|
B100 COMPARED WITH LOW SULFUR DIESEL
OBJECTIVES
A REVIEW OF LITERATURE The EPA procedures for measuring exhaust emissions from diesel engines require the monitoring of both gaseous and particulate emissions. Exhaust emissions testing for a diesel engine utilizes the constant volume sampling concept of measuring emissions, it requires a positive displacement pump-constant volume sampler (PDP-CVS) or critical flow venturi (CFV) with heat exchanger, and must be connected to a dilution tunnel for sampling particulate emissions. Specific operating requirements are detailed by the EPA for both systems. Compression Ignition (CI) engines require a heated flame ionization detection sampler for hydrocarbon analysis. The HFID must be taken directly from the diluted exhaust stream through a heated probe in the dilution tunnel. Other analyzers are required for carbon monoxide, carbon dioxide, and oxides of nitrogen. One such facility is described in detail by Dunlap et al. (1993).
Emissions testing is a case where the protocol is so restrictive as to limit testing to only a very few laboratories. To counteract the official level, many scientists use techniques of their own to give "relative" data. As the data is examined, it is easy to see that these many different local procedures shed more dark on the question than light. Acceptable research protocols should be suggested within the reach of the average engine test cell that would bring uniformity to this maze of tests. The reports reviewed can be broken into a range of sophistication. Many labs have only an opacity or smoke meter available and report this as emissions testing. Some have maintenance shop type instruments for the regulated emissions total hydrocarbons (HC), carbon monoxide (CO), carbon dioxide (CO2), oxides of nitrogen (NOx), and particulate matter (PM). These instruments are usually operated with little or no calibration and offer, in many cases, too little precision and/or sensitivity to give real "relative" data. The validity of most of those tests should be suspect. The literature clearly shows that smoke meter or opacity data indicate only visible smoke and are not necessarily related to the regulated emission called "particulate matter" (PM). Particulates are defined by the EPA as any diesel exhaust effluent collected on a binderless glass fibre filter in a dilution tunnel at temperatures below 52 degrees C. Hence, particulates include not only solid carbon, but sulfate emissions and soluble organic fractions (SOF). The SOF found in diesel particulates are comprised of unburned fuel and lube oil, and partially oxidized fuel and lube oil (Hunter et al., 1989). Barenescu (1994) and Holmberg and Peeples (1994) report an inverse correlation between emissions of NOx and PM. If an engine is optimized to reduce NOx emissions, each unit of NOx reduction will routinely result in a unit increase of PM. Similarly, a reduction in PM will result in an increase in NOx. Although the inverse correlation holds true for B100 as well, the severity of the trade-off is not as significant as in petrodiesel. The next level of sophistication is from tests which have good quality analyzers with heated lines and, hopefully, good calibration procedures for HC, CO, CO2, and NOx. In every case, these test cells did not have access to a dilution tunnel for measuring PM. In most cases they still use opacity or smoke for PM. These test cells also do not have capability for transient cycles and report steady state data. In some cases they use the 13 mode test, in most cases they use a protocol of their own. This is most generally a constant speed, variable load test. This data is often referred to as being "relative" to the transient cycle required by EPA. The third level of sophistication is a testing lab with a transient capable dynamometer, a dilution tunnel, calibration and all the required sophistication to be called "EPA Certified" for emissions testing. Even in these labs, however, there is not uniformity of data because cycles differ, reference fuels differ, engines differ, methods of changing fuels differ and there is a wide selection of vegetable oil ester fuels from different origins have been used as test fuels. Another problem is the need for "cold start" and "hot start. tests. These requirements are probably reasonable for certifying an engine if the lab is available. However, when one does a comparison of many different fuels, additives, engine adjustments, or blends it becomes an impossible task to create an adequate scientific experiment given a normal budget and time to use a certified emissions testing facility. The last level of emission testing involves laboratories that also measure the unregulated aldehydes, ketones, and polycyclic aromatic hydrocarbons (PAM) that are possibly potential health problems. Only a few studies report these constituents; however, most show B100 produces PAH's lower than those from diesel fuel. The overriding conclusion that was reached from this literature review was that B100 is essentially similar to diesel fuel in their emissions characteristics. Minor engine adjustments can bring about minor changes in results. Several studies show that NOx is increased. That may or may not be relevant, 10 or 15 % of the very small amount of NOx produced was accompanied by an equally important benefit such as reduced HC, CO, or PM. The chemistry of combustion requires emissions. Hydrocarbons are transformed into carbon dioxide and water accompanied by nitrogen in the air passing through the process. Incomplete combustion results in HC and CO. The high temperatures and pressures of combustion form NOx. Reductions in HC and CO must be accompanied by increases in CO2. Sales of diesel fuel in the U. S. are nearly 190 billion litres per year, 53.4 % goes to transportation according to the Energy Information Administration (1989). Diesel engines were designed over many years to operate on petroleum diesel. It is interesting that they perform as well as they do on B100 and that the emissions are comparable to that of commercial diesel fuel for which the engines were designed. In the U.S., the regulations that took effect in 1994 and 1998 pushed the design limitations of heavy-duty diesel engines. The Federal heavy-duty diesel engine emission regulations for 1991 and 1994 prompted a comprehensive re-examination of the diesel combustion process. These regulations require major reductions in the NOx and particulate emissions of diesel engines used in trucks over 3855 kg gross vehicle weight (GVW). According to EPA 86.094-11 exhaust emissions from new 1994 and later model year diesel heavy-duty engines shall not exceed the following: HC 1.3 gm/bhp-hr; CO 15.5 gm/bhp-hr; NOx 5.0 gm/bhp-hr; PM for diesel engines to be used in urban buses 0.07 gm/bhp-hr; PM for all other diesel engines 0.10 gm/bhp-hr. The opacity of smoke emissions shall not exceed the following: 20 % during the acceleration mode; 15 % during the lugging mode; 50 % during the peaks in either mode. Zhang et al. (1988) in two replicates of 200 hour tests reported "based on evaluations of engine performance, wear and combustion chamber component condition as indicators of engine durability, B100appears to be equivalent to number 2 diesel." See reports in left column> Schumacher et al. (1994) reported on tests with two 5.9 L Cummins engines in on-road vehicles that the black exhaust smoke normally observed when a diesel engine accelerates was reduced as much as 86 % when the diesel engine was fueled with 100 % soy B100. Humke and Barsic (1981) in tests with degummed crude soybean oil and crude soybean oil found that nozzle deposit formation after 10 hours caused engine performance to decrease and emissions to merease. Schumacher et al. (1993) studied engine exhaust from four farm tractors using portable exhaust emission testing equipment. As fuel changed from diesel to soy B100, CO dropped from 0.67 to 0.04 %, CO2 exhibited no clear trends, NOx increased from 843 ppm to 1006 ppm, while opacity decreased from 15.6 to 8.0 %. NOx exhaust emissions tended to be lower when the engines were fueled with 10-40 % soydiesel/diesel blends as compared to 100 % diesel or 100 % soy B100. Zubik et al. (1984) used a Case 188D, 4 cylinder, naturally aspirated, 4-stroke diesel engine to study exhaust emissions of diesel fuel, a blend of 25 % sunflower oil in diesel fuel and 100 % sunflower B100. Emissions were measured with continuous flow, heated line analyzing instruments. Particulates were measured with a smoke meter. At a constant speed of 1800 rpm, CO and NOx emissions were essentially the same for all three fuels. Smoke was the lowest for the B100 and highest for pure diesel fuel. Throughout the operating range the pure diesel fuel gave the lowest HC values, however at the lightest loads HC from diesel was above the B100 and blend values. Ishii and Takeuchi (1987) used a one cylinder, Yanmar indirect injection engine to compare emissions for diesel and transesterified curcas oils. They used steady state measurements at 2400 rpm and different load settings. HC and CO were higher at light loads and lower at full load for the ester compared to diesel fuel. NOx was essentially unchanged. There was a slight reduction in black smoke concentration when operating on the B100 curcas oils compared to diesel. Muryama et al. (1985) compared rapeseed and palm oils and their B100 version with diesel fuel for performance, emissions, and carbon deposits in a one cylinder, M diesel engine. It was found that shortening the combustion duration was effective in reducing smoke regardless of fuel. Reece and Peterson (1993) measured smoke opacity using the snap idle test with a 1992 Dodge with a Cummins 5.9 L, DI, turbocharged and intercooled engine comparing B100 and diesel fuel. They found a decrease in opacity of 63 %, and 9 % reduction in hp for the B100 when compared with diesel. Geyer et al. (1984) operated a single cylinder, 0.36 L, DI Diesel engine on certified no. 2 diesel, cottonseed oil, sunflower seed oil, B100 of cottonseed oil, and B100 of sunflower seed oil to compare performance and emission data. A primary objective was to assess aldehyde emissions and the potential health effects of particulate emissions when operating a diesel engine with vegetable oils. The engine was operated at 2400 rpm and load conditions of 1/3, 2/3 and full rack. They collected data for CO, HC, NOx, and total aldehydes, as well as individual aldehyde concentrations from formaldehyde through heptaldehyde. They found that the gas phase emissions were slightly higher for the vegetable oils. NOx was significantly higher for the methyl esters at all rack settings. They found that total aldehydes increased dramatically with the vegetable oil when compared to diesel; the averages for the B100 were slightly higher than the neat oils. The amount of formaldehyde increased with rack setting and was consistently higher for the B100 than for diesel. The diesel had a large increase in aldehydes with rack settings while the vegetable oils did not. Overall the aldehydes averaged 12 % for diesel and 31 % for the vegetable oils. They indicate significant variability in data "this may indicate some problem in the collection or extraction methods". Goering et al. (1984) tested diesel oil, a blend of 25 % sunflower oil in diesel, and sunflower B100. They reported that "all behaved similarly in terms of brake thermal efficiency, ignition delay, exhaust temperatures and exhaust emissions of carbon monoxide and nitrous oxide." B100 produced lower smoke than the blend and considerably less than diesel. Wagner et al. (1985) reported that all regulated emission levels for the ester fuels were similar to diesel fuel except for NOx. They state, "Smoke was definitely less visible under full rack conditions for the B100 compared to diesel fuel, but the butyl ester was greater ...." Esters tested were purchased from Emery industries and an additive package purchased from Ethyl Corporation was added. The test engine was a John Deere 4239TF. Emissions measurements were steady state and no PM equipment was available. Carbon dioxide, carbon monoxide, and oxygen emissions were very similar for all four fuels at both the 2200 and 1500 rpm speed settings. The NOx emissions for B100 measured 2 to 5 times that of the diesel fuel. B100 produced the highest levels, the butyl esters the next highest, with the ethyl ester producing the lowest levels of NOx emissions at both engine speed settings. Visible smoke was less for the B100 at full rack compared to diesel fuel. Clark et al. (1984) used a 3.9 L, John Deere 4239TF, 4-cylinder, direct injected (DI), turbocharged, CI diesel engine on a stationary dynamometer equipped with a constant speed or constant load mode of automatic control to study emissions from B100. Exhaust gas measurements for HC, CO, CO2, NOx, and O2 were made according to SAE J1003. They found that CO was very low; however, the CO emissions were too low to be reliable on their instrumentation. Soy B100 had slightly lower HC levels than the reference diesel fuel. Soy B100 had consistently higher levels of NOx than the reference fuel. Ethyl esters were lower in NOx than B100. Atfuso et al. (1993) used a DL CI diesel engine for emissions tests with the ECE 15, a nonstandard STOP and GO test cycle, and the European 13 mode test procedure. Fuel tested was B100. Tests indicated that B100 promoted a rise in NOx, decreased HC and CO, and reduced smoke. PM produced by B100 in transient cycles was higher than that obtained with diesel fuel. They also measured Polycyclic Aromatic Hydrocarbons (PAM) by taking a sample of diluted exhaust through a glass-fiber filter and then two cold traps in series that were respectively at 0 degrees C and 20 degrees C. The PAH measured with three or four rings, that at ambient temperature are still in the gas phase, were collected into cold traps. PAH with 5 rings or more were mainly collected on the filter. The volume of extract from the filter and the condensed phase were reduced by rotary evaporation and then analyzed by GC-MS equipment. This technique allows for the analysis of more than 20 compounds. They observed that NOx was increased, smoke was lower for the B100, and in effect increased with load. It is well known that PM emissions may not follow the trend of smoke because of the contribution of SOF. Particulate matter is composed of dry soot and a condensed phase. At low loads the contribution of SOF is higher. At light loads B100 produced more SOF and particulate matter than the diesel fuel, while in proximity of full load the trend was the opposite. HC for diesel was higher at light loads and less at high loads. CO was about the same for light and high loads with diesel, but was reduced with B100. NOx production was generally higher for the biofuel, but this effect was more important for the 13-mode cycle than for the transient one. It can be observed that dangerous PAM's did not show significant differences in ECE 15 tests, while they were lower for biofuel on the STOP and GO tests. Vander Griend et al. (1988) used a laboratory and modeling study with the KIVA combustion model to study combustion of vegetable oil fuels. They reported "The B100 appeared to be an ideal fuel under the conditions tested; it displays short ignition delay, sufficient charge for an extensive premixed burn, and a smooth diffusion burn stage. Therefore, there exists an opportunity for the optimization of parameters such as injector timing and swirl. Advanced timing should result in a slightly earlier and more vigorous premixed burn. Since B100 shows signs of a vapor rich core during injection, increased swirl should result in some increase in the amount of premixed burn and in an enhanced rate of diffusion burn." Sholl and Sorenson (1993), in tests with soy B100, found that HC emissions were 50 % less than those of the reference diesel fuel; smoke for soy B100 was generally lower than for the diesel reference fuel. They used a 4-cylinder, 4-stroke, normally aspirated, direct injection diesel engine. They only had 3.8 L and took their data over 100 consecutive engine cycles at 1800 rpms and 50, 150, 300,450, and 600 kPa BMEP. At all loads the HC emissions from soy B100 were about one-half of those from diesel fuel. CO emissions were slightly lower than for soy B100 except for the very lightest loads, where they were slightly higher. NOx was essentially unchanged. Conditions which have the highest peak pressures and rate of pressure rise at a given load and timing tended to have the highest NOx. Scholl and Sorenson show that the differences in NOx emissions are attributable to changes in ignition delay and burning rate only. This study showed reduced NOx at equal BMEP, and more reduction at 5 degrees retarded timing with SME compared to diesel. Bosch smoke numbers were lower for soy B100, but when using a small injector nozzle at retarded timing, the two fuels gave equal results. They state that "the correlation between smoke number and particulate emissions is tenuous at best.... Their conclusions were "In terms of combustion behavior and exhaust emission characteristics, soybean oil methyl ester can basically be regarded as interchangeable with diesel fuel." NOx emissions for the two fuels are comparable (but slightly lower for soy B100) and are related to the peak rate of pressure rise which occurs during the initial portion of the combustion process. Smoke numbers for the soy B100 were lower than those for the diesel fuel. Xiso (1993) tested emissions in a Detroit Diesel 6V-92 engine rated at 277 BHP at 2100 rpm. Fuels used were house number 1 diesel (DF#1) and number 2 diesel (DF#2), both low sulfur fuels. They found that 100 % soy B100 reduced baseline DF#1 and DF#2 total particulates by 35 %, linear with percentage of soy B100 in the fuel. Volatile particulate portions remained at a constant level so particulate reduction was mainly due to the reduction in soot (up to 65 % with 100 % soy B100). NOx was increased by up to 8.5 % for DF#2 blend and 19 % for DF#1 blends. The emission effects of 20 % soy B100 blends were small. Mittelbach and Tritthart (1988), in tests with methyl esters of used frying oil, reported slightly lower HC and CO emissions, but increased NOx. Particulate emissions were significantly reduced, especially the insoluble portion. These two researchers say that the reduction of particulate emissions can be explained by the oxygen content of the used frying oil fuel. They reported that PAH emissions are, in general, slightly higher than those from number 2 diesel fuel, but that the differences are within tolerance limits. Hemmerlein et al. (1991), in tests with neat rapeseed oil, reported increased HC (up to 290 %), increased carbon monoxide (up to 100 %), lower nitrogen oxides (up to 25 %), lower soot emissions (reduced Bosch number of 0.1 to 0.4 depending on engine type), and particulate emissions reduced 30 to 50 % with "divided" chamber and increased 90 to 140 % with a DI engine. The soluble organic fraction of particulate emissions increased by up to 15 % when rapeseed oil was used. They found that total emissions of aldehyde and ketones were 30 to 330 % higher with rapeseed oil; emissions of aromatic hydrocarbons were significantly higher with rapeseed oil. Emission profiles of particulate bound polycyclic aromatic hydrocarbon emissions (PAH) showed anthracene and phenanthrene had the highest concentrations followed by pyrene, chrysene, and fluoranthene. PAH emissions "were reduced to about one-third with big pre-combustion and swirl chamber engines. They increased with rapeseed oil in DI engines and the small swirl chamber engines by 10 to 140 % compared with diesel fuel. Marshall (1993) reported on using B100 in a stationary Cummins L-TOE engine and a Cummins 5.9 L in a Dodge utility tested on a chassis dynamometer. The L-1OE engine was tested using the 13-mode emissions test and the utility with a transient Federal Test Procedure test. They reported "Emission trends were very similar to those reported for other heavy duty diesel engines. That is, emission levels of HC, CO, and particulates decreased significantly for the SoyDiesel blends compared to the diesel fuel. The decrease from the base fuel for these three components was about 5% per 10% incremental addition of methyl soyate to the fuel. NOx emissions tended to increase linearly with the level of methyl soyate in the fuel-approximately 1.4% per 10% incremental addition of soy B100. They hypothesized that the increase in NOx was associated with cetane number or ignition delay characteristics. They reported that fuels with very high cetane numbers are associated with very high NOx emissions. They further comment that there is much information in the technical literature associating increased NOx emissions with lower cetane fuels. However, they say the low cetane number was created by increasing the aromatics content, which was not done in their study. Marshall (1993) found the following for the utility truck. There was no regular, systematic effect of soy B100 level on HC and NOx, particulates increased with the content of soy B100 and reached 70% above the low sulfiur diesel, and CO decreased with increasing soy B100 level. They reported quite high NOx -- 6 grams per mile (gpm) whereas the EPA standard for vehicles of this weight class is 1.7 8pm. Emission levels of HC, CO, and particulates were at or below the EPA standards. In this study, 5 % by volume soy B100 resulted in a 1 °/O increase in BSFC, 18 % reduction in HC, 8 % reduction in CO, 0.7 % reduction in NOx, and a 4.5 % decrease in PM. Marshall (1993) reported that there did not appear to be any fuel effect on aldehyde emissions. The levels with the 5.9 L Cummins engine were quite high for all three fuels - approximately 90 milligrams per mile compared to levels of 3 to 10 mgpm for spark ignition vehicles. One possible reason is the difference in the test procedures for light and heavy duty vehicles. Mittelbach et al. (1985) used a 2.3 L turbocharged, four cylinder, four stroke, DI prototype diesel engine with exhaust gas recirculation (EGR) under transient operating conditions on a chassis dynamometer for emissions tests with B100 and unrefined rapeseed oil(URR). Two U.S. diesel fuels and an Austrian diesel fuel were used for controls. They state, "It is common knowledge that by heating fats and oils various volatile compounds like aldehydes and ketones are produced." When using rape oil and B100 as diesel fuels, the typical smell of burnt fat can be detected, which could originate from unsaturated aldehyde like acrolein. They determined 10 polycyclic aromatic hydrocarbon compounds. For all fuels, the tests showed high formaldehyde (40 %)and very low acetaldehyde. For the sum of C-3 aldehydes, URR was highest and B100 was similar to US fuel B, whereas US fuel A and Austrian diesel fuel were lowest. The transformation of rapeseed oil into the B100 at least reduces the emissions of C-3 aldehydes significantly. In these tests, HC was reduced, CO unchanged, and NOx increased. NOx of the ester fuels was 0.56-0.62 gm/km while NOx of the diesel fuels was 0.47 -0.5 gm/mile. They conclude, "rapeseed oil methyl esters do not lead to irregular PAH-emissions." Tests by ORTECH (Goetz, 1993) using a 20 % blend of B100 and 80 % diesel fuel showed that a 3-degree timing change decreased NOx by 4.6 %, HC by 14.5 %, CO by 9.8 %, and PM by 17.2 % compared to diesel with no timing changes. Humphrey and Schumacher (1994) reported on emissions testing conducted by Cummins Engine Company in an EPA certified testing laboratory. Cummins tested a 5.9 litre, turbocharged, intercooled, direct injected diesel engine. The engine was fueled on B100 and reference diesel fuel. The tests revealed the following significant results:
Manicom et al. (1993) reports that tests were conducted on a 1991, DDC 6V-92 TA engine using the EPA Heavy-Duty Transient Test Cycle. They tested blends of 10, 20, 30, and 40 % soy B100 in comparison with a diesel control fuel. FMD subcontracted ORTECH International to carry out the testing. The diesel control fuel used was ESSO Diesel # 1. One cold start transient test and four hot start transient tests were completed on each fuel. The results showed nearly linear changes with percent fuel. The 40 % blend of B100 had the following results: HC down 39.5%; CO down 29.1%; NOx up 15.0%; PM down 17.8 %. Increased blend levels increased NOx while reducing PM. Proportionately, PM reductions were slightly more than the increase in NOx. The reduction in PM was attributed to the oxygen in the fuel. Total particulates for a 20 % blend are reduced by 11.1 %, but the soluble fraction increases from 40 % to 51 %. Manicom et al. (1993) reported that "the addition of B100 to diesel fuel had the effect of lowering particulates, hydrocarbons, and carbon monoxide emissions while increasing NOx emissions. The optimum blend of methyl soyate and diesel, based on the trade-off of particulates decrease with NOx increase, was 20 % methyl soyate in diesel. Due to the increase in NOx emissions in parallel with decreased exhaust temperatures, it is postulated that the soy B100 is acting as a fuel cetane improver. This improvement in cetane is resulting in a reduced ignition delay time and thus an effective advance in injector timing." Mills and Howard (1983) conducted an investigation of polynuclear aromatic hydrocarbon emissions with unmodified sunflower, rapeseed and soybean oils and B100 from sunflower oils. Three engine load/speed conditions were assessed for each fuel and emission levels were determined for 20 PAH compounds. Samples were generated using a probe in the exhaust pipe. The side stream of exhaust was first filtered to collect particulate bound PAH. Following filtration, the exhaust was transferred to two identical condensation traps for the removal of the vapor phase PAH. PAH profiles were strongly influenced by combustion conditions for all fuels tested. PAH profiles from the B100 from sunflower oil exhaust extracts showed similarities with diesel but were lower in total quantity. At 1500 rpm and no load, total PAH emissions (microgram/m3) for B100 were 289 ug/m3 compared to 3344 ug/m3 for "gas oil" and at 2350 rpm and full load 560 ug/m3 for B100 and 2405 ug/m3 for "gas oil". This work was conducted on a Petter BA2, DI diesel engine. Instruments were available for CO, CO2, and Bosch smoke number. Only 3.78 litres of each fuel was available which restricted the number of test conditions. Test conditions used were 1500 rpm, no load; 2350 rpm, half load; and 2350 rpm, full load. Liotta and Montalvo (1992) used a 1991 prototype Detroit Diesel Series 60 engine installed in a transient capable test cell to study the effect of 8 oxygenates, one of which was B100. They showed that PM emissions was directly related to the concentration of oxygen in the fuel; higher oxygen results in lower PM. PM reductions were accompanied by small increases in NOx emissions. Addition of the oxygenate to the fuel reduced CO and HC emissions; non-regulated aldehyde and ketone emissions were also reduced with the addition of an oxygenate. The composition of the PM from the 11 oxygenated fuels was compared with the reference fuel. The composition remained essentially unchanged. Volatile aldehydes and ketones found in the engine exhaust were measured for each of the fuels. The aldehydes that were speciated and measured include formaldehyde, acctaldehyde, accolein, propionaldehyde, crontonaldehyde, hexanolaldehyde, isobutyraldehyde, and benzaldehyde. Two ketones, acetone, and methyl ethyl ketone also were measured. All of the oxygenated fuels, except for the fuel containing the aliphatic alcohol, had lower aldehyde and ketone emissions. Typically, total aldehydes and ketones were decreased 10 to 25 %. Unfortunately, in this study no data aldehyde and ketone data were taken for the soy B100. This review of literature on emissions test results with vegetable oil fuels brings one to the conclusion that ethyl and methyl esters of vegetable oils are essentially similar to diesel fuel in their emissions characteristics. In some studies, HC is reduced as much as 50 %; CO is reduced by as much as 10 %; NOx and PM are related and tend to change inversely with each other, differing from diesel by at most 10-15 %. Generally, NOx was found to be slightly higher than diesel and PM slightly lower than diesel, although this differs with particular conditions. In this case, NOx was about 0.56 gm/mile[1] for B100 and 0.48 gm/mile for diesel. The last note on this literature review is that of the approximately 20 papers reviewed, no two used exactly the same protocol so as to be directly comparable.
MATERIALS AND METHODS
Chassis Dynamometer
Test Cycles Figure 1. Modified Arterial Cycle. Figure 2. EPA Cycle. Road Load Simulation Figure 3. Coast-down data for 1994 Dodge utility RL = F0 + F1V + F2Vn + Idv/dt + mg sin (grad) where: RL = Road Load (force at roller surface) The EPA and Arterial test cycles do not include any gradients (uphill or downhill) therefore, the Road Load equation becomes:
RL = F0 + F1V + F2Vn + Idv/dt To aid in the development of a set of coefficients, University of Idaho personnel conducted a coast down evaluation of the test vehicle in Idaho prior to the scheduled ETF test date, Figure 3. The coast down test involves accelerating the vehicle to 96 km/h and then letting it coast (decelerate) while recording the time for each speed until the truck comes to a stop. Based on the coast down data, LA-MTA personnel developed a set of coefficients as a starting point for road load model development. Once the vehicle was installed on the dynamometer, coast downs were conducted and the model was refined to match the average on-road data. This refinement process was necessary to "factor out" the internal dynamometer resistance. The coefficients developed for this program are Fo = 22.7 kg, F(1) = 0.237 kg/km, F(2) = 0.000025 kg/km(2) and I = 3583 kg.
Fuels Tested (1) Phillips 2-D low-sulfur diesel control fuel (2-D) The B100 and REE were produced in the Biological and Agricultural Engineering Laboratory at the University of Idaho. Results for fuels 8, 9, and 10 are beyond the scope of this article and are not reported. A complete set of fuel characterization data for each fuel reported is given in Table 1. Table 1. Fuel characterization data. Emissions Analyzers Nitrogen oxides (NOx) were measured using a chemiluminescent analyzer which monitored the chemiluminescent reaction of ozone (03) with nitric oxide (NO). This analyzer measured a wide range of NO and nitrogen dioxide (Nl )2) concentrations with negligible interference from other gas components. The analyzer was equipped with an ozone generator, which produced ozone for the reactions (03) from oxygen, or air using an ultraviolet ozone generating method. Carbon monoxide (CO) and Carbon dioxide (CO2) were measured using non-dispersive infrared (NDIR) detection. The ETF analytical bench was equipped with a total of four NDIR analyzers and was therefore able to measure both low and high concentrations of CO and CO2. The analyzer that measured low CO concentrations incorporated a unique dual-detector optical bent to obtain high sensitivity and accuracy, even in the presence of water vapor, carbon dioxide, and other interfering gases, eliminating the need for chemical reagents and gas dryers to remove water vapor and CO2 from the sample for analysis. Oxygen (02) concentrations were measured during raw, or non-dilute, testing using a magnetopneumatic sensing method. The analyzer provided an accurate and continuous determination of O2 with negligible interference from other gaseous components. Since current CARB testing protocol emphasizes dilute bag results, no raw testing was done on this test vehicle. (Dunlop, 1994)
Test Vehicle The engine was not modified in any way for use with the B100 fuels. The fuel delivery system was modified for convenience of changing fuels between test runs. Fuel delivery and fuel return lines were broken and 3-way, manually operated valves were installed so that stub lines with quick couplers could be installed on one part of the 3-way valves. Individual 19 litre fuel tanks were modified with fuel filter and flexible lines which could be connected to the 3-way valves. During normal operation, fuel was delivered and returned to the vehicle tank. During testing the valves were switched to the external lines to which the correct test fuel was connected. For the tests, the fuel filter assembly mounted on the engine was removed and replaced with an aluminum block with internal connecting ports. This change was necessarily to minimize the amount of fuel in the system when a fuel switch was required. Fuel filters were provided on each fuel delivery tank. Timed practice sessions with fuels of different colors showed that a minimum of 20 s was required for the return lines to be clear of the previous fuel. During actual testing, the return line was directed to a waste fuel tank while the engine was operated for 50 s at which time the return line was directed back into the test fuel tank. The low standard deviations in emissions data between tests of the same fuel were indicative of the success of the procedure for changing fuels.
Vehicle
Fuel Mass Flow Rate
Test Design Table 2. Test plan for emissions tests. A Fisher's Protected LSD analysis using SAS (Statistical Analysis System) was carried out on the data.
Test Procedure
Tests were conducted on March 14-18, 1994. Twenty-six modified arterial cycles and 13 EPA test cycles were completed. On one day, a shortage of gas for the flame ionization detector caused some delay and a need to replicate one run. Otherwise, the tests proceeded very smoothly and with very good repeatability.
PRESENTATION AND DISCUSSION OF RESULTS
Summaries of these data are shown in three ways. First, averages by arterial and EPA cycle for each fuel and for each of the regulated exhaust emissions were computed using Fisher's Protected LSD. This procedure makes adjustments for the unequal sample numbers. Statistical significance is indicated in each of he summary tables. The statistical analysis for days of test showed that the EPA cycle was significantly different from the arterial cycle for all parameters measured (not shown). All of the measured emissions were higher for the EPA cycle than for the arterial cycle. This was probably due to the increased idling time and nonrepetitive nature of the EPA. Figure 4. Total HCs in gm/mile for the modified arterial cycle. Figure 5. Total HCs in gm/mile for the EPA cycle.
Figures 4 through 13 show the data for each test parameter. The graphs show 100% diesel to 100% vegetable oil as the ordinate and the particular emissions parameter reported in gm/mile as the abscissa. Figures 10-14 through 16 show a summary of the data, averaged for B100 in both cycles, with CO and CO2 plotted together on one graph, HC and NOx on a second graph, and PM on a third graph. These data show similar trends for all fuels with a slight reduction in CO and NOx for ethyl ester compared to methyl ester. They also show a significant reduction in HC, CO, and NOx as percent of vegetable oil was increased and a non-significant increase in PM. The PM data had the most variability of the data and an examination of the raw data showed that there was scatter sufficient to neutralize differences.
Figure 6. Carbon monoxide in gm/mile for the modified arterial cycle.
Figure 9. Carbon dioxide in gm/mile for the EPA cycle.
As stated earlier, Barenescu (1995) and Holmberg and Peeples (1994) have indicated that NOx and PM are inversely related. That may explain the apparent difference in results from these tests. More generally, it has been found that esters increase NOx and decrease PM. However, in these tests, the reverse was generally true. One might speculate that this trend was due to the fatty acid constituents of rapeseed B100 tested or that it was a characteristic of this particular engine. In either case, the result was consistent for both B100's and for both cycles tested. Ethyl B100 at the 100% and 50% level were lower in NOx than methyl B100. Repeatability for HC, CO, CO2 and NOx was very good. In these experiments, setting up the test procedure was extremely critical because of the limited number of possible runs in the available time. It may have been more desirable to have replicated each fuel two or three times in succession and to have randomly replicated each fuel in time. However, as a compromise, each fuel was tested in random order and then, during one block, fuels were tested in succession. The data show that the test procedure was successful in preventing mixing of fuels and that the choice of previous fuel did not affect the emissions measured for the succeeding fuel. CONCLUSION
REFERENCES
Barenescu, Rodica. 1994. Engine Warranty Policy. In Commercialization of B100: Establishment of Engine Warranties, National Center for Advanced Transportation, University of Idaho, Moscow, Idaho 83844-1026. Cho, H. K., Smith, R. J., and Marley, S. J. 1988. System Identification of a Diesel Engine: Throttle-Smoke Response. ASAE paper no. 88-1563, ASAE, St. Joseph, MI. Clark, S. J., L. Wagner, M. D. Schrock and P. G. Piennar. 1984. Methyl and Ethyl Soybean Esters and Renewable Fuels for Diesel Engines. JAOCS 61(10):1632-1638. Dunlap, Lauren S., Vince Pellegrin, Randal Ikeda, Ray Wilson, Sylvia Stanley and Harvey Porter. 1993. Chassis Dynamometer Emissions Testing Results for Diesel and Alternative Fueled Transit Buses. SAE Technical Paper Series 931783. SAE, Warrendale, PA 150960001. Dunlap, Lauren. 1994. Final Report to University of Idaho for Emissions Testing Conducted on B100 Fueled Pickup Truck. Los Angeles County Metropolitan Transit Authority Emissions Testing Facility, Los Angeles CA. Energy Information Administration. 1989. Sales of Fuel Oil and Kerosene in 1989. EDOE/EIA, Washington, D.C. Feldman, M. 1991. Optimization of a Direct Injection Diesel Engine for Operation on Winter Rape Methyl Ester. Unpublished M.S. thesis, University of Idaho, Moscow, ID 83844 Geyer, S. M., M. J. Jacobus and S. S. Lestz. 1984. Comparison of Diesel Engine Performance and Emissions from Neat and Transesterified Vegetable Oils. Transactions of the ASAE 27(2):375-381. Goetz, Wendell 1993. Evaluation of Methyl Soyate/Diesel Blend in a DDC 6V-92TA Engine: Optimization of NOx Emissions. Addendum to Final Report 93-E14-21. Ortech International, Mississaga, Ontario. Hemmerlein, Norbert, Volker Korte and Herwig Richter. 1991. Performance, Exhaust Emissions and Durability of Modern Diesel Engines Running on Rapeseed Oil. 1991. SAE Technical Paper No. 910848, SAE, Warrendale, PA 15096-0001. Holmberg, William C. and James E. Peeples. 1994. B100, A Technology, Performance and Regulatory Overview. National Soydiesel Development Board, Jefferson City, MO. Huehn, W., and Sauerteig, J. E. 1989. The New Deutz Particulate Trap System for Trucks and Buses. SAE Technical Paper series no. 902075. Humke, A. L. and N. J. Barsic. 1981. Performance and Emissions Characteristics of a Naturally Aspirated Diesel Engine with Vegetable Oil Fuels - Part 2. SAE Technical Paper 810955, SAE Warrendale, PA 15096-0001. Humphrey, J. Kevin and Leon G. Schumacher. 1994. Review of Literature Related to Fueling Used Engines with Methyl Ester of Soybean Oil. Project Report for the American Soybean Association, St. Louis, MO. Ishii, Yukitsugu and Ryuzo Takeuchi. 1987. Transesterified Curcas Oil Blends for Farm Diesel Engines. Transactions of ASAE 30(3):605-609. Liotta, Frank J. jr. and Daniel M. Montalvo. 1992. The Effect of Oxygenated Fuels on Emissions from a Modem Heavy-Duty Diesel Engine. SAE Technical Paper 932734, SAE, Warrendale, PA 15096-0001. Manicom, Brian, Chris Green and Wendel Goetz. 1993. Methyl Soyate Evaluation of Various Diesel Blends in a DDC 6V-92 TA Engine. Final Report No. 93-E14-21 to Fosseen Manufacturing and Development, Ortech International, Mississaiga. Ontario. Marshall, William F. 1993. B100, Commercialization of a Renewable Fuel. Technical Report No. 1 prepared for the U.S. D.O. E. and NSDB. National Institute for Petroleum and Energy Research, Bartlesville, OK. Mills, G. A. and Howard, A. G. 1983. Preliminary Investigation of Polynuclear Aromatic Hydrocarbon Emissions from a Diesel Engine Operating on Vegetable Oil-based Alternative Fuels. Mittelbach, Martin, Peter Tritthart and Hans Junek. 1985. Diesel Fuel Derived from Vegetable Oils, II:Emission Tests Using Rape Oil Methyl Ester. Energy in Agriculture 4:207-215. Elsevier Science Publisher, Amsterdam. Mittelbach, Martin and Peter Tritthart. 1988. Diesel Fuel Derived from Vegetable Oils, II. Emission Tests Using Methyl Esters of Used Frying Oil. JAOCS 65(7):1185-1187. Muryama, Tadashi, Young-taig Oh, Noboru Miyamoto, Takemi Chikahisa, Nobukazu Takagi and Koichiro Itow. 1984. Low Carbon Flower Buildup, Low Smoke and Efficient Diesel Operation with Vegetable Oils by Conversion to Mono-Esters and Blending with Diesel Oil or alcohols. SAETechnicalPaper841161. SAE, Warrendale, PA1096-0001. Reece, Daryl L. and Charles L. Peterson. 1993. A Report on the Idaho On-Road Vehicle Test with B100 and Neat Rapeseed Oil as an Alternative to Diesel Fuel. ASAE paper no. 935018. ASAE, St. Joseph, Ml. Sholl, Kyle W. and Spencer C. Sorenson. 1993. Combustion of Soybean Oil Methyl Ester in a Direct Injection Diesel Engine. SAE Technical Paper 930934, SAE, Warrendale, PA 15096-0001. Schumacher, L. G., Borgelt, S. C., Hires, W. G., and Humphrey, J. K. 1993. B100 on the Road - A Report from Missouri. ASAE paper No. 93-5017. ASAE, St. Joseph, Ml Schumacher, L. G., S. C. Borgelt, W. G. Hines, C. Spurling, J. K. Humphrey and J. Fink. 1993. Fueling Diesel Engines with Esterified Soybean Oil - Project Update. ASAE Paper No. MC93-101, ASAE, St. Joseph, MI 49085-9659. Schumacher, Leon G., William G. Hires, Steven C. Borgelt. 1994. Fueling Diesel Engine with Methyl-Ester of Soybean Oil. Department of Agricultural Engineering, University of Missouri Columbia, MO. Stansvaag, John-Mark, Oren C. N. 1993. Clean Air Act: Law and Practice Vander Griend, L., M. Feldman and C. L. Peterson. 1988. Modeling Combustion of Alternate Fuels in a DI Diesel Engine using KIVA. Transactions of the ASAE, ASAE, St. Joseph, MI. Wagner, Larry E., Stanley J. Clark and Mark D. Schrock. 1985. Effects of Soybean Oil Ester on the Performance, Lubricating Oil and Wear of Diesel Engines. SAE Technical Paper 841385. SAE, Warrendale, PA 15096-0001. Xiao, Yun. 1993. Report to the National SoyDiesel Development Board, Jefferson City, Missouri 65110. Zhang, Q., M. Feldman and C. L. Peterson. 1988. Diesel Engine Durability when fueled with Methyl Ester of Winter Rapeseed Oil. ASAE paper 88-1562. ASAE, St. Joseph, ML Zubik, J. S. C. Sorenson and C. E. Goering. 1984. Diesel Engine Combustion of Sunflower Oil Fuels. Transactions of the ASAE 27:1252-1256. [1] EPA suggests grams per mile abbreviated 8pm be used in reporting emissions data. Because of the possible confusion with gallons per mile the abbreviations chosen in this paper are those used by Dunlap et al. (1993) and Dunlap (1994) gm/mile to clearly indicate grams per mile. |
Comprehensive tests were perfomed with B100 on an on-road vehicle in cooperation with
the Los Angeles County Metropolitan Transit Authority emissions test facility.
All tests were with a transient chassis dynamometer. Tests included both a
double arterial cycle of 768 s duration and an EPA heavy duty vehicle cycle of
1,060 s duration. The test vehicle was a 1994 utility truck with a 5.9 L
turbocharged and intercooled, direct injection diesel engine. B100 and blends were compared with low sulfur diesel
control fuel. Emissions data includes all regulated emissions: hydrocarbons
(HC), carbon monoxide (CO), carbon dioxide (CO2), oxides of nitrogen (NOx) and
particulate matter (PM).
Revised 20010118