model 107B, mw, nm) with a Beam Splitter Module was used to produce, thanks to a cylindrical converging lens (300 mm focal length), two crossing coherent laser beams in order to generate a control volume in the area of interest. Figure 2 Laser Doppler Anemometry Set up A Photomultiplier tube (EMI 9871 A) was used for the detection of laser light scattered by the tracers. Two lenses (425 and 80 mm focal length) was used to focus the light scattered into the pinhole (200 μm), while a diaphragm limits the amount of light that reaches the PM tube. A Digital Oscilloscope (Tektronix model TDS2022C) was used to digitilize the signal coming from the PM tube. We have used the Welch method ir order to process the signal and to obtain the Doppler frequency, and hence the flow velocity. The raw signal was captured using a sampling rate of 5 Megasamples/sec with 10 to 20 acquisitions of 2500 elements each, giving an average data series of to elements for each analyzed point. A measuring volume of approximately 200 μm was estimated with the chosen configuration, and the micro-positioning system gave us a translational accuracy of 100 μm, while the area of interest, a mixture of air and biodiesel droplets, measured about 15 mm. As a result, a high spatial resolution was obtained. Moreover, with a sampling frequency of 5 MHz, 10 times greater than the detected frequencies, and setting the displacement of the Bragg cells in 200 KHz, we have a high temporal resolution. With this configuration we can measure velocities in the range between and 9.5 m/s. Figure 3 show the velocity profiles of Flames I, II and IV, respectively, in the axis direction at a constant distance of 2 mm from the upper burner, in order to verify that the biodiesel droplets are good tracers and move at the same speed as the air, in addition we can see that the velocity profile is uniform and that the edge effect is almost negligible. As Flames I and III have the same air flow rate (see Table 1) they have also the same velocity profiles. Finally, Figure 4 show the axial velocity profiles of Flames I, II and IV, respectively, from 2 mm from the upper burner outlet to the area immediately prior to the vaporization zone. Making a quick comparison of the 3 Figures we can say that the position of the biodiesel flame is mostly related to the strain rate. So, Flame II is the one that stabilizes the farther from the upper burner. As we can see the profiles are not perfectly continuous, which may be due Copyright 2011 FLUVISU Page 4

5 to small fluctuations in the biodiesel flames, however the trend is clear enough for the comparisons with the numerical results. Figure 3 Velocity profile of Flames I, II and IV at a constant distance of 2 mm from the upper burner Copyright 2011 FLUVISU Page 5

6 Figure 4 Axial velocity profile of Flames I, II and IV from 2 mm from the upper burner outlet to the area immediately prior to the vaporization zone 3.2 Emission spectroscopy Emission spectroscopy is a 2-D nonintrusive diagnostic technique that offers spatially resolved data for combustion optimization and control. The UV and visible chemiluminescence of the excited radicals CH(A 2 Δ,B 2 Σ - ) and OH(A2 Σ + ) is studied experimentally in opposed-flow premixed flames. Two 500-mm focal length plano convex synthetic fused silica ultraviolet lenses was used to focus the light coming from the flame into the slit of a high resolution grating (1200 gr/mm) spectrometer (Acton Research, Model SpectraPro 2750). A periscope is located between the burner and the first lens because the spectrometer has only one vertical slit. The emission spectrum at each point along the burner axis was recorded by a scientific grade cooled CCD camera (Roper Scientific, Inc.). During the experiments the first lens was positioned at 50 cm apart from the flame axis, taking into account that the periscope is midway, while the second lens is located at 50 cm from the spectrometer slit. The distance between the two lenses was 10 cm. A correction in the detector wavelength response with a Tungstene lamp was done, as well as a substraction of the ambient background. We can take the methane flame as a reference for the comparisons with the simulations, so no need to make a calibration in intensity. The CO 2 substraction was done for the CH* according to «Hardalupas and Orain et al. [2004]». The OH* population was obtained by making an integration between 300 and 320 nm, while the CH* population was acquired between 420 and 440 nm. As we can see in figure 1.B both flames have a slight curvature, which creates an artificial signal that increases the measured concentration and produces the asymmetric profiles. The curvature of the biodiesel flame and related strong asymmetry of CH* profile are clearly visible in Fig. 5. To solve this problem, an Abel inversion of a 2-D image captured with a CCD camera/uv lens assembly was performed. Narrow-band interference filters were interposed along the optical path for capturing the excited CH and OH emission. The filter used for excited CH has 60% transmission and a 10-nm-wide bandpass centered around 430 nm. The second filter is Copyright 2011 FLUVISU Page 6

7 centered at 313 nm, 10-nm bandpass and 65% transmission in the maximum. The Abel inverted data were used for postprocessing the spectrometer data in order to correct the effect of curvature (Fig. 5). The reactants composition and velocities are metered by the mass flow controllers (Bronkhorst), with an accuracy of ±1% and a repeatability of ±0.5% of the full scale. A B Figure 5 (A) Original and Abel inverted image of CH* emission from Flame V (Biodiesel flame), (B) axial CH* emission profiles of the original and Abel inverted data normalized with the peak values. A B Figure 6 (A) Axial CH* emission profile for Flame V obtained with the spectrometer (B) Zoom of the axial CH* emission profile around the biodiesel flame. In Figure 6 we can see the axial CH* emission profiles of Flame V, before the corrections of the curvature. The difference in the maximum CH* emission is because the biodiesel flame is much less intense compared to the methane flame, because the atomizer does not produce enough droplets taking into account the air flow injected. Figure 7 CH* emission profile for Flame V (around the biodiesel flame) of the original and Abel inverted data. Copyright 2011 FLUVISU Page 7

8 In Figure 7 is presented the CH* emission profile around the biodiesel flame that has been postprocessed with the Abel inverted data, however it remains a slight asymmetry in the profile in the upper burner direction. In Figure 8 is presented the axial OH* emission profile of Flame V, where one can note the expected widening of the OH* peak, however because of the difference in the intensity of the two flames, as mentionned before, the OH* peak of the biodiesel flame can not be identified. A B Figure 8 (A) Axial OH* emission profile obtained with the spectrometer (B) Zoom of the axial OH* emission profile around the biodiesel flame. 4. Conclusion The experimental study of biodiesel (rapeseed and soybean) combustion has been performed in a counterflow burner with two premixed flames configuration. The objective was to gather information for validating the kinetic scheme that will be developed later. To this end, the velocity profiles obtained with LDV can be very useful, as it is a non intrusive technique and diferent profiles were studied. In the other hand, more species emission profiles for diferent initial conditions (equivalence ratio and strain rate) need to be obtained in order to evaluate the evolution of these species emission for different biodiesel flames. Réferences Candel, S., Lacas, F., Rolon, J.C. and Darabiha, N. (1999). Group combustion in spray flames. Multiphase Science and Technology, 11: pp Darabiha, N., Lacas, F., Rolon, J.C. and Candel, S. (1993). Laminar counterflow spray diffusion flames : A comparison between experimental results and complex chemistery calculations. Combus. and Flame, 95: pp P. Dagaut, S. Gaı l, M. Sahasrabudhe, Proc. Combust. Inst. 31 (2007) Y. Hardalupas and M. Orain. Combustion and Flame, 139 (2004), pp O. Herbinet, P.M. Marquaire, F. Battin-Leclerc, R. Fournet, J. Anal. Appl. Pyrol. 78 (2006) O. Herbinet, W.J. Pitz, C.K. Westbrook, Combust. Flame 154 (2008) Copyright 2011 FLUVISU Page 8

9 K. D. Jensen, ENCIT th Brazilian Congress of Thermal Sciences and Engineering, 2004, Rio de Janeiro, RJ, Brazil. R.L. McCormick, J.D. Ross, M.S. Graboski, Environ. Sci. Technol. 21 (4) (1997) Rolon, C., Veynante, D., Martin, J.P. and Durst, F. (1991). Counter diffusion flows. Exper. in Fluids, 149: pp K. Seshadri, T. Lu, O. Herbinet, S. Humer, U. Niemann, W.J. Pitz, C.K. Law, Proc. Combust. Inst. 32 (2009) M.E. Tat, P.S. Wang, J.H. Van Gerpen, T.E. Clemente, J. Am. Oil Chem. Soc. 84 (2007) A. Tsolakis, A. Megaritis, M.L. Wyszynski, K. Theinnoi, Energy 32 (2007) C.K.Westbrook,W.J. Pitz, H.J. Curran, J. Phys. Chem. A (110) (2006) C.K. Westbrook, W.J. Pitz, O. Herbinet, H.J. Curran, E.J. Silke, Combust. Flame 156 (1) (2009) Copyright 2011 FLUVISU Page 9