1
Matt Basinger 1 and Vijay Modi 2 1 Eng.Sc.D. Candidate in Earth and Environmental Engineering at Columbia University 2 Professor of Mechanical Engineering at Columbia University Though many plant oils have a similar energy density to fossil diesel fuel, several properties of plant oils are considerably different from those of diesel. Engine modifications can overcome some of these differences. An engine modification kit has been designed and tested for a slow speed, stationary, indirect-injection diesel engine - the Lister-type CS 6/1, common throughout the developing world. The kit allows vegetable oil fueling with similar performance to that of diesel fueling. The kit's simple yet robust design is targeted for use as a development mechanism, allowing remote farmers to use locally grown plant oils as a diesel substitute, as shown in Figure 1. The viscosity of plant oils is much higher then diesel at room temperature, but as temperature increases viscosity exponentially decreases, as shown in Figure 2. For this reason a preheater was designed to utilize the unique COV design of the Listeroid engine, as shown in Figure 3. The preheater is passive, utilizing waste heat from the engine, making it ideal for a remote, rural setting. In addition to the preheater, the modification kit also includes the tuning of the injector pressure and timing to better atomize given the unique properties of straight plant oils. Figure 2: Plant Oil Viscosity versus Temperature 0 10 20 30 40 50 60 70 80 90 100 0 50 100 150 200 Viscosity (cP) Temperature (C) Soybean Canola Peanut Palm Jatropha Soy Shortening Pre-WVO WVO Diesel Figure 3: (a) Lister Engine Head with Change Over Valve. (b) COV Plug Preheater Design replaces COV Figure 1: Multi-Function Platform (MFP) in Rural Mali 385 395 405 415 425 435 445 18 20 22 24 26 28 30 32 System BSFC (g/kWH) Injector Timing (ºBTDC) 9 MPa IVOP 12 MPa IVOP 15 MPa IVOP 0.50 0.52 0.54 0.56 0.58 0.60 18 20 22 24 26 28 30 32 Equivalence Ratio (Φ) Injector Timing (ºBTDC) 9 MPa 12 MPa 15 MPa 0 2 4 6 8 10 12 18 20 22 24 26 28 30 32 Opacity (%) Injector Timing (ºBTDC) 9 MPa 12 MPa 15 MPa 600 700 800 900 1000 1100 1200 18 20 22 24 26 28 30 32 NOx (ppm) Injector Timing (ºBTDC) 9 MPa 15 MPa 0 1 2 3 4 5 6 7 -90 -60 -30 0 30 60 90 Pressure (MPa) Crank Angle Diesel WVO WVO w/mod 0 0.05 0.1 0.15 0.2 0.25 0.3 -12 -10 -8 -6 -4 ∆P/∆CA Crank Angle Diesel WVO WVO w/mod 0 200 400 600 800 1000 1200 0% 20% 40% 60% 80% 100% System BSFC (g/kWH) Engine Load (%) Diesel WVO WVO w/mod 0.2 0.3 0.4 0.5 0.6 0% 20% 40% 60% 80% 100% Equivalence Ratio (Φ) Engine Load (%) Diesel WVO WVO w/mod 8 108 208 308 408 508 608 708 808 0% 20% 40% 60% 80% 100% CO (ppm) Engine Load (%) Diesel WVO WVO w/mod 0 200 400 600 800 1000 1200 0% 20% 40% 60% 80% 100% NOx (ppm) Engine Load (%) Diesel WVO WVO w/mod A Listeroid CS 6/1, 650 RPM, 4-stroke, 114.3 mm x 139.7 mm bore/stroke, water cooled, IDI diesel engine was used to drive an ST-5 5kW generator head which was loaded by a bank of light bulbs. The ST- 5 was chosen to provide the load to the engine because it is a generator commonly paired with Listeroids throughout the developing world. Engine load was measured from a power meter which logged volts, amps, frequency, and power factor. Engine speed was measured with a Hall Effect sensor and a magnet on the flywheel. The engine coolant system was comprised of a 55 gallon drum filled with water, circulating via a passive thermal siphon cycle. Two type J thermocouples were used to measure water coolant level entering and exiting the engine. A thermostat was used in the engine coolant exit to speed up the rate at which the engine reached steady state (the water coolant temperature leaving the engine was stable and consistent, ~90C). The air intake flow rate was measured via an orifice plate & pressure transducer. A plenum chamber was utilized to attenuate the air flow pulses, sized per SAE standards. Ambient temperature, pressure, and relative humidity were measured near the plenum chamber entrance. A dual fuel tank approach was utilized – the engine was started on diesel to allow the preheater to come up to temperature and shutdown on diesel so that the high pressure fuel lines and pump were purged of the waste vegetable oil (WVO). Each fuel tank had its own filter. The WVO was pre-filtered to 1-micron before being added to the fuel tank. The WVO’s “on-engine” filter (post fuel tank) was 80 mesh, to reduce a pressure drop that could result in starving the engine of fuel. To minimize mixing between the two fuels, the injector’s fuel return line was not routed back to either fuel tank, but instead directly to the juncture of the two fuel lines (SVO and diesel) situated just before the fuel pump. This point is labeled as the “fuel cross” in Figure 4. The opacity of the exhaust was measured using an AutoLogic model # 310- 0332 opacity meter to take readings at 16 Hz that were then averaged across a 10 minute sampling window. To measure gaseous emissions, an Enerac 700 integrated emissions system was used to measure O2, CO, CO2, unburned hydrocarbons (UHCs), NO, NO2, and SO2. Data was collected once per second as parts per million (ppm) or vol%, depending on the concentration. Readings were averaged across a ten minute steady state window (steady state referring to both engine stabilization and the stabilization of the gas analyzer measurements). Figure 4: Experimental Setup (a) (b) A pressure transducer provided by Kulite Semiconductor Products Inc. was used to map the pressure in the pre-combustion chamber versus the crank angle. Three cases were tested: • diesel at standard (unmodified) conditions (room temperature fuel, IVOP of 9 MPa, and injector timing of 20 ºBTDC) •WVO under unmodified conditions (room temperature fuel, IVOP of 9 MPa, and injector timing of 20 ºBTDC) •WVO under modified conditions (fuel heated to 100°C, IVOP of 15 MPa, and injector timing of 25 ºBTDC). The engine was loaded to 75% and run at 650 RPM. The pressure transducer was mounted into the COV plug preventing its use as a preheater, instead the high pressure line was heated electrically. The pressure traces are shown in Figure 5. Ignition is usually identified by heat release, specifically the initial spike on the rate of heat release (RoHR) curve. To accurately calculate the heat release of an IDI engine both the prechamber and main chamber pressures must be measured. Though heat release can be calculated with only one or the other pressure, the calculation can have as large as a 25% error, especially during early combustion. Due to the availability of only the prechamber pressure, heat release was not calculated. Instead, ignition was approximately identified by the rapid change in pressure per change in crank angle (∆P/∆CA). From Figure 6 ignition can be identified in this way for each of the three cases. For the purpose of this study the ignition point was specifically defined as the point where the ∆P/∆CA exceeded 0.15 for 2 crank angle degrees. The ignition point and peak pressure point of the diesel and unmodified WVO cases are nearly identical, though the peak pressure value of the unmodified WVO case is lower. The modified WVO case showed a slightly earlier ignition which resulted in an earlier peak pressure point, and higher peak pressure value. The lower viscosity from preheating, advanced injector timing and increased IVOP are likely all contributing factors to this enhancement to the ignition quality. Figure 5: Pre-combustion Chamber Pressure Traces Figure 6: Change in Pressure per Change in Crank Angle in the pre-combustion chamber In order to tune the engine’s timing and IVOP to optimal conditions for the WVO, three timing settings (20 ºBTDC, 25 ºBTDC, and 30 ºBTDC) and three IVOPs (9 MPa, 12 MPa, and 15 MPa) were tested. Higher IVOPs were tested previously but erratic performance precluded further exploration past 15 MPa. Each of these nine settings was tested three times each at 75% engine load (+/- 1%) and 650 RPM. For all of the 27 tests, the fuel was preheated via the COV plug design detailed above. The summary of the engine’s performance across these tests is detailed below. Error bars signify standard error across the three repetitions, for each of the 9 different settings. Figure 7 shows various engine performance parameters. All four parameters point toward an optimized tuning of the injector timing at 25 ºBTDC with an IVOP of 15 MPa. Figure 7 (a) shows the brake specific fuel consumption for the system (generator and electrical losses were not subtracted). The best performance was found at the timing of 25 ºBTDC with an IVOP of 15 MPa. These results relate well to the findings of other studies where advanced timing of an IDI engine resulted in lower BSFC at low engine speeds . The advanced timing aids combustion because the plant oils have longer ignition delay. The equivalence ratio is the actual fuel to air ratio divided by the stoichiometric fuel to air ratio and is a direct indicator of the quality of combustion. Figure 7 (b) shows the equivalence ratio across the nine test points. In strong agreement with all other engine performance parameters, the equivalence ratio map reinforces the hypothesis that when fueling the Listeroid CS 6/1 on WVO, optimal tuning occurs at the timing of 25 ºBTDC with an IVOP of 15 MPa. The measured opacity values are reported in Figure 7 (c). The results show that at the most advanced injector timing (30 °BTDC), regardless of IVOP, opacity is worse than at stock timing (20 °BTDC). However, in agreement with the engine performance data increasing the IVOP improves the opacity. This vividly shows an engine tuning "envelope" effect - advanced timing and increased IVOP improves combustion, but only to a point, where it then begins to hinder combustion. When injection is advanced too far cylinder conditions are not optimal for good atomization. Temperatures and pressures rise rapidly close to TDC. These conditions are important for appropriate mixing and vaporization which lead to good combustion. Advancing the injection too much introduces the fuel spray into cylinder before these conditions are available. The NOx values between the different IVOPs are not dramatically different, as shown in Figure 7 (d). Increasing the IVOP from 9 to 15 MPa does not then drastically increase NOx. Timing on the other hand does seem to noticeably affect NOx. The envelope is not as sharp, NOx somewhat levels off as timing is increased past 25° BTDC. NOx formation is commonly attributed to the Zeldovich mechanism (a thermal mechanism). With regard to advanced injection timing NO formation may increase in two ways. As timing is advanced ignition delay increases but less so than the actual advancement (in terms of crank angle), resulting in earlier autoignition. Higher NO formation is then related to the longer ignition delay as it allows for more fuel evaporation and mixing in the lean flame region of the spray. But in other spray regions NO may also increase due to higher temperatures. Figure 7: Injection Tuning Performance and Emissions. (a) Brake Specific Fuel Consumption. (b) Equivalence Ratio. (c) Exhaust smoke opacity. (d) Oxides of Nitrogen (c) (d) (a) (b) Figure 8: Multi-Load Performance and Emissions. (a) Brake Specific Fuel Consumption. (b) Equivalence Ratio. (c) Carbone Monoxide. (d) Oxides of Nitrogen. (c) (d) (a) (b) From the IVOP-timing maps and pre-combustion traces it was found that an appropriate modification kit would include a COV plug preheater (injector temperature ~90°C), advanced injection timing (25 ºBTDC), and increased IVOP (15 MPa). This modification kit was then tested across all engine loads, 10%-90%, and compared to diesel and WVO without the modification kit. The diesel and WVO were run under stock conditions - no injector preheating, injector timing of 20 ºBTDC, and an IVOP of 9 MPa. Data was taken for each of the three test cases only after bringing the engine to “steady state” conditions at 650 RPM at each of the five loads tested. Data for each case and for each load was then averaged across a one hour “steady state” window. These results are detailed in Figure 8. Generally diesel performed the best, with the modified WVO case performing better then the unmodified case, as shown by the BSFC. The equivalence ratios for each of the three cases are nearly indistinguishable from one another, though at high loads the unmodified WVO case does have slightly poorer values. At all loads diesel CO is the lowest. At low loads unmodified WVO is relatively flat, then at mid load the CO begins to increase. The modified WVO case is in between the diesel and unmodified case. This is likely because the modification improves mixing and evaporation, but the heavier WVO molecules with their slower evaporation rate still do not burn as completely as the diesel. The NOx values show the diesel to have the highest values across all loads, while both WVO cases are nearly identical. Like the CO, this points to a more complete combustion in the diesel case, which in turn provides hotter temperatures for greater thermal NOx generation. 0 10 20 30 40 50 60 70 0 100 200 300 400 500 Lub. Oil Consump. (mL/Hr) Hour 0 200 400 600 800 1000 1200 0 2 4 6 8 10 0 100 200 300 400 500 CO (ppm) Opacity (%) Hours Opacity CO Figure 9: Lubrication oil Consumption Rate Figure 12: Exhaust Smoke Opacity 7 8 9 10 11 12 13 0 10 20 30 40 50 60 70 80 90 100 0 100 200 300 400 500 Chromium (ppm) Iron (ppm) Hours Iron Chromium -0.410 -0.400 -0.390 -0.380 -0.370 -0.360 0 100 200 300 400 500 Crank Case Pressure (PSI) Hours Figure 11: Engine Crank Case Pressure (gauge) Figure 10: Lubrication Oil Iron and Chromium Content Short-run tests (1 to 10 hours in duration) are valuable for understanding the immediate possibilities of straight plant oil fueling, however questions of durability are left largely unanswered. In order to understand potential endurance issues a 500 hour longevity test was carried out. The engine was run 10 hours a day, five days a week for 10 weeks, at 75% engine load, 650 RPM. Each day the engine started and stopped on diesel (15 minutes each) to allow the preheater to come up to temperature and to purge the fuel lines. In between the startup/shutdown the engine was run for 10 hours on WVO (waste vegetable oil). The 500 hour test lasted approximately 10 weeks. At the start and end of the test a detailed physical inspection was made of the engine in which the cylinder bore was measured and important components were photographed. Every 100 hours during the test the pre-combustion pressure was measured. Every 100 hours emissions were measured. Every 25 hours lubrication oils samples were taken. Every 10 hours (daily) fuel consumption and lubrication oil consumption was measured. Several run time parameters were also measured each second, including various temperatures and pressures throughout the engine. Preliminary analysis has illuminated an important trend: there seems to be an initial "primary" break-in period followed by a longer more subtle "secondary" break-in period. During the first 5 to 50 hours lubrication oil consumption was very high and emissions were poor, but after this "primary" break-in engine performance continued to slowly and steadily improve for another 200 to 300 hours at which point it leveled off. This is seen most clearly in Figure 9. (Note that the first 50 hours are not represented due to lubrication oil measurement and consumption issues.) There is a steady decrease in the lubrication oil consumption rate until the 300 hour point, at which point consumption rates seem to stabilize. This is likely due to the piston rings still breaking in, which is supported by the chromium and iron measured in the lubrication oil shown in Figure 10. The chromium and iron content is high initially, while engine parts are still breaking-in, then a low is reached at 300 hours. The crank case pressure data also supports this two phase break-in observation. The crank case pressure is an indication of blow-by. As the pressure decreases, less blow-by is likely occurring. Figure 11 shows a trend almost identical to the lubrication oil consumption. During the first 300 hours a stead decline is experienced, then at 300 hours the pressure levels out. Figure 12 shows the carbon monoxide emissions and exhaust smoke opacity. Both of these emissions constituents improve during the first 200 to 300 hours and then level off. These emissions are an indication of the incompleteness of combustion, lower values indicate better combustion. These results show that at least during the first several hundred hours of operation, when fueled on straight plant oil via the modification kit, engine endurance issues are not significant. Quite to the contrary - the engine actually experiences a steady improvement in performance during the first few hundred hours of operation. Figure 13: Engine Head and bottom of valves after 500 hours Figure 14: Injector Tip before 100 hours 200 hours 300 hours 400 hours 500 hours Figure 16: Piston Top after 500 hours Figure 15: Valves before and after 500 hour test Intake Valve Before Intake Valve After Exhaust Valve Before Exhaust Valve After Figure 18: Main Bearing Shells after 500 hour test Bottom Top Figure 19: COV plug after 500 hours compared to an unused COV plug After Unused Figure 17: Cylinder after 500 hours Several critical engine components were photographed before and after the 500 hour test. Figure 13 shows the bottom of the engine head and valves after the 500 hour test; though there is carbon buildup, it does not appear excessive. Figure 14 shows the injector tips after each 100 hour period of the test. Every 100 hours the injectors were pulled and cleaned (when the oil was changed). The photos show significant buildup on the injector tips, without the 100 hour cleaning fouling and atomization problems may occur. Figure 15 shows before and after photos of the intake and exhaust valves. The exhaust valve did experience some pitting, which in time could cause seating and sealing problems. Figure 16 shows the piston top after the test. There is carbon buildup but it was surprisingly small. Figure 17 shows a top portion of the cylinder from a side view. This wear is surprisingly noticeable, the rest of the cylinder appeared in good shape but this top portion, where the intensity of compression and combustion takes place, experienced significant wear, scratching and carbon buildup. Figure 18 shows the main bearing shells (the lubrication oil was wiped off for the photo). There was surprisingly little wear, no major scratching, only minor scuffing. Figure 19 shows the COV plug after the 500 hour test compared to an unused COV plug. There is a large amount of carbon buildup on the COV plug. This portion of the COV plug sits inside the pre-combustion chamber near the injector tip. This may point to the need to remove and clean the COV plug regularly, similarly to the cleaning regiment for the injector tip.
