Microscopic Spray Investigation of Karanja Biodiesel and Its Effects on Engine Performance and Emissions

ILASS-Asia 2016, 18 th Annual Conference on Liquid Atomization and Spray Systems - Asia, Chennai, India Microscopic Spray Investigation of Karanja Biodiesel and Its Effects on Engine Performance and Emissions Chetankumar Patel, Nikhil Sharma, Nachiketa Tiwari and Avinash Kumar Agarwal * Engine Research Laboratory, Department of Mechanical Engineering Indian Institute of Technology Kanpur, Kanpur-208016, India Abstract Diesel engines are widely popular for small to large size vehicles and agricultural/ construction/ power generation machines. Biodiesel from non-edible oils such as Karanja can be used as an alternative fuel without modifications in the engine hardware. Fuel atomization is largely affected by higher viscosity of biodiesel. Microscopic spray investigations are conducted to understand the spray behavior of the biodiesel. This investigation carried out to measure spray droplet s sauter mean diameter (SMD), droplet velocity and its variations with respect to time at lower injection pressure of 200 bar (Typical fuel injection pressure in mechanical fuel injection equipment equipped engines) at atmospheric condition for Karanja biodiesel (KB100) and mineral diesel. Higher SMD and slightly lower velocity were observed for KB100 vis-à-vis baseline mineral diesel. Engine performance and emissions were measured for these test fuels at constant engine speed of 1500 rpm. Higher CO emission and lower HC and NO X emissions were observed from KB100, primarily due to differences in spray characteristics and fuel properties. Keywords: Microscopic Spray Characteristics; SMD; Droplet Velocity; Biodiesel; Emissions. * Corresponding author: akag@iitk.ac.in

ILASS-Asia 2016, 18 th Annual Conference on Liquid Atomization and Spray Systems - Asia, Chennai, India Introduction Biodiesel from non-edible vegetable oils has emerged as a strong alternative to replace diesel as an alternate fuel in near future. However biodiesel has higher viscosity, and density, which largely affects fuel injection equipment and ultimately affects fuel spray atomization. Engine emissions largely depend on the fuel atomization. Microscopic and Macroscopic spray investigations have drawn large attention of researchers in last few decades, which have helped understand the behavior of alternative fuels in engine combustion chamber. Agarwal et al. [1] reported that cavitation, turbulence and velocity of the fuel at the nozzle exit are critical parameters for spray atomization. These parameters influence fuel atomization characteristics. Som et al. [2] investigate biodiesel and diesel flow in the injector and observed that biodiesel cavitate less compared to mineral diesel because of its lower vapour pressure. They also found that higher viscosity of biodiesel results in reduction in injection velocity. Faria et al. [3] observed that increasing biodiesel content in biodiesel/ diesel blend resulted in reduction in spray atomization quality. They also observed that higher average sauter mean diameter (SMD) resulted in reduction in mass transfer during combustion process. Chong and Hochgreb [4] reported 23% smaller spray droplet sizes for Jet A-1 fuel compared to Rapeseed Methyl Ester (RME). This was due to lower viscosity and higher volatility of JET A-1 fuel, which promoted spray vaporization. Gao et al. [5] reported higher SMD of non-edible oil based biodiesel blends due to higher viscosity and surface tension compared to conventional diesel. Kim et al. [6] investigated the effects of DME blending with biodiesel on the SMD. They reported that droplet size of DME blended biodiesel was similar to that of diesel, regardless of fuel injection pressure. Monirul et al. [7] conducted investigations in a single cylinder diesel engine using 10% and 20% palm, Jatropha and Calophyllum inophyllum biodiesel blends and reported higher BSFC and lower BTE in case of biodiesel compared to baseline diesel. Usta [8] conducted investigations on a turbo-charged in-direct injection (IDI) diesel engine by using tobacco seed methyl ester. He reported reduction in CO and slightly increase in NO X emissions. Lesnik and Bilus [9] investigated the effect of rapeseed biodiesel on engine performance and emission characteristics and reported higher BSFC and lower NO x and CO emissions, when higher biodiesel blends were used. Few other researchers conducted investigations using varying fuel injection quantity and low injection pressure. The investigations were carried out to measure spray droplet size and velocity distributions for mineral diesel and Karanja biodiesel under atmospheric conditions 40 mm downstream of the fuel injector. Experimental Set up Droplet size and velocity distributions were measured by Phase Doppler Interferometry (PDI) instrument (Figure 1a). Spray measurement directions are shown in Figure 1b. Spray droplet velocity components were measured in X, Y and Z directions. Table 1 shows the specifications of the PDI instrument. PDI experimental set up consists of two transmitters (1 and 2), a receiver, three advanced signal analyzers and the processing software. Three diode pump solid state (DPSS) lasers are used in the PDI experiment. Transmitter 1 emits two lasers, green (532 nm) and blue (491 nm); while transmitter 2 emits yellow laser (561 nm). A beam splitter splits each of the laser beams into two beams of nearly equal intensity and a grating is used for creating a slight phase shift in these two beams of the same wavelength. These six laser beams were then carefully aligned to meet at a single point, which is called Probe volume. Alternate dark and bright fringes develop in the probe volume, where two laser beams of same wavelength with same phase shift intersect. When a spray droplet passes through the probe volume, changes in phase difference and frequency variations are observed. The changes in frequency are directly proportional to droplet velocity components, while the phase difference is directly proportional to the droplet diameter. A solenoid injector injected the fuel spray in a constant volume spray chamber (CVSC) and the injector was controlled by an injector driver module (NI; 9411). Fuel was pressurized by a pneumatic high pressure fuel pump (Maximator) and then supplied to the injector via high pressure pipe lines. Investigations were carried out 40 mm downstream of the injector nozzle under atmospheric conditions. This investigation was carried out at lower fuel injection pressure of 200 bar (typical of mechanical fuel injection systems) with varying fuel injection quantities (12 to 32 mg/ injection). Only one of the holes of the injector was used for generating spray plume. Other nozzle holes were covered by a customized cap. Table 1: Technical Specifications of PDI Droplet size range 0.5 to 2000 μm Estimated accuracy ± 0.5 μm Estimated resolution ± 0.5 μm Velocity measurement range -100 to 300 m/s Velocity accuracy ± 1% Volume flux accuracy ± 15% Focal length of receiver lens 350, 500, 750, 1000 mm Focal length of transmitter 350, 500, 750, 1000 mm lens Laser type Diode pumped solid state (DPSS) Laser wavelength Blue: 492 nm; Green: 532 nm; Yellow: 660 nm

Test Fuel Table 2: Test fuel properties Calorific Value (MJ/kg) Density (g/cm 3 ) Mineral 43.22 0.820 2.67 Diesel KB100 40.34 0.886 5.74 Kinematic Viscosity @ 40 C (cst) Figure 2 shows the schematic of the engine experimental setup. A single cylinder constant speed compression ignition engine (Kirloskar; DM-10) was used for determining the performance and emission characteristics of test fuels. The tests were conducted at constant engine speed of 1500 rpm at varying engine loads (0% to 100% load). Engine emissions were measured by a five gas emission analyzer (AVL, 4000) and smoke opacity was measured by smoke opacimeter (AVL, 437). The engine was coupled to an AC alternator and the power generated by the alternator was dissipated in resistive load, thus loading the engine. The engine was equipped with fuel flow rate measuring system, air flow rate measuring system, smoke opacimeter and raw exhaust gas emission analyzer, as shown in the schematic (figure 2).Table 2 shows the properties of the test fuels. Results and Discussion Microscopic spray investigations were carried out for diesel and Karanja biodiesel by varying fuel injection quantities (12, 16, 20, 24, 28, 32 mg). The results of SMD, Droplet size distribution, velocity components distributions in different directions were experimentally determined and correlated with the engine performance and emission characteristics. Sauter Mean Diameter (SMD) Droplet size and droplet velocity distributions play important role in fuel spray atomization, mixing of fueland air and resulting combustion and emission characteristics of an internal combustion engine. Figure 3 shows the SMD of Karanja biodiesel and baseline mineral diesel and Figure 4 shows the SMD variations with time after the start of solenoid energizing. SMD of biodiesel was relatively higher than baseline mineral diesel for all fuel injection quantities, which are representative of different engine loads i.e. higher fuel quantity is injection in an engine cycle in the combustion chamber at higher engine load. Karanja Biodiesel also showed higher SMD after energizing of solenoid injector w.r.t. time. Higher SMD of the biodiesel was observed primarily due to relatively higher fuel viscosity and surface tension in case of biodiesel, which affected spray breakup and spray evolution after the fuel injection. Higher viscosity also resulted in higher van-der-wall forces among the biodiesel fuel molecules, which led to larger droplet size distributions. Droplet Velocity Distribution Velocity variation of spray droplet was divided into injection duration, time taken for detection, head (constant velocity region) and tail (decreasing velocity region) sections [10]. Figure 5a shows the droplet velocity variations in X direction for diesel and Karanja biodiesel for 12 and 32 mg fuel injection quantities. Head section duration increased with increasing fuel injection quantity. It was primarily due to the higher fuel injection duration for injection of higher fuel quantity. Velocity in X- direction ranges from 0 to 80 m/s for all test fuels for 12 and 32 mg fuel quantity. Figure 5b shows the average droplet velocity variations in X-direction for 12 to 32 mg fuel quantity per injection. Maximum average velocity observed was ~60 m/s for both test fuels. There were no significant differences observed in average velocity of the Karanja biodiesel and diesel droplets at lower fuel injection quantities (12 and 16 mg per injection). Slightly lower average droplet velocity was observed for Karanja biodiesel for higher fuel quantities of 20, 24, 28 and 32 mg per injection. This may be due to the reduction of injection velocity at the exit of the nozzle for biodiesel fuel due to higher aerodynamic drag, which is observed due to larger droplet sizes [2]. Figure 6a shows the droplet velocity variations in Y-direction for 12 mg and 32 mg fuel per injection. Head section duration was 1 ms for 12 mg fuel per injection in case of diesel and KB100, both. Droplet velocity range was from 0 to 50 m/s for 12 mg fuel per injection for diesel as well as KB100. Tail section starts 2 ms after energizing the injector. The tail section than becomes constant beyond 5 ms. Head section duration for mineral diesel was 2.3 ms with droplet velocity ranging from 0 to 50 m/s, while KB100 showed similar head section duration with droplet velocity ranging from 0 to 40 m/s, for 32 mg fuel injection quantity. Figure 6b shows the average droplet velocity variations in Y-direction. Average velocity in Y-direction showed trends similar to X-direction but with reduced maximum velocity. Diesel droplets exhibited maximum velocity in the range of ~30-40 m/s while it was ~25-30 m/s in case of KB100.

Figure 7a shows droplet velocity distribution in Z- direction for 12 and 32 mg fuel per injection. Negative droplet velocity reflects that velocity component was detected in opposite direction. Detection time was larger for higher fuel injection quantity. Numbers of droplets passing in this direction were also lower compared to the other two directions. This was due to injector position of the experiment. Figure 7b showed that maximum average droplet velocity was ranging from ~20 to 30 m/s for all test fuels and all fuel quantities injected. These spray investigations were followed by performance and emission characteristics of biodiesel w.r.t. baseline mineral diesel. Engine Experiments Engine experiments were carried out for mineral diesel and KB100. Experimental investigations included performance and emissions characterization of the engine at 0, 20, 40, 60, 80 and 100% loads at 1500 rpm engine speed. Figure 8 shows the effect of Biodiesel on engine performance parameters such as brake thermal efficiency (BTE), brake specific fuel consumption (BSFC), brake specific energy consumption (BSEC) and exhaust gas temperature (EGT). It was observed that BTE of the engine increased with increasing engine load. BTE was relatively lower for KB100 compared to baseline mineral diesel. It was due relatively higher fuel viscosity of Karanja biodiesel, which led to higher droplet size distribution in fuel spray (Figure 3 and 4), finally leading to relatively more inefficient combustion. BSFC and BSEC were relatively higher at lower engine loads and decreased with increasing engine load. BSFC was relatively higher for biodiesel due to its relatively lower calorific value. KB100 showed higher BSFC and BSEC compared to baseline mineral diesel. EGT didn t show any significant difference among the two test fuels and showed almost similar trend at all loads. Figure 9 shows CO, CO 2, HC, NO X emissions and smoke opacity. Higher viscosity of biodiesel resulted in higher SMD and lower droplet velocity distribution compared to baseline mineral diesel. Biodiesel showed relatively inferior atomization due to these two important factors (Higher SMD, and lower droplet velocity). In-cylinder temperature of biodiesel during combustion remains lower due to lower heat release rate. Brake specific CO emission, although very low, were observed to be relatively higher at lower engine loads for all test fuels. KB100 showed marginally higher CO emission at all loads due to higher fuel viscosity, which resulted in higher SMD and lower droplet velocity distribution, which was responsible for relatively inferior mixing of the test fuels with ambient air. This resulted in marginally higher CO formation in case of biodiesel Brake specific CO 2 emissions were relatively higher at lower engine loads and decreased with increasing engine loads for both test fuels. CO 2 emissions were almost similar for baseline mineral diesel and KB100. Brake specific HC emissions were observed to be relatively higher at lower engine loads but decreased with increasing engine load for all test fuels. HC emission levels were insignificant for any meaningful comparison though. NO x emissions were also relatively higher at lower engine loads and they decreased with increasing engine load. Biodiesel showed lower NO x emission compared to baseline mineral diesel. This was mainly due to lower in-cylinder temperatures, caused by lower HRR max which resulted in lower brake specific NO X emissions in case of biodiesel. Smoke opacity of the exhaust was relatively lower at lower engine loads and increased with increasing engine load. Fuel quantity injected increased with increasing engine load, which resulted in higher particulate formation therefore higher smoke opacity. Smoke opacity is an indication of presence of larger particles in the exhaust. Mineral diesel showed relatively higher smoke opacity at all loads compared to KB100, except for lower loads. This was primarily due to oxygen present in biodiesel fuel molecules, which reduced particulate formation in the combustion chamber. Conclusions Microscopic spray investigations were carried out for measurement of droplet size distribution and droplet velocity distribution in different directions for varying fuel injection quantities of Karanja biodiesel and compared with baseline mineral diesel. Engine performance and emissions investigations were carried out for the same test fuels at similar injection pressures. Conclusions from these experiments are as follows: 1. SMD and SMD distributions were relatively higher for biodiesel compared to baseline mineral diesel, due to higher viscosity and surface tension of KB100. 2. Average spray droplet velocities were similar for both test fuels at lower fuel injection quantities of 12 and 16 mg in X and Y-directions. It slightly decreased for KB100 for higher fuel injection quantities of 20, 24, 28 and 32 mg. 3. Head duration (constant velocity region) increased with increasing fuel injection quantity in X and Y direction. 4. Droplet velocity variations in Z direction were lower because majority of the spray zone was in X and Y direction. 5. Higher CO and lower HC and NOX emissions were observed for KB100 compared to baseline diesel. Higher BSFC and lower BTE were observed for KB100. These were primarily due to higher viscosity of

biodiesel, which led to larger droplet size distribution, lower droplet velocities and relatively inefficient mixing of oxygenated fuel and air. References 1. Agarwal, A.K., Som, S., Shukla, P.C., Goyal, H. and Longman, D., Applied Energy, 156:138-148(2015). 2. Som, S., Longman, D.E., Ramírez, A.I. and Aggarwal, S.K., Fuel, 89(12):4014-4024(2010). 3. Faria, M.D.C., Valle, M.L.M. and Pinto, R.R.D.C., SAE technical paper 2005-01-4154. 4. Chong, C.T. and Hochgreb, S., Fuel, 115:551-558(2014). 5. Gao, Y., Deng, J., Li, C., Dang, F., Liao, Z., Wu, Z. and Li, L., 27(5):616-624(2009). 6. Kim, I., Hyun, G., Goto, S. and Ehara, R., SAE Technical Paper 2001-01-3636. 7. Monirul, I.M., Masjuki, H.H., Kalam, M.A., Mosarof, M.H., Zulkifli, N.W.M., Teoh, Y.H. and How, H.G., Fuel, 181:985-995(2016). 8. Usta, N., Energy Conversion and Management, 46(15):2373-2386(2005). 9. Lešnik, L. and Biluš, I., Energy Conversion and Management, 109:140-152(2016). 10. Patel, C., Sharma, N., Tiwari, N. and Agarwal, A.K., SAE Technical Paper 2016-01-0994. Nomenclature PDI Phase Doppler Interferometry SMD Sauter Mean Diameter HC Hydrocarbon CO Carbon Monoxide NO X Oxides of Nitrogen CO 2 Carbon Dioxide KB100 Karanja Biodiesel ms Milli-second µm Micro-meter DME Dimethyl Ether BSFC Brake Specific Fuel Consumption BSEC Brake Specific Energy Consumption EGT Exhaust Gas Temperature

ILASS-Asia 2016, 18 th Annual Conference on Liquid Atomization and Spray Systems - Asia, Chennai, India Figure 1a: Schematic of the experimental setup for microscopic spray characterization using PDI Figure 1b: Spray measurement directions for test fuel Figure 2: Schematic of the engine experimental setup

Figure 3: Sauter mean diameter of spray droplets of different test fuels for varying fuel injection quantities Figure 4: Average SMD variations with time after solenoid energizing for varying fuel quantities (12 to 32 mg per injection)

Figure 5a: Spray droplet velocity distribution in X-direction for 12 and 32 mg fuel per injection Figure 5b: Average spray droplet velocity variations in X-direction for 12 to 32 mg fuel per injection

Figure 6a: Spray droplet velocity distribution in Y-direction for 12 and 32 mg fuel per injection Figure 6b: Average spray droplet velocity variations in Y-direction for 12 to 32 mg fuel per injection

Figure 7a: Spray droplet velocity distribution in Z-direction for 12 and 32 mg fuel per injection Figure 7b: Average spray droplet velocity variations in Z-direction for 12 to 32 mg fuel per injection

Figure 8: Biodiesel fuelled engine s performance characteristics Figure 9: Biodiesel engine s exhaust emissions characteristics