Evaluation of Emissions and Performance of NJ TRANSIT Diesel Locomotives with B20 Biodiesel Blends

Transcription

1 Evaluation o Emissions and Perormance o NJ TRANSIT Diesel Locomotives with B2 Biodiesel Blends Final Report: December 1, 29 Anthony J. Marchese, Principal Investigator Krishan K. Bhatia, Co-Investigator Robert P. Hesketh, Co-Investigator David McKenna, Graduate Student Department o Mechanical Engineering Rowan University 21 Mullica Hill Rd. Glassboro, NJ

2 Evaluation o Emissions and Perormance o NJ TRANSIT Diesel Locomotives with B2 Biodiesel Blends Final Report: December 1, 29 Executive Summary This report summarizes the inal results o an NJDEP grant to Rowan University, which was established to quantiy the exhaust emissions and perormance characteristics o 2% soy methyl ester biodiesel blends (B2) in diesel locomotives representative o the NJ TRANSIT commuter leet. Testing was perormed with #2 diesel summer blend, #2 diesel winter blend, ultra low sulur diesel (ULSD) summer blend, ULSD winter blend and B2 blends with each o these uels. Tests were perormed on two dierent diesel locomotive types to determine the dierences in perormance and emissions between older and newer locomotive engines when operating on biodiesel blends. Speciically, tests were perormed on a GP4FH-2 locomotive equipped with an EMD engine manuactured rom a 196 s design and a recently manuactured ALSTOM PL42AC locomotive equipped with an EMD engine. The tests were perormed by operating the diesel engines statically (using a load bank) over the ull test matrix o 8 uels. During each test, brake speciic exhaust emissions and uel consumption were computed or each uel blend using the Line-Haul Duty Cycle as outlined in the CFR Part 4 Title 92 Federal Test Procedure. The Line-Haul Duty Cycle was slightly modiied given our inability to measure engine horsepower at idle conditions. Each uel/locomotive test combination was perormed 3 times to ensure repeatability. To accurately quantiy the exhaust emissions, measurements were made using a Sensors SEMTECH-D mobile emissions analyzer to measure CO, CO 2, NO, NO 2, O 2, and total unburned hydrocarbons (HCs), along with a Wager 65RR Railroad Opacity Meter. Instantaneous uel consumption was monitored using two AW Company JV-KG positive displacement low meters, which measure the supply and return uel low rate, respectively. GP4FH-2 Locomotive Results For the GP4FH-2 locomotive, all B2 blends resulted in comparable horsepower, decreased exhaust opacity and decreased greenhouse gas CO 2 emissions with respect to pure petroleum diesel. Using greenhouse gas emission actors or soy biodiesel, the estimated results suggest that NJ TRANSIT would realize an approximately 7% decrease in greenhouse gas emissions using B2 blends in their older locomotive leet. The summer B2 blends exhibited NO x and unburned hydrocarbon (HC) increases o up to 15% and CO decreases o up to 43%. The winter B2 blends contained signiicant quantities o kerosene in order to meet NJ TRANSIT s winter cloud point requirement as F max, which is the highest ambient temperature or a uel to become cloudy. The winter B2 blends exhibited decreases in NO x o up to 1%, along with increased HC o up to 5%. The carbon monoxide emissions varied widely or the winter blends, with B2/#2 winter blend exhibiting a decrease in CO o 13% and B2/ULSD/winter showing increases in CO o 4%. The emissions results or the GP4FH-2 locomotive are summarized in the Table E1. The total mass emissions or each pollutant are calculated based on a weighted average o all notch settings using weighting actors developed rom actual NJ TRANSIT notch data. The HC emissions measurements with B2-Summer blend was not perormed because o equipment ailure during the tests. Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 2

3 Table E1. Summary o emissions measurements rom GP4FH-2 locomotive. Mass emissions are based on a weighted average o all notch settings using weighting actors developed rom NJ TRANSIT notch data. CO 2 NOx HC CO Estimated Soy Biodiesel Greenhouse CO 2 Fuel kg/hr % change g/hr % change g/hr % change g/hr % change kg/hr % change #2 Summer ULSD-Summer % % % % % B2-Summer % % % % ULSD-B2-Summer % % % % % #2 Winter ULSD Winter % % % % % B2-Winter % % % % % ULSD-B2-Winter % % % % % PL42AC Locomotive Results For the PL42AC locomotive, all B2 blends also resulted in comparable horsepower, decreased exhaust opacity and decreased greenhouse gas CO 2 emissions with respect to pure petroleum diesel. Using greenhouse gas emission actors or soy biodiesel, the estimated results suggest that NJ TRANSIT would realize an approximately 2% decrease in greenhouse gas emissions using B2 blends in their newer locomotive leet. For the PL42AC locomotive, the summer B2 blends exhibited NO x decreases o 15.5%, unburned hydrocarbon (HC) decreases o 6.8 %and CO decreases o 27.3%. The winter B2 blends contained signiicant quantities o kerosene in order to meet NJ TRANSIT s winter cloud point requirement o F max as the highest ambient temperature or a uel to become cloudy. The winter B2 blends exhibited decreases in NO x o up to 17.8% and decreased CO o 5.9%, but showed an increase in HC o up to 29%. The emissions results or the PL42AC locomotive are summarized in the Table E2. The total mass emissions or each pollutant are calculated based on a weighted average o tested notch settings using weighting actors developed rom actual NJ TRANSIT notch data. The CO emissions measurements with ULSD-B2-Winter blend could not be not perormed because o equipment ailure during the tests. All emissions measurements were carried out with this locomotive at idle, and notches 5-8, because o technical limitations. Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 3

4 Table E2. Summary o emissions measurements rom PL42AC locomotive. Mass emissions are based on a weighted average o tested notch settings using weighting actors developed rom NJ TRANSIT notch data. CO 2 NOx HC CO Estimated Soy Biodiesel Greenhouse CO 2 Fuel kg/hr % change g/hr chang e g/hr % change g/hr % change kg/hr % change LSD Summer LSD-B2-Summer % % % % % LSD Winter ULSD Winter % % % % % LSD-B2-Winter % % % % % ULSD-B2-Winter % % % % Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 4

5 1. Project Personnel The study described herein was perormed by a team o aculty, graduate students and undergraduate students at Rowan University. The aculty and graduate students are listed below: Pro. Anthony J. Marchese *, PI Pro. Krishan Bhatia, co-pi Pro. Roberth Hesketh, co-pi Tim Vaughn, Engineering Technician David McKenna, graduate student David Martinez, graduate student Chris Rowen, graduate student Kyle Fitzpatrick, graduate student 2. Introduction Biodiesel is a renewable alternative uel that is produced rom raw animal or vegetable ats via a chemical reaction with an alcohol (typically methanol). This reaction results in a mixture o methyl esters o varying carbon chain length and this mixture o methyl esters is what is known as biodiesel [1,2]. Biodiesel is generally considered to be a renewable uel because the carbon present in the vegetable or animal eedstocks originates rom carbon dioxide already present in the air. However, it should be noted that a complete lie cycle analysis that takes into account the greenhouse gas emissions rom agricultural, transport and processing activities shows that biodiesel is not completely carbon neutral [3]. Biodiesel has also been shown to be highly eective in reducing carbon monoxide (CO), unburned hydrocarbons (HC), and particulate matter (PM) emissions rom diesel engines [4]. Results have shown, or example, that 1% biodiesel can reduce PM emissions by as much as 5% with respect to petroleum diesel. Reduction in PM emissions is a key beneit and motivation or using biodiesel in public leets such as buses or locomotives since diesel exhaust has been classiied as a probable human carcinogen by the World Health Organization and the USEPA, and recent studies have linked diesel PM to heart disease [5]. The NJ TRANSIT diesel locomotive leet currently consumes 12.3 million gallons o petroleum diesel per year. Because o rising petroleum prices, concerns about disruptions in supply and environmental considerations, NJ TRANSIT was currently evaluating the easibility o supplementing a substantial component o their total diesel consumption with biodiesel. Since biodiesel is typically blended at 2% with petroleum diesel (such a ormulation is designated as B2), a ull deployment o B2 by NJ TRANSIT would reduce NJ TRANSIT s ossil uel consumption by nearly 2.5 million gallons per year. Such a ull scale deployment could potentially have substantial impact on the environment as it has been shown in other literature that biodiesel reduces emissions o PM, CO, HC s, greenhouse gases (CO 2 ) and sulur containing compounds. The purpose o the proposed investigation was to demonstrate the use o biodiesel/ulsd blends in NJ TRANSIT diesel locomotives. The investigation was perormed at the NJ TRANSIT Meadows Maintenance Complex and on selected NJ TRANSIT rail lines. At the inception o the study, NJ TRANSIT maintained an inventory * In January 28, PI Marchese departed rom Rowan University to accept a aculty position at Colorado State University. Marchese continued to work on the project on an advisory basis until its completion in January 29. Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 5

6 o 122 diesel locomotives, the majority o which were line haul locomotives with head end power units. Most o the line haul locomotives are older locomotives based on the General Motors Electro-Motive Division GP4 platorm with model types GP4PH, GP4FH and F4PH, which are powered by EMD 645 Series engines. NJ TRANSIT also has recently begun to acquire newer PL42AC line haul locomotives manuactured by ALSTOM, which are powered by EMD 71 series diesel engines. In addition to the line haul locomotives, which are used or passenger service, NJ TRANSIT also maintains 4 SW15 switcher locomotives used on-site at the Meadows Maintenance Complex in Kearny, NJ and 4 GP4 work train locomotives, which are used or nonpassenger rail activities. A summary o NJ TRANSIT s current leet is as ollows: 19 Line Haul Locomotives with EMD engine 5 Line Haul Locomotives with EMD engine 4 Switcher Locomotives with EMD 567 engine 4 Work Locomotives with EMD engine The line haul locomotives are also equipped with an additional auxiliary diesel engine called the Head End Power Unit (HEP), which is used to power auxiliary systems needed or passenger rail service. The HEP is typically a 6-cylinder diesel engine (Cummins or Caterpillar) similar to those used in heavy-duty on-road diesel vehicles. It should be noted that the HEP engine and the main engine share a common uel tank so both engines were tested with biodiesel blends. Perormance o the HEP engine was less o an issue as these types o engines have been tested extensively with B2. With the exception o the EMD SW15 switcher locomotives, all o the diesel locomotives currently operated by NJ TRANSIT are powered by turbocharged EMD engines or EMD The ormer engines are used in the older GP4 series locomotives, while the latter are used in the newer PL42AC locomotives. The EMD engine is manuactured by EMD (ormerly the Electro-Motive Division o General Motors. Although the EMD is available in either turbocharged or roots blower coniguration, all o the EMD engines operated by NJ TRANSIT are turbocharged. The EMD engine is a 2-stroke, 16-cylinder engine. Each cylinder has a displacement 645 in³ with a bore o 9.1 inches and a stroke is 1 inches. The compression ratio is 16 to 1. The EMD is rated at 3 HP. The EMD engine is also a turbocharged, 16-cylinder engine but has a displacement o 719 in 3 per cylinder. It has a bore o 9.1 inches and a stroke o 11 inches. It also has a compression ratio o 16 to 1. Table 1 o the Appendix contains detailed speciications o both the 645 and 71 series engines. 2.1 Emissions rom Diesel Locomotives Although a set o minimal emissions standards were put in place in 1997 or new diesel locomotive engines, diesel locomotive engines continue to be signiicant contributors to air pollution in many o our nation's most populated areas. Speciically, they continue to emit large amounts o nitrogen oxides (NO x ) and particulate matter (PM), both o which have been shown to contribute to serious health problems [6]. The relative contribution in emissions rom diesel locomotive and marine engines is expected to grow due to the expected uture growth in the use o these engines. The USEPA has estimated that, without new controls, locomotives and marine diesel engines will contribute 27% o the total NO x emissions and 45% o the total ine diesel particulate matter (PM2.5) emissions rom all mobile sources combined [6]. Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 6

7 Millions o Americans continue to live in areas with unhealthy air quality that may endanger public health and the environment. Exhaust rom diesel engines contributes to unhealthy concentrations o ine particles and ozone which are linked to serious problems, including premature death, increased risk o lung cancer, heart disease, aggravated asthma and other respiratory conditions. In addition, PM, NO x, and ozone adversely aect the environment in various ways including visibility impairment, crop damage, and acid rain. Although locomotive engines being produced today must meet relatively modest emissions requirements set in 1997, diesel locomotive engines continue to be signiicant contributors to air pollution in many o our nation s most populated areas. Speciically, they continue to emit large amounts o oxides o nitrogen (NOx), and particulate matter (PM), both o which have been shown to contribute to serious health problems. In 24, as part o the Clean Air Nonroad Diesel Rule, the USEPA inalized a new standard or nonroad diesel uel that will decrease the allowable levels o sulur in uel used in locomotives by 99 percent. The majority o diesel locomotives currently in service by NJ TRANSIT are approximately 3 years old and these engines are not covered by the new USEPA standards unless these locomotive engines are rebuilt, which will trigger upgrading to proposed standards. Moreover, attempts to retroit the older locomotive engines with exhaust ater-treatment technology (i.e. particulate ilters) would be costly. Thereore, in 27, a grant was awarded rom the NJ DEP to Rowan University to perorm an experimental study on exhaust emissions o NJ TRANSIT diesel locomotives operating on biodiesel blends with the ollowing objectives: To determine the emissions beneits (i.e. reduced CO, HC s, soot) o using biodiesel in the NJ TRANSIT locomotive leet, To quantiy any emissions drawbacks, such as increased NO x emissions, To determine i there are any other potential diiculties in using biodiesel, such as reduced power and/or storage/handling issues. This report summarizes the activities that have been perormed on behal o this grant during the period rom June 27 to January Test Plan Tests were perormed on two dierent diesel locomotive engines to determine the dierences in perormance and emissions between newer and older locomotive engines when operating on biodiesel blends. The GP4FH-2 is an older locomotive built in It has a 3 HP Turbocharged EMD engine, which is a V-16 coniguration with a 645 CID. The PL42AC is a newer locomotive manuactured by ALSTOM in France. NJ TRANSIT began accepting delivery o these locomotives in 25. These locomotives are powered by a 365 HP Turbocharged EMD 16-71, which is a V-16 coniguration with 71 CID cylinders. Tests were perormed with 8 dierent uel blends: Summer Blend (#2 diesel, 5 ppm sulur) Winter Blend (3% kerosene, 7% #2 diesel) ULSD Summer (<15 ppm sulur) ULSD Winter (4% kerosene, 6% ULSD) B2 Summer (2% biodiesel, 8% #2 diesel) B2 Winter (2% biodiesel, 56% kerosene, 24% #2 diesel) Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 7

8 B2 ULSD Summer (2% biodiesel, 8% ULSD) B2 ULSD Winter (2% biodiesel, 56% kerosene 24% ULSD) Tests were perormed by operating both diesel engines statically (using a load box) over the ull test matrix o 8 uels. During the tests, brake speciic exhaust emissions and uel consumption were computed or each uel blend using the line-haul and switcher duty cycles as outlined in the CFR Part 4 Title 92 Federal Test Procedure. Each locomotive/uel blend combination was tested 3 times, resulting in a complete static test matrix o 48 tests. The completed test matrix can be viewed in the Appendix in Table Instrumentation and Equipment The locomotive s gaseous emissions were measured with the Semtech-D mobile emissions analyzer manuactured by Sensors, Inc. An FEM-3 Fuel Monitoring System manuactured by AW Company was used to measure the instantaneous volumetric uel low rate. An opacity meter rom Wager Company was used to quantiy exhaust opacity. An Avtron Load Box at NJ TRANSIT MMC was used to dissipate the electrical power produced by the diesel locomotive alternator during static testing. Instantaneous engine horsepower was calculated by measuring the instantaneous alternator voltage and current. The instantaneous uel low rate data, opacity data, load box data and additional temperature data were acquired and stored using an Agilent 3497A data logger and PC notebook. 4.1 Gaseous Emissions Concentrations A SEMTECH-D portable emissions measurement system by Sensors Inc. was used to measure real-time exhaust emission concentrations or CO, CO2, NO, NO2 and HC as well as ambient temperature and relative humidity. The SEMTECH-D is a portable PCbased data acquisition system capable o measuring emission levels along with several vehicle and engine parameters. The Semtech-D includes the ollowing measurement subsystems: heated lame ionization detector (FID) or total hydrocarbon (THC) measurement, non-dispersive ultraviolet (NDUV) or nitric oxide (NO) and nitrogen dioxide (NO 2 ) measurement, non-dispersive inrared (NDIR) or carbon monoxide (CO) and carbon dioxide (CO 2 ) measurement, and electrochemical sensor or oxygen (O2) measurement The Semtech-D also contains a module or wireless communication or remote monitoring using PC or a personal digital assistant (PDA), a global positioning system (GPS), and a weather probe or ambient temperature and humidity measurement. For the locomotive application, a bracket was abricated to locate and attach the heated sample line and sample probe to the exhaust stack to secure it during tests and ensure it is in the same location or every test. Figure 1 shows the SEMTECH-D in test coniguration during a static load box test. A heated sample line is run rom the ront o the machine to the exhaust stack. A bracket was abricated to locate and attach the sample probe to the exhaust stack to secure it during tests and ensure it is in the same location or every test conducted. Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 8

9 Figure 1 (a) Static load bank emissions measurements on GP4FH-2 locomotive using SEMTECH-D portable emissions measurement system and (b) opacity meter and SEMTECH-D heated sample line installed on the exhaust stack o the GP4FH-2 locomotive. 4.2 Fuel Mass Flow Rate To calculate the mass emissions rate o each gaseous species (g/hr) rom the measured concentrations in ppm (parts per million), total exhaust low rate or uel mass low rate must be measured. Because o the large size o the exhaust stack, measuring exhaust low rate is extremely diicult so the decision was made to employ the latter method and measure uel consumption rate. A complete Fuel Monitoring System rom AW Company was used to measure uel low rates and uel consumption. The system uses two JVA- 3KG positive displacement low meters and an FEM-3A2 Flow Transmitter. The FEM- 3A2 low transmitter has both a display or instantaneous readings and analog output (4 to 2 ma) or external data logging. Type K thermocouples were installed directly upstream o the low meters, which along with density vs. temperature uel property data will be used to convert volume low rate into mass low rate. The low meter controllers output a 4-2mA signal to an Agilent 3497A data logger that sends it to a PC where the current signal is converted to a volume low rate. Measuring the volumetric low rates, then subtracting the return line low rate rom the supply line low rate produces the total volumetric uel consumption rate. The volumetric low rate is then converted to mass low rate by multiplying by the uel density. To account or variation in uel density with temperature, a thermocouple was placed in the uel supply line to measure the uel temperature directly upstream o the supply line low meter. For the irst set o tests, only the supply line uel temperature was measured. Moreover, the AW Company Fuel Monitoring System technique o subtracting volume low rates implicitly assumes that the uel supply and return temperatures are identical. Accordingly, a second thermocouple was installed into the uel return line, which enabled the team to account or density changes between the uel supply and return lines. 4.3 Exhaust Opacity As shown in Fig. 1(b), the opacity o the locomotive exhaust plume was quantiied using a Wager 65RR Railroad Opacity Meter. The 65RR is a complete system with a custom mounting rame or locomotive applications, heavy-duty optics to withstand high temperatures and maintain accuracy while spanning longer distances and the ability to send data to an acquisition system, or make instantaneous measurements. The emitter Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 9

10 and detector o the opacity meter are mounted on an aluminum bracket that is bolted to the roo o the locomotive surrounding the exhaust stack. This meter also utilizes a compact control unit that outputs a -1V signal to the Agilent 3497A data logger that correlates to -1% opacity. The unit is calibrated in the ield by adjusting the emitter and detector such that the beam is perectly aligned. Ater alignment, calibration points are taken or % opacity and 1% opacity, respectively. The latter calibration point corresponds to the detector being completely covered. The opacity readings are corrected to % and 1% opacity at each o these calibration points via adjustable potentiometers on the emitter and detector. 4.4 Load Box Static locomotive testing is perormed using an AVTRON Load Box located at the NJ TRANSIT Meadows Maintenance Complex (MMC). The Load Box dissipates the electrical power produced by test locomotive s main alternator, thereby simulating an actual operating load while the locomotive remains stationary. The magnitude o the electrical power dissipation is quantiied by measuring the main alternator voltage and current during static loading. Using Association o the American Railroad correction actors or air temperature, altitude, uel temperature and uel speciic gravity, the measured voltage and current are used to calculate the horsepower produced by the main traction engine. The alternator voltage and current are calculated rom reerence voltage and current readings taken rom the Morrison Knudsen Corporation W14-MK card located in the cabin o the locomotive by the Agilent 3497A data logger and converted to horsepower using a dedicated PC that is coupled with the data logger. A complete instrumentation list is included in Table 2 o the Appendix. 5. Test Logistics Since the locomotives that were tested are part o the NJ TRANSIT operating leet, the logistics involved in conducting each test were substantial. Speciically, prior to each test, the speciic uel blend had to be ordered rom Sprague Energy and a date or delivery to MMC had to be scheduled. Because there were no holding tanks available or these tests, delivery o the test uel had to be coordinated with NJ TRANSIT so that the locomotive could be pulled out o service prior to the uel delivery and the test uel could be pumped directly into the locomotive uel tank. Since the locomotive is a revenue generating system, it is typically only available or a window o 3 to 5 days or each test. Because o this narrow window o opportunity, any locomotive mechanical issues or test instrumentation ailures result in a high risk o aborting an entire 3-point test sequence and subsequent loss o 1 gallons o diesel uel. During the course o the grant, we experienced clogged uel injectors, heavy rain and several emission analyzer ailures (a broken heated sample line and a system leak). In each case, NJ TRANSIT personnel were extremely accommodating and they did everything in their power to hold the locomotive out o service or extended periods o time. By the end o the matrix, the Rowan team, coupled with helpul NJ TRANSIT employees, were able to execute up to 6 tests in a single day, testing two dierent uels. Through various correspondence and conerence calls, there was a mutual decision to reduce the matrix rom the one that was originally proposed. An increase in uel costs o nearly 1% rom the time the grant began, in addition to time constraints rom all three involved parties, made it necessary to reduce the matrix to what was deemed as absolutely necessary or the study. Table 3 shows the completed tests o the inal matrix. Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 1

11 6. Test Procedure Each static test was conducted according to the test protocol summarized in Table B o o CFR Title 4, Part 92 or locomotive testing. This table is reproduced in the Appendix as Table 4. The test procedure entails operating the locomotive or a speciied period o time in each notch setting, while the power rom the main alternator is dissipated at the load box. During each o the test modes, exhaust gas concentrations, exhaust opacity, uel low rate, uel temperature, alternator voltage and alternator current data are acquired and stored at 1 Hz. Data rom the irst two test modes (Notch 8 and lowest idle) are used to ensure the locomotive is at adequate operating temperature. The locomotive is then operated in each notch (idle to notch 7) or 6 minutes each and inally in notch 8 or 15 minutes. Figure 2 is a plot o instantaneous concentration o NOx, CO and CO 2 as measured by the SEMTECH-D or an entire static test sequence or the GP4FH-2 locomotive operating on #2 summer blend diesel uel. Figure 3 is a plot o instantaneous horsepower, uel low rate and exhaust temperature or the same test sequence Concentration (ppm) CO ppm NOx CO2 % Time (s) Figure 2. Instantaneous concentration o NOx, CO and CO 2 during a static load box test sequence or the GP4FH-2 locomotive operating on #2 summer blend diesel uel. Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 11

12 Figure 3. Instantaneous horsepower, uel volume low rate and exhaust temperature during a static load box test sequence or the GP4FH-2 locomotive operating on #2 summer diesel uel. The SEMTECH-D portable emissions measurement system measures exhaust emission concentrations in parts-per-million (ppm). By perorming an overall carbon balance, it is possible to convert each gaseous concentration into a uel speciic mass emission. For example, the equation below shows such a conversion or CO: g_co ( CO) MW CO CO s = ( CO 2 ) + ( CO) + ( THC) g_uel MW uel where CO s is the uel speciic emissions o CO, (CO) the concentration o CO in ppm, (CO 2 ) the concentration CO 2 in ppm, (THC) the concentration o total unburned hydrocarbons in ppm, MW CO the molecular weight o CO in g/mol and MW uel the molecular weight o the uel based on its C:H ratio. It should be noted that the MW uel term is not the average uel molecular weight, but rather the molecular weight o a CH x, where x is the hydrogen to carbon ratio. The uel speciic emissions or each measured gaseous species are then converted to a mass emissions rate (g/hr) by multiplying by the measured uel mass low rate. As described above, the mass low rate o the uel is simply the dierence o the volumetric low rates multiplied by their respective densities as show in the ollowing equation:. g.. = ρ S V S hr m CO ( ρ R V R )CO s Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 12

13 where m& CO is the mass emission rate or CO, ρ the density o supply uel, ρ the density S R o return uel, V S is the volume low rate o supply uel, V R is the volume low rate o return uel, and CO s is the uel speciic emissions o CO. 6.1 EPA Duty Cycles and Calculation o Overall Brake Speciic Emissions (g/bhp-hr) To calculate break speciic emissions (g/bhp-hr), the mass emissions rate is then divided by the Brake Horsepower (BHP) or each notch, which is calculated rom the measured alternator current and voltage. Calculation o brake speciic emissions is necessary to compare the measured emissions with current and uture EPA standards. For direct comparisons with EPA standards, a weighted average o the brake speciic emissions measured during each mode in Table 4 is computed. The EPA deines two separate cycles: the Line-Haul Duty Cycle and the Switcher Duty Cycle. Each o these duty cycles has their own weighting parameters. The weighting parameters or each notch are shown in Table 5 in the Appendix or both EPA cycles. It should be noted that the highest weighting parameters or the Line-Haul Duty Cycle correspond to normal idle and notch 8, respectively. Based on conversations with NJ TRANSIT personnel, as well as actual notch data supplied to us rom NJ TRANSIT, the locomotives do spend the overwhelming majority o their operating time in idle and notch 8. Accordingly, the Line-Haul Duty cycle is the more relevant o the two EPA cycles or the purposes o comparing the emissions rom biodiesel in NJ TRANSIT locomotives to current and uture EPA standards. In the present study, a Modiied Line-Haul Duty Cycle was used, wherein exhaust emissions during idling conditions are not considered (See Table 5). The idle emissions data were not used in the overall duty cycle calculations because the horsepower is not directly measured during idle since the main traction alternator is not engaged. Rather, during idle, the engine horsepower must be estimated based on estimated parasitic loads (cooling ans, etc.). Furthermore, the measured uel mass low rate has its highest uncertainty during idling conditions as a result o subtracting supply and return uel low rates which are very close in magnitude during idle conditions. 6.2 Estimated Total Annual Mass Emissions o Criteria Pollutants and Greenhouse Gases Ater measuring the mass emissions (g/hr) o each gaseous species at each notch or every uel/locomotive combination, it was possible to estimate the potential impact o B2 biodiesel blends on total annual mass emissions [kg/year] o criteria pollutants and greenhouse gases or the NJ TRANSIT leet. These calculations were perormed by multiplying the measured mass emissions (g/hr) at each notch by a weighting actor or that notch which was developed rom actual NJ TRANSIT notch data summarized in Table 6 in the appendix. The resulting NJT-weighted mass emissions (g/hr) were then multiplied by an estimated annual operating hours per locomotive and number o locomotives. 7. Test Results: GP4FH-2 As summarized in Table 3, twenty eight ull test sequences were conducted on locomotive 4142, model GP4FH-2. During the period o June 18 to June 22, 27, three tests were conducted on locomotive 4142 with #2 summer blend. During the period o August 2 to August 3, 27, three tests the irst biodiesel blend tests were conducted with B2/#2 summer blend. As discussed below, these tests showed decreased horsepower due to injector clogging and thereore were not counted against Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 13

14 the test matrix. On September 2 and September 21, a second set o B2/#2 summer blend tests were completed successully. A ourth test was completed during this last period because it was discovered that the uel low meter cables were connected improperly during the irst ew notches o the irst test on September 2. During November 8-9, three successul tests were completed using Summer ULSD uel. On November 3, three Summer ULSD B2 tests were made with the 4142 locomotive. Ater experiencing a ew equipment and locomotive ailures, testing resume in March 28, testing #2 Winter baseline rom March 4-6. Three successul Winter B2 runs were completed on April 22. Winter ULSD was tested in three consecutive runs on June 27. The inal test with the 4142 locomotive using Winter ULSD B2 was completed on July 3, 28. A brie summary o the tests conducted to date ollows. The baseline #2 summer tests conducted in June were highly successul. All test instrumentation unctioned as anticipated and the locomotive was in excellent running condition. These tests proved that the test protocol and experimental setup described above results in suicient accuracy and repeatability. NJ TRANSIT replaced all the uel and oils ilters prior to these tests. The measured locomotive horsepower was consistent with NJ TRANSIT expectations and the raw gaseous emissions data and brake speciic mass emissions data were consistent with those reported elsewhere [7]. Prior studies have shown that biodiesel has a tendency to act as a detergent and loosen deposits in uel systems [1]. Accordingly, prior to acquiring emissions data on the irst biodiesel blend (B2/#2 summer), locomotive 4142 was irst operated statically on the MMC load box or 8 hours on in the idle notch, periodically throttling it up to notch 8 or a ew minutes to ensure proper operation. Ater the 8 hour period, the uel ilters were removed, disassembled and inspected by NJ TRANSIT personnel and showed little or no deposits. New uel ilters were installed and the irst set o emissions tests were then conducted with the B2/#2 summer blend. These tests showed a substantial decrease in horsepower with respect to the #2 baseline tests. For example, in notch 8, the average horsepower dropped rom 29 to 26 horsepower. A 1% decrease in horsepower was unexpected and potentially problematic or uture plans to deploy B2 into the NJ TRANSIT locomotive leet. Based on previous studies (both with locomotives and heavy duty diesel engines), a to 2% decrease in horsepower had been expected since the lower energy content o the oxygenated biodiesel uel is oset by its higher density, resulting in minimal reduction in delivered power (or, in the case o on-road vehicles, volumetric uel economy in MPG). Ater trouble-shooting this problem, it was hypothesized that the uel injectors may have been clogged due to the detergent eect o the B2 blend described above. Ater pulling a random selection o injectors, it was ound that they were in act clogged and they were replaced with new units. The next tests showed an immediate increase in horsepower to levels that were comparable with those o the #2 summer uel. Unortunately, there was not suicient B2/#2 summer uel let in the tank to complete a ull test sequence during the window o opportunity and the tests had to be rescheduled or a later date. The remainder o the B2/#2 summer tests were completed in September. Preliminary results o data analysis comparing the baseline #2 and B2/#2 tests are reported below and tabulated in Table 6 o the Appendix. 7.1 GP4FH-2 Horsepower Results The data presented herein represents the inal calculated and corrected results rom the GP4FH-2 test matrix. Figures 4 and 5 are bar graphs comparing the measured horsepower at each notch setting or the summer and winter uel blends, respectively. Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 14

15 Each bar represents the average o three separate test points. As shown in Figs. 4 and 5, the horsepower levels in all o the notches were very comparable, showing only slight variations between the summer blends and slightly larger variations with the winter blends, with the lowest power rom the ULSD winter uel. However, the horsepower drop in Winter USLD and Winter ULSD B2 uel can be partially attributed to the necessity to test year round in order to complete the matrix. As such, both o these winter uel blends were tested during the summer months. Higher ambient air and uel eed temperatures and during these test thus had a negative impact on horsepower, and the drop shown in Figure 5 can not be entirely attributed to uel blend dierences alone BHP (horsepower) Notch Summer #2 Summer B2 ULSD Summer ULSD Summer B2 Figure 4. Measured horsepower or each notch setting during static load box testing or summer uel blends BHP (horsepower) Notch Winter #2 Winter B2 ULSD Winter ULSD Winter B2 Figure 5. Measured horsepower or each notch setting during static load box testing or winter uel blends. Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 15

16 7.2 GP4FH-2 Opacity Results Figures 6 and 7 are bar graphs showing the exhaust gas opacity or the same test conditions as Figs. 4 and 5. As shown in Fig. 6 the measured opacity levels in most o the notches showed a substantial decrease or the alternative summer blends in comparison with the baseline #2 summer tests. Speciically, the B2/#2 showed reductions o up to 3% with respect to the #2 and over 5% reduction with ULSD/B2 in several o the notches. In notches 2 and 3, the ULSD blends showed slight increases in opacity. Fig. 7 suggests the exhaust opacity or all winter blends are relatively close, and generally lower than the summer blends. Again, the ULSD blends actually exhibited slightly higher opacity than the baseline LSD blends, which was an unexpected result. The opacity results or all the test blends < 1% in the idle notch setting. A previous study with low sulur diesel uels on a 2-stroke, roots-blown, diesel locomotive engine showed that PM emissions are relatively unresponsive to uel type, since PM emissions rom 2-stroke roots blown engines are expected to be dominated by lubricating-oil derived components [7]. For the 2-stroke, turbocharged engines tested herein, the PM emissions are expected to have a lower soluble organic raction than that o the roots blown engine, particularly at the higher notch settings when the turbocharger becomes ully engaged thereby resulting in more complete combustion. Indeed, as shown in Fig. 6, the opacity measurements or the lower notch settings are less sensitive to uel type than the higher notch settings or the summer blends. Figure 7 shows winter blends or the GP4FH-2 locomotive exhibited increased opacity or the lower sulur blends in most notches in comparison with the higher sulur diesel. Typically, or these 2-stroke, turbocharged diesel engines, the CO and opacity max out in notches 5 and 6 and are much lower in notches 7 and 8. This is because, at lower notches an overrunning clutch spins the turbocharger rom the engine's own gear train. But when the power is increased, the turbine takes over and spins the compressor aster than the engine's drive, so the overrunning clutch releases and the turbo is ully powered by exhaust. The net result is that the CO and opacity decrease at the high notches because o more complete combustion when the turbo boost pressure is maximum. However, as shown in Figs 7 and 11, the CO and opacity or both the ULSD winter and ULSD-B2-winter appear to have a maximum in notch 7, instead o notch 5 or 6, as was observed or all o the summer blends. Figure 5 shows a drop in horsepower or the ULSD winter and ULSD Winter B2 blends as well, so it is possible that the engine power/speed was not high enough or the turbo to ully engage until Notch 8. Furthermore, Table 3 shows that the ULSD-B2-Winter tests were perormed on July 3, 28 when the ambient temperature was above 8 degrees F, resulting in decreased engine horsepower due to decreased air and uel density. One o the inherent drawbacks o the test plan was it was necessary to perorm testing over the entire calendar year. To meet NJT's specs, we had to ormulate a winter blends that met the cloud point requirements. So, tests were perormed in the summer with uels that were very high in kerosene. The test results perormed with the GP4FH-2 turbocharged locomotive with summer B2 blends exhibited clear dierences in opacity. This result is contrary to another study perormed by Fritz and coworkers [8] who reported little dierence in opacity with the addition o ULSD in the same roots blown EMD E engine employed in [7]. Again, the dierence between the tests perormed in this study and those cited in [7,8] is that the tests perormed herein were perormed on a turbocharged engine. It should Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 16

17 also be noted that direct PM mass emissions were not measured in the present study, but rather only opacity. While opacity may correlate with total PM mass emissions, this is not necessarily the case Opacity (%) Idle Notch Summer #2 Summer B2 ULSD Summer ULSD Summer B2 Figure 6. Measured opacity or each notch setting during static load box testing or summer uel blends Opacity (%) Idle Notch Winter #2 Winter B2 ULSD Winter ULSD Winter B2 Figure 7. Measured opacity or each notch setting during static load box testing or winter uel blends. Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 17

18 7.3 GP4FH-2 Brake Speciic Emissions Results Because the EPA Line-Haul Duty Cycle heavily weights the data rom the idling conditions, the calculated EPA Line-Haul average or the alternative uels would be higher than that calculated or #2 summer diesel. It should be noted, however, that the horsepower is not directly measured during normal idle because the main traction alternator is not engaged. Rather, the engine horsepower is estimated based on estimated parasitic loads (cooling ans, etc.). Moreover, the measured uel mass low rate has its highest uncertainty during idling conditions as a result o subtracting supply and return uel low rates which are very close in magnitude during idle conditions. An uncertainty analysis shows that an uncertainty o ± 4% exists on the uel mass low rate measurements during the idle conditions. That uncertainty reduces to 1.3% or Notch 8. When combined with the uncertainty o the estimated horsepower at idle, the total uncertainty in brake speciic NO x emission at idle is ± 48%, which reduces dramatically to 3.6% at Notch 8. Accordingly, the team evaluated whether the EPA Line-Haul Duty cycle was the most appropriate cycle or these tests. It is clear that the accuracy o the reported data would increase dramatically i the idle data were not included in the overall calculations. The brake speciic emission plots contain error bars to represent the percentage uncertainty in each notch, as well as the total weighted uncertainty or the averaged data. Figure 8 shows the brake speciic NO x emissions in g/bhp-hr or each notch setting or the summer uel blends. The igure shows that the individual brake speciic NO x emission values were slightly higher or the alternative summer blends at most notches, except or notch 1, which showed signiicant NO x emissions. Figure 9 shows the brake speciic NO x emissions in g/bhp-hr or each notch setting or the winter uel blends. The igure suggests that the individual brake speciic NO x emission values were comparable or the alternative winter blends at most notches, with the exception o ULSD Winter B2, which showed a slight increase in notch 1. Figure 1 shows the brake speciic CO emissions in g/bhp-hr or each notch setting or the summer uel blends. The igure shows that the CO emissions or the alternative uels were higher in notch 1 and substantially lower in the middle notches (5 and 6), showing reductions o over 5% in some cases with the ULSD/B2 blend. Figure 11 shows the brake speciic CO emissions in g/bhp-hr or each notch setting or the winter uel blends. The igure shows that the CO emissions or the alternative uels were greater in the higher notches, showing increases as large as 1% over the baseline data with ULSD uels. ULSD Winter B2 increased CO emissions substantially in notch 1 as well. As noted earlier, the results or a roots blown 2-stroke engine [7] may be dierent than those observed with a turbocharged engine tested here. As shown in Fig. 1, or the summer blends, dramatic decreases in brake speciic CO mass emissions were observed or the ULSD and B2 blends with respect to the baseline. While these results are higher than what might be expected, they are qualitatively consistent with the opacity results. Figure 12 shows the brake speciic HC emissions in g/bhp-hr or each notch setting or the summer uel blends. The emissions or the alternative uels were roughly the same in all notches with the exception o Notch 1, which showed an increase o over 1% with the ULSD/B2 blend. Figure 13 shows the brake speciic HC emissions in g/bhp-hr or each notch setting or the winter uel blends. The emissions or the alternative uels were roughly the same in all notches. Figures 14 and 15 show the brake speciic CO 2 emissions in g/bhp-hr or each notch setting or the summer and winter uel blends, respectively. The CO 2 emissions in all o the notches were comparable, with the Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 18

19 exception o the lower notches with the ULSD/B2 summer blend, which exhibited increases with respect to the baseline #2 uel NO x (g/bhp-hr) MLH- AVG Notch Summer #2 Summer B2 ULSD Summer ULSD Summer B2 Figure 8. Brake speciic NO x emissions (g/bhp-hr) or the GP4FH-2 locomotive operating the summer blend uels NO x (g/bhp-hr) MLH- AVG Notch Winter #2 Winter B2 ULSD Winter ULSD Winter B2 Figure 9. Brake speciic NO x emissions (g/bhp-hr) or the GP4FH-2 locomotive operating the winter blend uels. Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 19

20 12 1 CO (g/bhp-hr) MLH- AVG Notch Summer #2 Summer B2 ULSD Summer ULSD Summer B2 Figure 1. Brake speciic CO emissions (g/bhp-hr) or the GP4FH-2 locomotive operating the summer blend uels CO (g/bhp-hr) MLH- AVG Notch Winter #2 Winter B2 ULSD Winter ULSD Winter B2 Figure 11. Brake speciic CO emissions (g/bhp-hr) or the GP4FH-2 locomotive operating the winter blend uels. Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 2

21 THC (g/bhp-hr) MLH- AVG Notch Summer #2 ULSD Summer ULSD Summer B2 Figure 12. Brake speciic THC emissions (g/bhp-hr) or the GP4FH-2 locomotive operating the summer blend uels. (Faults with the THC sensor in the emissions equipment voided the summer B2 data.) THC (g/bhp-hr) MLH- Notch AVG Winter #2 Winter B2 ULSD Winter ULSDWinter B2 Figure 13. Brake speciic THC emissions (g/bhp-hr) or the GP4FH-2 locomotive operating the winter blend uels. Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 21

22 25 2 CO 2 (g/bhp-hr) MLH-AVG Notch Summer #2 Summer B2 ULSD Summer ULSD Summer B2 Figure 14. Brake speciic CO 2 emissions (g/bhp-hr) or the GP4FH-2 locomotive operating the summer blend uels CO 2 (g/bhp-hr) MLH- Notch AVG Winter #2 Winter B2 ULSD Winter ULSD Winter B2 Figure 15. Brake speciic CO 2 emissions (g/bhp-hr) or the GP4FH-2 locomotive operating the winter blend uels. 7.4 Estimated Average Total Mass Emissions or GP4FH-2 To estimate the average mass emissions (g/hr) or each gaseous species or the GP4FH-2 locomotive, the measured mass emissions at each notch are multiplied by a Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 22

23 weighting actor or that notch which was developed rom actual NJ TRANSIT notch data as tabulated in Table 6 in the appendix. Figures 16 through 19 show the NJT-weighted average mass emissions rates o NOx, CO, HC and CO 2, respectively. The data which is shown graphically in Figs. 16 through 19 is tabulated in the Executive Summary in Table E1. The greenhouse CO 2 emissions estimates were made assuming carbon intensity numbers o g CO 2 e/mj or petroleum diesel and g CO 2 e/mj or soy biodiesel as proposed by the Caliornia Air Resources Board [3]. As shown in Figure 16, the summer blend results predict an increase o 7.2% NO x and 14.9% NO x or B2-summer and ULSD-B2-Summer, with respect to the baseline, respectively. These increases with respect to the baseline are slightly higher than that which would be expected rom previous studies on heavy duty diesel engines [3]. However, very little work has been done on large 2-stroke, turbocharged diesel engines such as these. It should also be noted, however, that ULSD alone resulted in an increase in NO x o 13.25%. Indeed, i we compare the ULSD-B2-summer to the ULSDsummer, the presence o the biodiesel resulted in only a 1.5% increase. Such a 1.5% increase is consistent with many previous studies on B2 [4]. In terms o the winter blends, all o the blends tested in the GP4FH-2 resulted in decreases in NO x. Speciically, the ULSD-winter, B2-winter and ULSD-B2-winter resulted in NO x decreases o 1.3%, 5.7% and 9.6% respectively. In contrast to the summer blends, the presence o the ULSD appears to have lowered the NO x. Indeed, a comparison o ULSD-winter to ULSD-B2-winter suggests that the presence o the biodiesel resulted in an increase in NO x o.76%, which is also consistent to what might be expected rom prior smaller diesel engine studies with B2 [4]. The dramatic dierences in the eect o ULSD and B2 on NO x between the summer and winter blends is surprising. It should be noted, however, that all winter blends contained very high percentages o kerosene as necessary to meet NJ TRANSIT s winter cloud point requirement as F max which is the highest ambient temperature or a uel to become cloudy The presence o the kerosene might, in act, be the dominating actor. Reerring back to Figs. 6 and 7, the dierence between the summer and winter blends is also evident when examining the opacity results. For the summer blends, the ULSD and biodiesel resulted in substantial decreases in opacity, as expected. However, all o the winter blends (including the baseline #2 winter) resulted in decreased opacity with respect to the summer blends. This result could also be due to the presence o the kerosene. Figure 17 shows that the ULSD-B2-Summer blend resulted in a 43% percent decrease in CO, whereas the ULSD-B2-Winter blend resulted in increased CO emissions. Figure 18 shows that that the ULSD-B2-Summer blend resulted in a 15% increase in HC emissions, whereas all o the winter blends showed very little variation in HC emission with respect to the baseline. Figure 19 shows the measured CO 2 mass emissions in kg/hr, along with the estimated greenhouse CO 2 emissions. As shown, all o the B2 blends result in decreased greenhouse CO 2 emissions. Speciically, the results suggest that NJ TRANSIT could realize a 7% decrease in greenhouse gas emissions using B2 blends in their older locomotive leet. Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 23

24 NO x (g/hr) #2 Summer ULSD- Summer B2- Summer ULSD- B2- Summer #2 Winter ULSD Winter B2- Winter ULSD- B2- Winter Fuel Figure 16. Mass emissions rate or NO x emissions (g/hr) or the GP4FH-2 locomotive operating the summer and winter blend uels CO (g/hr) #2 Summer ULSD- Summer B2- Summer ULSD- B2- Summer #2 Winter ULSD Winter B2- Winter ULSD- B2- Winter Fuel Figure 17. Mass emissions rate or CO emissions (g/hr) or the GP4FH-2 locomotive operating the summer and winter blend uels. Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 24

25 6 5 4 THC (g/hr) #2 Summer ULSD- Summer ULSD-B2- Summer #2 Winter ULSD Winter B2-Winter ULSD-B2- Winter Fuel Figure 18. Mass emissions rate or THC emissions (g/hr) or the GP4FH-2 locomotive operating the summer and winter blend uels CO 2 (kg/hr) #2 Summer ULSD- Summer B2- Summer ULSD- B2- Summer #2 Winter ULSD Winter B2- Winter ULSD- B2- Winter Fuel Total CO2 Greenhouse CO2 Figure 19. Mass emissions rate or CO 2 emissions (kg/hr) and greenhouse CO 2 emissions or the GP4FH-2 locomotive operating the summer and winter blend uels. Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 25

26 8. Test Results: PL42-AC The second locomotive tested in this study was a PL42-AC model. Locomotive 414 was allocated or our testing purposes and released or normal revenue service when alternative uel tests were not scheduled. All tests were perormed between September 28 and January 29. The procedure and equipment used in the PL42-AC tests were very similar to the GP4FH-2 tests with just a ew minor exceptions. Because the PL42-AC model is computer controlled with electronic uel injection, it was not possible to tap into the main traction engine s alternator to read voltage and current. Our horsepower readings were obtained rom the PL42-AC s on-board computer system, recorded rom an LCD monitor readout in the cabin. The PL42-AC is also equipped with a sel load eature. This was convenient because it allowed tests to be conducted wherever the locomotive could be saely parked, as opposed to the GP4FH- 2, which required the use o the load box. One downside o the sel load eature is that it was not possible to load the locomotive at throttle notches 1 through 4. Accordingly, the test results or the PL42-AC locomotive only contain brake speciic data rom notches 5 through 8. And, because loading was only possible in notches 5 through 8, it was not possible to calculate an overall Modiied Line-Haul Cycle average. The Semtech-D analyzer was used in the same manner as it was during the GP4FH-2 tests, recording gaseous emissions concentrations in percentage and ppm. The uel low meter control units were reused, in conjunction with larger positive displacement low meters to accommodate the higher uel low and uel pressure o the larger engine. The opacity meter with abricated locomotive-speciic bracket was also reused on the PL42-AC with minor modiication. O the eight original uels planned or testing, one was dropped (Summer ULSD) due to insuicient time in the summer 29 season to complete that uel. Furthermore, due to uel gelling, testing this blend in the winter months was impractical. In addition, during the testing o Summer ULSD-B2 blend, the Semtech-D gaseous analyzer suered a ailure which was not apparent until months later during data analysis. As such, or that uel, no gaseous emissions are reported. However, horsepower and opacity results or Summer ULSD-B2 are reported. 8.1 PL42AC Horsepower Results Figure 2 shows the measure horsepower rom the tests conducted with the PL42-AC locomotive. These igures were recorded rom the readout in the cabin and were subsequently modiied using AAR correction actors or ambient temperature, altitude, uel speciic gravity and uel temperature were applied. The alternative summer blends showed a slight decrease in power, with Summer ULSD B2 decreasing the most at 1.8%. The alternative winter uels all saw an increase in power over the baseline, with the largest gain being 1.9% rom Winter ULSD B2. It should be noted, once again, that the winter blends contained high quantities o kerosene. Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 26

27 HP Notch 7 8 Baseline LSD Summer LSD B2 Summer ULSD B2 Summer LSD Winter LSD B2 Winter ULSD Winter ULSD B2 Winter Figure 2. Brake Horsepower measurements rom PL42-AC locomotive or test uels. 8.2 PL42AC Opacity Results Figure 21 shows the measured exhaust opacity or the entire set o PL42-AC tests. The results or baseline summer and Summer LSD B2 were similar in most notches, however, a substantial decrease (over 65%) in some notches was seen with Summer ULSD B2. Winter ULSD B2 opacity was slightly higher than the baseline, while Winter LSD B2 and Winter ULSD showed decreased opacity up to 8% in some cases or the latter uel. The ULSD B2 Winter showed increased opacity with respect to the ULSD winter, which was unexpected. Please note that no idle data or the ULSD B2 Winter uel is plotted due to insuicient idle opacity data. Idle data or the other six uels is included even though or several the opacity levels were nearly zero or below the detection limit o the opacity meter. Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 27

28 Figure 21. Exhaust Opacity measurements rom PL42-AC locomotive or test uels. 8.3 PL42AC Brake Speciic Emissions Results Brake speciic emissions were calculated or notches 5 through 8 or the dierent pollutants. No line-haul average was applied to this data because only having data rom our made it diicult to come up with representative weighting actors or the runs. The error bars in brake speciic emissions plots indicate the percent uncertainty in the calculation or each notch. Figure 22 shows the brake speciic NO x emissions or the test uels in the PL42-AC locomotive. Summer LSD B2 showed similar results to the baseline, some notches slightly higher, other slightly lower. The alternative winter uels showed a decrease in emissions in all cases except or notch 5, where the baseline was slightly lower. The largest decreases, up to 25%, were seen with Winter ULSD. Figure 23 shows the brake speciic CO emissions or the test uels in the PL42-AC locomotive. Summer LSD B2 showed a slight decrease in emissions except or notch 6. All the alternative winter uels decreased CO emissions with the largest decrease being 5% rom LSD B2. Figure 24 shows the THC emissions or the PL42-AC locomotive. Summer LSD B2 values were comparable to the baseline. ULSD Winter was very similar to the baseline winter uel, however, LSD B2 and ULSD B2 were both slightly higher in all the notches. The reductions in THC and CO illustrated in Figures 23 and 24 due speciically to sulur content ollow similar trends between our high sulur diesel (2ppm), LSD (5PPM) Evaluation o B2 Biodiesel Blends in NJ TRANSIT Locomotives pg. 28