Home

ABSTRACT SANDHU, GURDAS SINGH. Methods for

1. i aB 2 on E Q a 120 100 D A 80 E S 60 40 20 m B20 m B20FA 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 l Normalized Manifold Absolute Pressure mB20 m B20FA 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 Normalized Manifold Absolute Pressure GB20 m B20FA 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 l Normalized Manifold Absolute Pressure 86 150 GB20 WB20FA T amp 100 oF Z 50 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 Normalized Manifold Absolute Pressure a B20 mB20FA E C ae 5 5 z A oS 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 Normalized Manifold Absolute Pressure B20 20 Biodiesel 80 Petroleum Diesel B20FA B20 with Fuel additive Figure A 2 Emission Rates of Each Pollutant on a Per Fuel Basis for Engine Based Modes for Combination Truck 215 5715 87 HC gallon CO g gallon 0 1 0 1 0 2 0 2 0 3 0 4 0 5 0 6 0 7 Normalized Manifold Absolute Pressure 0 3 0 4 0 5 0 6 0 7 Normalized Manifold Absolute Pressure 88 0 8 GB20 BB20FA 0 9 1 mB20 WB20FA 0 8 0 9 l Fuel Use g sec CO g sec NO mg g sec BB20 WB20FA 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 l Normalized Manifold Absolute Pressure 80 BB20 BB20FA 60 40 20 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 Normalized Manifold Absolute Pressure 250 GB20 BB20FA 200 150 100 50 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 Normalized Man
2. System Environmental Science amp Technology Volume 41 Number 18 p 6554 6561 Dearth M A Butler J W Colvin A Gierczak C Kaberline S Korniski T 2005 SemtechD The Chassis Roll Evaluation of a Commercial Portable Emission Measurement System PEMS SAE Technical Paper No 2005 01 0673 Society of Automotive Engineers Warrendale PA 78 Durbin T D Johnson K Cocker III D R Miller J W 2007 Evaluation and Comparison of Portable Emissions Measurement Systems and Federal Reference Methods for Emissions from a Back Up Generator and a Diesel Truck Operated on a Chassis Dynamometer Environmental Science amp Technology Volume 41 Number 17 p 6199 6204 Frey H C Unal A Chen J Li S Xuan C 2002a Methodology for Developing Modal Emission Rates for EPA s Multi scale Motor Vehicle amp Equipment Emission System EPA420 R 02 02 Prepared by North Carolina State University for Office of Transportation and Air Quality U S Environmental Protection Agency Ann Arbor MI 2002 Frey H C Unal A Chen J 2002b Recommended Strategy for On Board Emission Data Analysis and Collection for the New Generation Model Prepared for U S Environmental Protection Agency Office of Transportation and Air Quality Ann Arbor MI by the Computational Laboratory for Energy Air and Risk Department of Civil Engineering North Carolina State University Raleigh NC 2002 Frey H C Unal A Rou
3. Install Montana in vehicle seat insert bottom padding if required Connect power cable Montana to DC splitter Connect MAP cable to Montana Connect Sensor Array cable to Montana Connect Gas and PM sampling lines to Montana Connect exhaust lines 3 and zeroing line 1 to Montana Connect Montana GPS to data port and power port Secure sampling lines so that they do not interfere with gear shift Check MAP RPM Temperature values Check Gas concentrations and PM level Check Montana GPS fix Take Odometer reading Figure 23 Example of a Check List for Installation of the Montana System on a Vehicle 45 3 4 Quality Assurance and Quality Control For quality assurance purposes the combined data set for a vehicle run is screened to check for errors or possible problems If errors are identified they are either corrected or the data set is not used for data analysis First the types of errors typically encountered are described followed by a discussion of methods for making corrections The predominant types of errors or problems include Engine Data Errors On occasion communication between the vehicle s onboard computer and the engine scanner may be lost leading to loss of data Sometimes the loss of connection is because of a physical loss of electrical contact while in other cases it appears to be a malfunction of the vehicle s on board diagnostic system This rarely happens However when it happens this error can b
4. 2877 2883 Konstantas G Stamatelos A 2004 Quality assurance of exhaust emissions test data Proceedings of the Institution of Mechanical Engineers Part D Journal of Automobile Engineering Volume 218 Number 8 2004 p 901 914 Kubelt J and Bonnel P 2007 Portable Emission Measurements PEMS Data evaluation and post processing Manual for the Data Evaluation Software EMROAD Version 3 80 European Commission Joint Research Centre Institute for Environment and Sustainability ISSN 1018 5593 Ropkins K Quinn R Beebe J Li H Daham B Tate J Bell M Andrews G 2007 Real world comparison of probe vehicle emissions and fuel consumption using diesel and 5 biodiesel B5 blend Science of The Total Environment Volume 376 Issues 1 3 p 267 284 Vojtisek Lom M and Cobb J T 1997 Vehicle Mass Emissions Measurements Using a Portable 5 gas Exhaust Analyzer and Engine Computer Data Proceedings of the EPA A amp WMA Emissions Inventory meeting Research Triangle Park NC 1997 Vojtisek Lom M and Cobb J T 1998 On road Light duty Vehicle Emissions Measurements Using a Novel Inexpensive On board Portable System Proceedings of the 8th CRC On road Vehicle Emissions Workshop San Diego CA 1998 Vojtisek Lom M and Allsop J E 2001 Development of Heavy Duty Diesel Portable On Board Mass Exhaust Emissions Monitoring System with NO CO 2 and Qualitative PM Capabilities SAE Technical
5. 3 and 4 increasing by 32 6 8 9 8 6 and 13 6 respectively At the same time fuel use rates decreased by more than 6 for VSP modes 10 11 and 12 and more than 16 for VSP modes 13 and 14 Results for VSP based fuel use when the OBD data is shifted are shown in Figure 11 MAP based fuel use results show lower sensitivity to OBD data shift When the OBD data is shifted by 5 seconds fuel use for normalized MAP mode 0 1 decreased by 4 7 and increased by 7 5 for the same mode when the data is shifted by 10 seconds When GPS data is shifted by 10 seconds fuel use increased by 26 3 10 8 and 9 4 for VSP modes 1 2 and 4 25 ma Ur EOBD no offset OBD 5 sec offset OBD 10 sec offset N Fuel Use g sec a a Ur l 2 3 4 5 6 7 8 9 10 11 12 13 14 Vehicle Specific Power Mode Figure 11 Effect of shifting OBD Data on VSP based Fuel Use Rate Results for VSP and MAP based NO emissions rate When OBD data is shifted by 5 seconds VSP based NO emissions rate increased by 15 7 10 4 and 77 3 for modes 1 2 and 3 respectively and decreased by 18 3 for mode 13 A 10 second shift to OBD data resulted in NO emissions rate increasing by 48 9 and 126 4 for VSP modes and 3 At the same time NO emissions rate decreased by 22 to 48 for modes 9 to 14 Results for NO emissions rate when OBD data 1s shifted are shown in Figure 12 A shift of 5 seconds to OBD data decreased the NO emissions rate by 13 6 for MAP m
6. Where r Equation 12 r Pearson s Coefficient of Correlation X Master data array with n elements X Mean of X Y Slave data array with n elements Y Mean of Y 4 3 Intake Air Temperature The algorithm for marking IAT errors was updated to more than a mere comparison of a pair of consecutive values since that simple algorithm results in false positives For example consider three consecutive IAT values in C of 60 71 61 where the 71 C is an error value If the algorithm is set to delete the former of the pair it would delete 60 C appearing before the 71 C which is incorrect If the algorithm is set to delete the latter of a pair and the data were 60 71 71 61 it would correctly delete the first 71 C value but keep the 16 second 71 C value and instead delete the valid 61 C value Errors in IAT values are uncommon but when they do occur each instance is typically 1 to 4 consecutive seconds long The new algorithm when it encounters an error value retains the last good value and makes an extended comparison till the next good value is encountered or 5 seconds have elapsed whichever is earlier AT is a slow moving variable and the sensitivity of fuel use and exhaust emission mass rates to IAT is less than 2 to a change of 10 in IAT value Thus to prevent loss of data due to IAT error it is safe to impute missing or error IAT values up to 5 consecutive seconds which is implemented
7. and road grade from GPS Frey et al 2002a In order to handle the large amount of data and parameters and multiple timestamps generated by PEMS OBD and GPS a computer based quality assurance method is needed 2 4 Need For Rapid Onsite QA Rapid on site quality assurance is important to capture and correct systemic failures in sensors or measurement setup Examples of such failures could be a sensor that is biased high or low air leak in the sampling line or loss of signal for intermittent periods Rapid on site quality assurance will help to quickly identify problems so that it is possible they can be fixed In turn accumulation of errors over multiple tests in sequence can be avoided by quickly diagnosing and correcting problems as they occur Other examples of test setups that require rapid onsite QA are e Time sensitive series of tests that are needed for comparisons to make sure baseline data are valid before proceeding with comparative tests Such as tests to compare fuel alternatives Petroleum Diesel vs B20 Diesel or effect of Fuel additives e Optimization tests such as engine performance versus emissions optimization tests In these tests one or more input parameters 1s varied e g for locomotive engine tested on a dynamometer fuel airbox pressure fuel injector type and or timing presence of heat shields and configuration of cooler water and the resulting engine power output and emissions are measured Give
8. vol CO produces about 5 5 change of same polarity in fuel use and CO2 mass emission rate about 4 5 change of reverse polarity in NOx HC and CO mass emission rates a 4 0 change of reverse polarity in exhaust air flow rate A 10 change in vol O results in about 5 5 change of reverse polarity in all output parameters except intake air which is unaffected A 10 change in IAT produces about 1 6 change of reverse polarity in all output parameters A 10 change in exhaust volume fractions of NOx HC and CO has no effect on any output parameters except an almost equivalent change in their own mass emission flow rate The sensitivity of HDD fuel use to change in input parameters in shown in Figure and graphs for effect on other parameters is included in the appendix For LDG a change in RPM MAP and NO HC and CO has similar effect on output parameters as with HDD A 10 change in vol CO gt has negligible effect on fuel use rate and CO mass flow rate but a almost 10 change of reverse polarity in NOx HC and CO mass emission rates and exhaust air flow rate A 10 change on vol Oz has negligible effect on all parameters The sensitivity of LDG fuel use to change in input parameters in shown in Figure 2 and graphs for effect on other parameters is included in the appendix In summary in general the output parameters are most sensitive to changes in RPM and MAP followed by CO and Oy and finally to a lesser degree IAT Thus fr
9. 0 7 Normalized Manifold Absolute Pressure 0 3 0 4 0 5 0 6 0 7 Normalized Manifold Absolute Pressure 96 GB20 BB20FA 0 8 0 9 l mB20 WB20FA 0 8 0 9 l Evaluation of Non Detected Measurements of Modal Average Gas Concentrations Table A 1 Average Concentrations for Each Pollutant Based on 215 5715 Combination Truck Fueled for B20 172 187 06 60 18 OS 535 13 5 188 203 07 660 22 0 005 547 132 204 219 08 669 21 0005 587 127 220 235 09 644 17 0006 642 19 Detection Limits for HC and CO are approximately 13 ppm and 0 012 vol Table A 2 Average Concentrations for Each Pollutant Based on 215 5715 Combination Truck Fueled for B20 with Fuel additive 172 187 06 63 18 002 595 128 204 219 08 693 20 0014 648 121 220 235 09 738 21 0014 72 11 Detection Limits for HC and CO are approximately 13 ppm and 0 012 vol 97 Table A 3 Average Concentrations for Each Pollutant Based on 215 6415 Combination Truck Fueled for B20 Cass 03 122 o 0009 344 159 156 176 04 137 SSS 405 S A 199 219 O6 156 7 0 008 5 64 13 4 242 262 0 8 183 6 0 008 642 124 263 284 0 9 193 7 0010S 667 122 285 306 1 202 6 000 734 114 Detection Limits for HC and CO are approximately 13 ppm and 0 012 vol Table A 4 Average Concentrations for Each Pollutant Based on 215 6415 Combination Truck Fueled for
10. A 4 Average Concentrations for Each Pollutant Based on 215 6415 Combination Truck Fueled tor B20 with Fueladditive i cies ahs cai bedicsacy ashes cents chacuateentien a E I 98 Table A 5 Average Concentrations for Each Pollutant Based on 215 6667 Combination Truck Pa NS UO B Oien a erent esr abit es ete tte etree eaten eects werent 99 Table A 6 Average Concentrations for Each Pollutant Based on 215 6667 Combination Truck Fueled for B20 with Fuel additive cccccccccccccsseeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeees 99 Table A 7 Summary of Fuel USC iccoitessc on taculacareswacadebaasatondeacedduaresontindubanatoccdelagae 100 Table A 8 Rate of Loss of Data Because of Data Quality Errors c cc ccessssesesseseeeeeeeeeeeees 101 vill LIST OF FIGURES Figure 1 Sensitivity of g s Fuel Use to Input Parameters for Heavy Duty Diesel Vehicle 19 Figure 2 Sensitivity of g s Fuel Use to Input Parameters for Light Duty Gasoline Vehicle 19 Figure 3 NO from PEMS and RPM from OBD before synchronization cccceeeeeeeeeeeees 20 Figure 4 NO from PEMS and RPM from OBD after synchronization cccceeeeeeseeeeeees 21 Figure 5 Correlation between NO PEMS and RPM OBD before synchronization 21 Figure 6 Correlation between NO PEMS and RPM OBD after synchronization 06 22 Figure 7 Vehicle Speed from OBD and Vehicle Speed from GPS before synchron
11. For large 2 stroke diesel locomotive prime mover engine the range could be 300 rpm to 1000 rpm For PHEVs the range typical range 1s similar to gasoline cars with the exception that when the vehicle is idling the engine can shut off and RPM can fall to zero An error in RPM reading is defined as a value below the RPM lower limit above the RPM upper limit or missing The error second of data is deleted 1f RPM is being used to calculate intake molecular air flow rate as per speed density method which in turn 1s used to calculate exhaust flow rate and subsequently pollutant mass emission rate IAT Error The intake air temperature is a slow changing variable and based on previous field data a change in IAT greater than 1 C between two consecutive seconds t and t 1 is considered to be an error Thus the data at time t 1 are marked as bad Additionally missing IAT values are marked as error The error timestamps are deleted if IAT is being used to calculate intake molecular flow rate as per speed density method and in turn exhaust flow rate and pollutant mass emission rate MAP Error For cases where MAP value is missing for up to 3 consecutive seconds an absolute relative difference ARD is calculated using the valid MAP values occurring immediately before and after the missing MAP values If the ARD is within 5 the missing MAP value is calculated using the two MAP values immediately preceding and following the missing value If MAP val
12. PEMS main unit If the CO gas concentration is very low then there could be a leakage in the sampling line and therefore inspection and repositioning of the sampling line may be indicated 3 2 Preparation for Field Data Collection Preparations for field data collection include three major components 1 verification of the status of the PEMS and that all necessary parts and consumables are available 2 laboratory calibration of the PEMS 3 completion of a field study design and 4 coordination with the vehicle owner operator regarding scheduling of the test and access to the vehicle As part of preparation NCSU ensured that the PEMS had appropriate electrochemical sensors for NO and Oy and that all consumables replacement parts were replaced such as filters in the sampling line NCSU conducted a calibration of the PEMS using a standard calibration gas Field study design includes specifying which vehicles are to be tested when they are to be tested what fuel will be used what type of duty cycle will be performed and who will operate the vehicle As part of this project NCDOT allowed NCSU to access selected vehicles for testing Three vehicles were tested Each vehicle was tested during one day when operated on B20 biodiesel without the Fuel additive and a repeat test was made on another day when operated on B20 biodiesel with the Fuel additive B20FA NCDOT obtained the fuel and fueled the vehicles 42 NCDOT provided a
13. Pollutant on a Per Time Basis for Engine Based Modes for Combination Truck 215 6415 0 0 0 0 ccccccceceseeeeeseeeseessesseseeseesesseseeeeeeas 89 Figure A 4 Emission Rates of Each Pollutant on a Per Fuel Basis for Engine Based Modes for ZDVS O4 gt COMI aOR TTC i ai iatessscesiauispae tetas aetcine aceon E eae 9 Figure A 5 Average Fuel Use and Emission Rates of Each Pollutant on a Per Time Basis for Engine Based Modes for Combination Truck 215 6667 ccccccceeeessseeeeeeeessseesssseeseseseeseeeeeas 93 Figure A 6 Emission Rates of Each Pollutant on a Per Fuel Basis for Engine Based Modes for Combination Truck 215 6007 resons ienn A SEEE A E EEE RENS 95 Figure A 7 Scatter plot of Actual Fuel Use vs Measured Fuel Use eeeeeeeeeeeeeeeeerrrrrrerrrrren 100 Figure A 8 Cumulative Frequency of Normalized Manifold Absolute Pressure for Truck 215 E H E A E N E A A E EEE S E E E A E AE E ES 103 Figure A 10 Cumulative Frequency of Normalized Manifold Absolute Pressure for Truck 215 GOOT arane E E a daieeah eet emenal Agaeueneueaed 104 Figure A 11 Cumulative Frequency of Normalized Vehicle Speed for Truck 215 5715 105 Figure A 12 Cumulative Frequency of Normalized Vehicle Speed for Truck 215 6415 106 Figure A 13 Cumulative Frequency of Normalized Vehicle Speed for Truck 215 6667 107 Chapter 1 Methods for Quality Assurance of Portable Emissions Measurement System Data 1 Introduction Use of
14. and corrected Dilution of the exhaust sample with ambient air does not affect the fuel based emission rates for NO CO or HC Hence a comparison is still made between the fuel with additive versus baseline fuel without additive for fuel based emission factors However the dilution of exhaust with ambient air will lower the estimates of mass per time fuel flow rate and emission rates and also will lead to underestimation of the mass per distance fuel flow and emission rates Although these results are reported they are footnoted to indicate that they are for informational purposes only and should not be used for comparison 4 7 Results of Fuel use and Emission Rates for Each Tested Truck Table 5 is a summary of the total fuel use and emissions for each of the three tested trucks based on the Overall real world duty cycle given in Figure 10 For example Truck No 215 6415 consumed 194 kg of fuel and emitted 598 kg of CO2 1 82 kg of NO reported as equivalent mass of NO2 30 6 g of particulate matter 152 g of hydrocarbons and 582 g of CO when operating on baseline B20 without Fuel additive 60 The HC and CO emission rates of diesel engines tend to be very low Many of the measured concentrations of these two pollutants were below the detection limits of the gas analyzers as documented in Tables A 1 through A 6 in the Appendix Thus comparisons of these emission rates for a given truck for with and without additive are typica
15. are not shown for this truck 4 8 Results of Fuel use and Emission Rates for Each Tested Truck Detailed results are given for time based fuel based and distance based for each of the three tested vehicles as follows Vehicle 215 5715 Table 6 Time based average fuel use and emission factors based on the real world typical duty cycles of Figure 10 The estimated mass of fuel use and emissions for Truck 5715 are biased low because of perforation of the exhaust pipe by rust that was detected partway through the test with Fuel additive The numbers reported for this truck are for informational purposes only Because the amount of dilution likely differed when comparing the baseline test and the test with additive ratios of the mass per time rates for with additive to without additive are not shown for this truck Table 7 Fuel based emission factors based on the real world typical duty cycles of Figure 10 The fuel based emission factors for NO CO and HC are not affected by the diluation of exhaust with ambient air because of perforation of the exhaust pipe with rust Hence ratios of these emission rates with additive to those without additive are shown However the PM emission rate is calculated in a different manner from those of the gases and these results are affected by dilution of the exhaust with ambient air Therefore ratios are not shown for PM Table 8 Travel distance based average fuel use and emission facto
16. driver operator for each vehicle NCSU instrumented each vehicle and used the instrument to observe the vehicle during normal activity 3 3 Field Data Collection Procedure NCSU used the OEM 2100 Montana system PEMS for data collection This PEMS and the key aspects of instrumenting a vehicle are described in the text below Figure 6 provides an example of a check list that is used during pre installation and on test day Field data collection includes the following main steps 1 pre installation 2 final installation 3 data collection and 4 decommissioning Pre installation was performed the morning or afternoon before a scheduled test This step involves installing on the vehicle the exhaust gas sampling lines power cable and engine data sensor array Exhaust gas sampling lines have a probe that is inserted into the tailpipe The probe is secured to the tailpipe using a hose clamp The sampling line is secured to various points on the chassis of the vehicle using plastic ties The sampling line is routed through the passenger side window of the truck cab so that it can be connected to the Montana main unit Likewise a power cable is routed from the cigarette lighter port An engine sensor array was used to measure manifold absolute pressure MAP engine revolutions per minute RPM and intake air temperature IAT MAP RPM and IAT are used in combination with the measured exhaust gas composition to estimate the fuel and air
17. factors for Vehicle 215 6415 are given in Table 10 There is less variability in the fuel based emission factors when comparing different components of the duty cycle versus the results for the time based emission factors For example whereas the time based NO emission factor varied by a factor of about 50 percent from the lowest to highest value among the three duty cycles the fuel based emission factor varies by less than 5 percent among these cycles Thus fuel based emission factors tend to be less sensitive to engine load than time based emission factors The comparison for additive versus without 1s less sensitive for NO HC and CO for the fuel based emission rates than for the time based emission rates in that there is less variability in these ratios in the former case The travel distance based emission factors for Vehicle 215 6415 are given in Table 11 The mileage based fuel use and emission rates are higher for the arterial cycle than for the highway cycle whereas the converse is the case for the time based fuel use and emission rates However the ratios of fuel use and emission rates with to without additive are the 73 same regardless of whether they are estimated from time or mileage based emission factors since the estimates are based on the same set of duty cycles that are given in Figure 10 Similar results are given for the other two vehicles as described previously For example for Vehicle 215 6667 as shown in Tab
18. flow through the engine The engine sensor array includes an MAP sensor that is connected to an existing port on the intake air manifold of the engine The RPM sensor is based on an optical device that detects the reflection of light from reflective tape that is placed on a pulley wheel that rotates at the same RPM as the engine IAT is measured with a thermocouple The amount of time for pre installation was approximately two hours per vehicle Final installation was performed in the morning prior to field data collection The Montana system was secured in the cab of the vehicle and was connected to the exhaust sample lines engine data cables and power cable In addition a GPS receiver was deployed As part of final installation the Montana system main unit was warmed up for about 45 minutes The 43 research assistant entered data into the Montana system regarding vehicle characteristics and fuel type Data collection involved continuously recording on a second by second basis exhaust gas concentration engine and GPS data The research assistant followed the test vehicle in a pick up truck and periodically checked on the status of the PEMS during a break in work activity in order to determine quickly if any problems arose during data collection that could be corrected For example sometimes there can be a loss of signal that can be corrected by checking connections in a cable Sometimes the gas analyzers freeze they fail to c
19. for RGstep values of 0 1 mile 0 05 mile and 0 25 mile Road Grade CDF 1 0 0 9 o T RG_0 1mi RG_0 05mi RG_0 25mi o 9S ion N f l Cumulative Frequency wn 17 0 14 0 12 0 10 0 8 0 6 0 4 0 2 0 0 0 2 0 4 0 6 0 8 0 10 0 11 0 Road Grade Figure 15 CDF for roadgrades calculated using RGgtep values of 0 1 0 05 and 0 25 mile 29 Effect of RGgtep on VSP Modal Results The effect of choice of RGstep value on VSP based modal model fuel use and pollutant emissions rate is evaluated Results with RGstep 0 1 mile is taken as basis for comparison of results with RGstep values of 0 05 mile and 0 25 mile RGgtep did not significantly impact VSP based fuel use CO and CO emission rates For both fuel use and CO emissions rate the highest impacts were 6 3 increase for VSP mode 1 when RGstep 0 05 mile and a 4 increase for VSP mode 9 when RGstep 0 25 mile For CO emissions rate were generally affect by less 5 except for a 24 jump for VSP mode 1 when RGstep 0 25 mile NOx emissions rate increased by 13 8 and 16 4 for VSP mode 1 and 3 when RGstep 0 05 and decreased by 11 for the same road grade criteria With RGgtep 0 25 mile NO emissions rate saw a 2 to 10 increase for VSP modes 1 to 10 and dropped by 16 for VSP mode 14 HC emissions rate increased by 11 for VSP mode 1 and 12 with RGgtep 0 05 mile An increase of 33 18 and 16 in HC emissions rate was observed for VSP modes 1 4 and 12 when RGs
20. in the updated algorithm 4 4 Road Grade Estimation When using a VSP based modal model for fuel use and pollutant mass emission rates road grade will influence the modal results For real world in use driving tests road grade is calculated from coordinate information and elevation data recorded by GPS unit Second by second road grade calculated as the change in elevation per second over the distance travelled in that second 1s noisy and leads to spurious road grade numbers characterized by 1 rapidly varying road grade even while driving over roads known to have little grade variation and 2 road grades frequently lying beyond the typical range of 10 To prevent VSP results getting biased due to noisy and incorrect road grade the road grade 1s calculated as the slope of the least square fit line for a set of consecutive seconds during which the vehicle has travelled a specified distance which is called here as RGgtep Values of 0 1 mile 0 05 mile and 0 25 mile have been used for RGstep and from preliminary review of road grade data calculated for a series of real world driving tests 0 1 mile 1s found to produce best approximation though this is part of an ongoing study and final results may alter this selection 5 Results 5 1 Sensitivity Analysis For HDD a 10 change in RPM and MAP results in about 10 change of same polarity in fuel use mass emission rates intake air rate and dry exhaust air rate A 10 change in 17
21. of driving plus cargo loading unloading at warehouses visited during the test time for combination trucks Thus the total data collected is between 18000 to 30000 seconds In a study of excavators data collection duration for the tests was typically 24 000 seconds and one test had 54 000 seconds Abolhasani 2008 In another study of cement mixer trucks measurement durations of 28 000 seconds are reported Frey and Kim 2009 2 2 Number Of Parameters In this section the parameters recorded by commercially available PEMS systems such as Axion and Montana CATI Buffalo NY Semtech D from Sensors Inc Saline MI and Horiba OBS 2000 Horiba Inc are described All subsequent sections use CATI systems as examples The Semtech D from Sensors Inc Saline MI PEMS for diesel vehicles measures CO CO2 using NDIR spectroscopy NO and NO using dual channel Non dispersive Ultra Violet NDUV resonant absorption spectroscopy HCs using a heated Flame Ionization Detector FID O2 using an electrochemical sensor and vehicle engine parameters using a OBD and vehicle location data using a GPS Dearth et al 2005 The Semtech G PEMS for gasoline vehicles is similar to Semtech D except that it has a single channel NDUV that measures only NO Gierczak et al 2006 Hart et al 2002 The latest generation Semtech DS uses the same type of sensors as the Semtech D Johnson et al 2009 The Horiba OBS 2000 measures CO and CO concentrati
22. of the 12V DC vehicle electrical system using the cigarette lighter outlet The power consumption is 5 8 Amps at 13 8 V DC The components of the sensor array including the MAP sensor engine RPM sensor and IAT sensor are briefly described 3 1 1 Manifold Air Boost Pressure Sensor In order to measure MAP a pressure sensor is installed on the engine For most heavy duty diesel engines there is a port on the engine after the turbocharger For example Figure 1 depicts the location of an existing port on the intake air manifold of a Cummins IS X 500 engine In a regular engine performance check this port is used for performance testing of the turbocharger An existing bolt is removed and a barb fitting is screwed into the port Plastic tubing is used to connect the MAP sensor to the barb fitting The MAP sensor is attached to a convenient location in the engine away from a hot surface using plastic ties The MAP sensor provides manifold air pressure data for the computer of the main unit through a cable that connects the sensor to the MAP port located in the back of the main unit 36 Figure 18 Placement of the Manifold Absolute Pressure Sensor on a Cummins ISX 500 Engine Truck 215 6415 3 1 2 Engine Speed Sensor The engine speed sensor is an optical sensor used in combination with reflective tape to measure the time interval of revolutions of a pulley that rotates at the same speed as the eng
23. on board portable emission measurements system PEMS for in use real world vehicle emissions measurement has gained importance in recent years The EPA has conducted several studies Hart et al 2002 on systems and methods for on board analysis and has used results from various studies in the development of its latest vehicle emission inventory model Mobile Vehicle Emission Modeling System MOVES Hart et al 2002 Baldauf et al 2001 On board systems allow for developing emission inventory for real world driving cycles from emission data collected on a second to second basis Some of the commercially available PEMS systems with second by second emission measurement capabilities are Clean Air Technologies Inc CATTI Sensors Inc and Horiba Inc In this paper the key research questions addressed are 1 What are the sources of errors if any in second by second real world on road vehicle emission and energy use data 2 What quality assurance procedure can be used to minimize the effect of these errors on results and 3 What is the effect of quality assurance procedure on modal fuel use and emission rates estimated from such data 2 Motivation 2 1 Data Sample Sizes In use real world should typically be conducted for four or more hours per vehicle for data sufficiency giveref Typical measurement conducted by various research groups involve five or more hours of driving for Light Duty Gasoline Vehicles LDGVs and up to eight hours
24. 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 Normalized Speed The normalization of speed is between 0 and 75 mph Sample Size Min Speed Max Speed Avg Speed SD of Speed B20FA 26841 _ oO a 43 13 29 27 Figure A 12 Cumulative Frequency of Normalized Vehicle Speed for Truck 215 6415 106 6667_B20 Cumulative Frequency 6667_B20 FA 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 l Normalized Speed The normalization of speed is between 0 and 73 52 mph a Sample Size Min Speed Max Speed Avg Speed SD of Speed B20FA 29254 2 73 52 38 93 30 78 Figure A 13 Cumulative Frequency of Normalized Vehicle Speed for Truck 215 6667 107
25. 20 Biodiesel 80 Petroleum Diesel was conducted without the Fuel additive for three tractor trailer trucks owned and operated by the North Carolina Department of Transportation NCDOT A second test was conducted to measure the fuel economy and emissions with the additive North Carolina State University NCSU has been a pioneer in the development and application of procedures for real world data collection of in use vehicles using Portable Emission Measurement Systems PEMS Beginning in 1999 NCSU has conducted field studies of the activity fuel use and emissions of light duty vehicles Frey et al 2003 Beginning in 2004 NCSU conducted field studies on comparison of B20 versus petroleum diesel for heavy duty diesel vehicles including dump trucks Frey and Kim 2006 Since 2005 NCSU has been conducting field studies on nonroad vehicles including bulldozers backhoes front end loaders motor graders excavators off road dump trucks and skid steer loaders Frey et al 2008a Frey et al 2008b NCSU has provided technical assistance on several other projects including assessment of activity fuel use and emissions of vehicles on dirt versus paved roads assessment of light duty diesel vehicle emissions in England and assessment of the effect of a Fuel additive on fuel use and emissions in progress 33 2 0 Background Commonly used methods for measuring vehicle energy use and emissions include engine dynamometers chassis
26. 2006 01 1081 Society of Automotive Engineers Warrendale PA Hart C Koupal J Giannelli R 2002 EPA s Onboard Analysis Shootout Overview and Results Assessment and Standards Division Office of Transportation and Air Quality U S Environmental Protection Agency EPA420 R 02 026 October 2002 Hawley J G Bannister C D Brace C J Cox A Ketcher D Stark R 2004 Vehicle modal emissions measurement techniques and issues Proceedings of the Institution of Mechanical Engineers Part D Journal of Automobile Engineering Volume 218 Number 8 2004 p 859 873 Hawley J Brace C J Cox A Ketcher D Stark R 2003 Influence of time alignment on the calculation of mass emissions on a chassis roll dynamometer SAE Technical Paper Series 2003 01 0395 2003 SAE World Congress Detroit Michigan March 3 6 2003 Horiba 2010 OBS 2000 Product Page http www horiba com us en automotive test systems products emission measurement systems on board systems details obs 2000 877 Accessed December 19 2010 Jimenez Palacios J L 1999 Ph D Thesis Massachusetts Institute of Technology Boston MA 1999 80 Johnson K C Durbin T D Cocker D R Miller W J Bishnu D K Maldonado H Moynahan N Ensfield C Laroo C A 2009 On road comparison of a portable emission measurement system with a mobile reference laboratory for a heavy duty diesel vehicle Atmospheric Environment 43 2009
27. 5715_B20 FA Cumulative Frequency 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 Normalized MAP The normalization of MAP is between 96 and 262 kPa Sample Size Min MAP Max MAP Avg MAP SD of MAP 22875 p98 262 154 Figure A 8 Cumulative Frequency of Normalized Manifold Absolute Pressure for Truck 215 5715 102 6415 B20 6415 B20 FA Cumulative Frequency 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 Normalized MAP The normalization of MAP is between 96 and 306 kPa Sample Size Min MAP Max MAP Avg MAP SD of MAP Figure A 9 Cumulative Frequency of Normalized Manifold Absolute Pressure for Truck 215 6415 103 6667_ B20 6667_B20 FA Cumulative Frequency 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 l Normalized MAP The normalization of MAP is between 93 and 306 kPa Sample Size Min MAP Max MAP Avg MAP SD of MAP Figure A 10 Cumulative Frequency of Normalized Manifold Absolute Pressure for Truck 215 6667 104 5715 B20 Cumulative Frequency 5715_B20 FA 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 l Normalized Speed The normalization of speed is between 0 and 72 72 mph i Sample Size Min Speed Max Speed Avg Speed SD of Speed 22875 0 70 87 41 75 27 16 Figure A 11 Cumulative Frequency of Normalized Vehicle Speed for Truck 215 5715 105 6415_B20 Cumulative Frequency 6415_B20 FA
28. 6 cylinder 500 hp engine with diesel particulate filter DPF All three trucks pulled 48 foot long trailers The unloaded weight of each 50 truck was approximately 34 000 to 37 000 Ibs The loaded weight was approximately 37 000 to 43 000 Ibs The vehicles carried miscellaneous cargos of parts and materials for delivery to NCDOT Division field sites at various locations in the state Figure 24 NCDOT Combination Trucks Examples of the loading unloading activities and the fueling of B20 for instrumented trucks 51 It was discovered during the second run with Fuel additive for Truck 5715 that the exhaust pipe had ruptured as a result of rust as shown in Figure 8 Rupture Figure 25 Ruptured Tailpipe of Truck 5715 from the B20FA run on July 8 2009 The exhaust pipe may have been perforated by rust prior to the rupture This likely lead to the introduction of ambient air into the tailpipe upstream of the exhaust sample probe Therefore during the baseline test and part of the Fuel additive test there was likely to have been excess ambient air introduced to the sample line Dilution of the exhaust sample with ambient air does not affect the estimate of gram per gallon emission rates for NO CO or HC These estimates depend only on the relative ratios of NO CO HC and CO in the exhaust The relative ratios among these components do not change because of dilution Hence comparisons can still be made on a g gallon ba
29. ABSTRACT SANDHU GURDAS SINGH Methods for Quality Assurance of Portable Emissions Measurement System Data and Methods for Field Comparison of Alternative Fuels Under the direction of Dr Christopher Frey The use of Portable Emission Measurement Systems PEMS has grown in popularity with many PEMS of various configurations in use with organizations such as universities governments consulting firms and others However in practice there is not a standardized methodology for processing of 1Hz data obtained from a PEMS which can lead to potential errors or inconsistencies in how data are used This work discusses specific quality assurance methods for identifying and where possible correcting data quality problems and procedures for formatting synchronizing and analyzing data A sensitivity analysis is conducted to show that fuel use and emissions rate numbers are highly sensitive to engine RPM and Manifold Absolute Pressure MAP for both diesel and gasoline vehicles An algorithm for finding and correcting errors in reported Intake Air Temperature values is presented Vehicle Specific Power VSP based modal model results are shown to be affected by road grade estimation and an algorithm is provided for estimating road grade using the slope of least square fit line for elevation data vs distance travelled Synchronization of data streams from multiple independent instruments is shown to significantly affect VSP based fuel use and NOx emissi
30. B20 with Fuel additive 91 112 01 159 os 17 O 185 134 155 03 158 G 352 152 156 176 04 160 8 0009 435 146 177 198 05 171 G 537 134A 199 219 06 181 ONT 5 94 127 242 262 08 21 7 000 675 17 263 284 09 2083 6 0009 706 13 285 306 1 220 S 6 771 10 5 Detection Limits for HC and CO are approximately 13 ppm and 0 012 vol 98 Table A 5 Average Concentrations for Each Pollutant Based on 215 6667 Combination Truck Fueled for B20 199 219 06 103 11 0 000 638 134 242 262 08 117 2 0 000 702 127 263 284 09 135 O37 123 Detection Limits for HC and CO are approximately 13 ppm and 0 012 vol Table A 6 Average Concentrations for Each Pollutant Based on 215 6667 Combination Truck Fueled for B20 with Fuel additive 113 133 02 90 19 GO S 199 219 06 12 20 ono 656 127 242 262 08 u7 f 19 0003 714 RI 263 284 09 139 18 0003 748 16 Detection Limits for HC and CO are approximately 13 ppm and 0 012 vol 99 Table A 7 Summary of Fuel Use on ast Menor a Asa tite ast Wear Dora Estimated fuel use is biased low because of exhaust pipe rupture Displayed in Figure A 7 but excluded from the regression analysis trendline NCDOT did tank up at start of test so actual fuel use is biased high Not displayed in Figure A 7 Based on fuel station refilling after test The truck is tanked up befo
31. ENDIX zadenn e TON ec eds 83 Appendix A Detailed Emission Factor Results for Individual Tested Vehicles 84 vi LIST OF TABLES Table 1 Sensitivity Analysis Base Case and Typical Range for Selected Input Parameters 14 Table 2 Frequently used dataset parameter pairs to synchronize datasets csseseeeeeeeeees 15 Table 3 Data Conec non Ss Chea le yisatast stale tects net Ged wets oh a tind ete 50 Table 4 Data Collection PIC Lofnnrsseroeinirn rin EA O EE 53 Tabie 5 S ummary of Data Collection ROUteS sranie E E 54 Table 6 Summary of Duty Cycle Time Fraction for MAP Modes 00nnnnnennnnnnnnnnnnnsssssssssesee 57 Table 7 Total Fuel Use and Emissions based on Overall Representative Real World Duty Cycle for Three B 20 Fueled Combination Trucks with and without Fuel additive 0 61 Table 8 Time Based Average Fuel Use and Emission Factors based on Representative Real World Duty Cycles for the Combination Truck 215 5 As oe eee enn eee eT eee ete 64 Table 9 Fuel Based Emission Factors based on Representative Real World Duty Cycles for the Combination Truck 215 5715 v cccccccccscsscsessssssecesssscsesessvsesecsssvsecassvssecavavsucasavsusecavavsesesavavaneasavees 65 Table 10 Travel Distance Based Average Fuel Use and Emission Factors based on Representative Real World Duty Cycles for the Combination Truck 215 5 TN e 66 Table 11 Time Based Average Fuel Use and Emission Factors based on Representat
32. Paper No 2001 01 3641 Society of Automotive Engineers Warrendale PA 8 1 Weilenmann M Soltic P Ajtay D 2003 Describing and compensating gas transport dynamics for accurate instantaneous emission measurement Atmospheric Environment 37 2003 5137 5145 Weilenmann M Bach C Rudy C 2001 Aspects of instantaneous emission measurement Int J Vehicle Design Vol 27 Numbers 1 4 2001 p 94 104 82 APPENDIX 83 Appendix A Detailed Emission Factor Results for Individual Tested Vehicles This appendix contains the following results for each of the three tested vehicles e Time based modal fuel use rate and emission factors with respect to normalized Manifold Absolute Pressure MAP for B20 with and without the Fuel additive e Fuel based modal emission factors with respect to normalized Manifold Absolute Pressure MAP for B20 with and without the Fuel additive In addition there are several tables that summarize the mean exhaust gas concentrations for each vehicle and fuel that was tested The purpose of these tables is to identify situations in which the mean pollutant concentration was at or below the gas analyzer detection limit in order to provide insight regarding data quality A comparison is given between the measured and actual fuel use for each of the 6 tests This comparison illustrates that there is excellent agreement between the fuel use measured based on the portable emission measurement
33. Quality Assurance of Portable Emissions Measurement System Data enaena T E tenet ea tee l l TOI O eea E E A E E re E A O E l 2 Monvao isoo a T E eee ee oe l 2e DataS ample SZE Sea cada evn T E l 223 Nom OFP AMES re ERE 2 2 95 sD VINCI OMIZAU O Merere E E E 4 24 Need For Rapid Onsite QA nesuuredenoianieionnaun annia Gane 7 gt Data Quality And Error Identification ee EE EOR EREA 8 4 INES mMOdOlO Vasra da eE vhost tune eee ee N E 14 Alig Sensitivity ATIALYSIS seseina cwsuslostedetedesvosamsenan Sevan nA aai EET 14 A2 VTA CMO MIAO eana oeeie iios 15 Ao Intake Air Temperatura E A 16 AA Road Grade ESMI O raae a E E E 17 5 Sei 3 Gc eae eee ce Nae rte ae Ee re Noa een rs MIO EAT oor ne nS RIOT een Core 17 Sla Sens Y PATA Y SUS ete isc els hans ot are tiara T cartel Sat eae eraeas ct 17 5 2 SyNCNTOMZA ONM sarena nea eaaa aa eaa aaee ESTA ea eS 20 Sde Anake Air Pemperauite eruo n E eee 27 34 R d Grade Sta ON esgor aT a T 28 6 CONCUSSIONS aneen T a E E a eee aves 32 Chapter 2 Methods for Field Comparison of Alternative Fuels 00 ccccee 33 1 0 PANCRO CNC HON 5226s secs ssehct serch e A TE weleoncaatah A aaah Beene eave 33 20 MACK COUN pct stash cesta a nace haar oda cata aaece deanna 34 3 0 TE CCNAIC AWA PLO ACW orco toasts ae ei dea Be Oe 35 3 1 Portable Emission Measurement System cccccccccceeeeeeeeeeeeeeeeeeeeeeeees 35 3 1 1 Manifold Air Boost Pressure Sensot ccccccccccccscssseseseeee
34. Road Grade Estimation Road grade calculated using GPS data from real world in use emissions testing of a LDGV with RGstep values of 0 1 mile 0 05 mile and 0 25 mile is shown in Figure 14 A smaller RGstep 1s better able to capture the change in road grade while a larger RGstep 1s more immune to noise in elevation and position data Road grades calculated using RGstep 0 05 mile have higher rate of change than actual and more seconds have values outside of the 10 range which is the typical range for city and highway roads RGstep value of 0 25 mile result in loss of resolution of actual road grade variation especially for hilly road sections In the field test the GPS unit recorded a total of 14623 seconds of data Road grade calculated with RGstep of 0 1 mile resulted in 0 34 seconds falling outside of 10 range Similarly for RGstep values of 0 01 mile and 0 25 mile 0 56 and 0 seconds had values outside of 10 range However road grade calculated as change in elevation per second over distance travelled per second produced 22 seconds with road grade outside of 10 range 28 A cumulative frequency of the road grades using different RGgtep values for the complete test duration is shown in Figure 15 Road Grade Timeplot 12 0 RG_0 1mi 10 0 R f RG_0 05mi Road Grade 12 0 T T T T T T T T T T 1 9200 9300 9400 9500 9600 9700 9800 9900 10000 10100 10200 Data Row Figure 14 Comparison of road grades
35. S e Ses CSTD Co oe CCBBOCOSIR AG OH Pearson s Correlation Coefficient 50 0 50 Slave Parameter Timestamp Shift s Figure 6 Correlation between NO PEMS and RPM OBD after synchronization In another example the OBD and GPS used while testing a Pontiac Grandprix 1999 model 3 1 liter engine 160 HP are synchronized using vehicle speed from OBD as master and vehicle speed calculated from GPS coordinate information as slave The GPS timestamp is advanced by 3 seconds to achieve synchronized datasets Figure 7 and Figure 8 show before and after synchronization time series plots Figure 9 and Figure 10 show PCC plots before and after synchronization 22 60 Speed OBD 50 AA P N Speed GPS 2 2 E E ro fal ay 7 3 7 7 S 20 3 N N 10 0 53020 53070 53120 53170 Timestamp sec Figure 7 Vehicle Speed from OBD and Vehicle Speed from GPS before synchronization 60 Speed OBD 50 F g i N Speed GPS 40 30 20 Speed_OBD mph 10 53020 53070 53120 53170 Timestamp sec Figure 8 Vehicle Speed from OBD and Vehicle Speed from GPS after synchronization 23 Speed_GPS mph Pearson s Correlation Coefficient 50 0 Slave Parameter Timestamp Shift s Figure 9 Coefficient of correlation for vehicle speed from OBD and GPS before synchronization Pearson s Correlation Coefficient 50 0 Slave Parameter Timestamp Shift s Figure 10 Coefficient of corr
36. ak in the exhaust sampling line upstream of the gas analyzers would result in excess air lower CO CO2 NO and HC mole fractions and increase of O2 and N gt 11 mole fractions Air leakage does not affect the mass emission rates for the pollutants because the decrease in pollutant mole fractions is balanced by the increase in dry exhaust molecular flow rate which is calculated from the intake air flow rate Frey et al 2008 The ratio of intake air mass flow rate to fuel use rate known as AFR g air g fuel is used to test for air leakage An acceptable AFR can be set based on field data For example for construction vehicles 99 9 of the s b s field data has AFR value within the range 25 to 150 g air g fuel Frey et al 2008 When AFR is within acceptable range the AFR value is used When AFR value is outside of acceptable range the random measurement error in the pollutant concentration may result in large uncertainty in the mass emission rate In such a scenario the concentration reading is compared to the sensor precision If pollutant concentrations are above the precision then the second of data is used otherwise the second of data is deleted Errors related to data from GPS are described below VSP Modal Binning Error Second by second GPS elevation data based on barometric pressure can be used to calculate road grade which in turn is used to calculate VSP Thus incorrect road grade estimates can affect the VSP based modal mod
37. ank for three of the tests For one test NCDOT did not record the actual amount of fuel added after the test therefore no comparison of estimated versus actual fuel use is possible For Truck 5715 as previously explained the conditions of the test do not permit an accurate estimate of the fuel consumption As shown in Figure A 7 in the Appendix the PEMS estimate of fuel consumed for this truck 1s biased low because of the effect of air leakage in the tailpipe due to the ruptured exhaust pipe which leads to an over estimate of the air to fuel ratio In turn this leads to an underestimate of the amount of fuel consumed 74 Fuel use during an actual duty cycle is sensitive to the distribution of time that the engine is subject to different loads Engine load is represented by MAP The frequency distributions of MAP for each truck comparing the tests with and without the additive are given in the Appendix in Figures A 8 to A 10 For example Figure A 9 is for Truck 6415 Although there are some minor differences in the distribution of MAP for each test the average MAP was nearly identical at 176 to 177 kPa and the variability in MAP was very similar with a standard deviation of 60 to 65 kPa and a range of 96 to 306 kPa Likewise for the other two trucks the average standard deviation and range of values of MAP were similar for both tests Hence we do not suspect that there was enough variability in the activity patterns of the duty cycle
38. asure emissions from any vehicle that passes through the infrared and if available ultraviolet beams that are used to measure pollutant concentrations Each measurement is only a snap shot at a particular location and thus cannot characterize an entire duty cycle 34 On board emissions measurement systems offer the advantage of being able to capture real world emissions during an entire duty cycle In particular Portable Emissions Measurement Systems PEMS that are more easily installed in multiple vehicles than complex on board systems are selected for use in this study 3 0 Technical Approach The general technical approach for this project involved four major components 1 the Portable Emission Measurement System PEMS instrumentation 2 preparation for field data collection 3 field data collection and 4 quality assurance and quality control Each of these components of the technical approach 1s described 3 1 Portable Emission Measurement System The OEM 2100 Montana system is comprised of two parallel five gas analyzers a PM measurement system an engine sensor array a global position system GPS and an on board computer The two parallel gas analyzers simultaneously measure the volume percentage of carbon monoxide CO carbon dioxide CO2 hydrocarbons HC nitrogen oxide NO and oxygen O2 in the vehicle exhaust The PM measurement capability includes a laser light scattering detector and a sample conditio
39. ccurate and the visual technique of matching rising or falling slopes is to be used as the ultimate check Improperly synchronized OBD and GPS datasets with respect to PEMS dataset can affect VSP based fuel use rate and NOx mass per time emissions rate by 33 and 49 MAP based fuel use rate show lesser sensitivity to time misalignment but even then the modal fuel use rates can be affected by up to 26 In all cases there is substantial reduction in the highest modal rates and less variability when comparing highest to lowest modal rates The use of a least square fit line approach provides road grade estimates that are more robust as compared to a simple second by second gradient calculation to elevation data noise and overall the estimates are consistently within actual road grade range The VSP modal results are not significantly sensitive to road grade estimates estimated using this approach which further highlights the usefulness of this technique 32 Chapter 2 Methods for Field Comparison of Alternative Fuels 1 0 Introduction The purpose of this study is to apply a methodology for real world evaluation of a fuel additive in order to assess the change in fuel economy and emissions before and after the additive is used The methodology features the use of portable emission measurement systems PEMS for the purpose of quantifying the activity fuel use and emissions of vehicles during actual duty cycles A baseline test for B20
40. cession and all three trucks were tested within the second and third week of June 2009 A period of approximately one month occurred between the initial baseline test for B20 and the comparison test with the Fuel additive This time period was used to break in and condition the vehicle with the Fuel additive Usage of the Fuel additive began immediately after the baseline test was completed Each truck was run through two tankfulls of B20 with Fuel additive before it was tested for B20 with Fuel additive Also the fuel filters were replaced a day before the test with Fuel additive Table 3 Data Collection Schedule Period Vehicle Phase Test Fuel 215 5715 215 6415 215 6667 B20 No Additive 06 09 2009 06 16 2009 06 10 2009 06 09 2009 06 16 2009 B20 B20 with Fuel 07 08 2009 08 11 2009 07 08 2009 07 14 2009 08 11 2009 additive B20FA B20 20 Biodiesel 80 Petroleum Diesel 4 2 Vehicle Characteristics and Test Conditions An example of a tested vehicle is shown in Figure 7 The detailed characteristics of each tested vehicle as well as the conditions of the 6 tests for each vehicle are given in Table 2 The vehicles tested included one 1999 model International 2574 6x4 tractor with a Cummins ISM 370 10 8 liter 6 cylinder 370 hp engine one 2004 International 9400I 6x4 tractor with a Cummins ISX 500 15 0 liter 6 cylinder 500 hp engine and one 2007 International 9200I 6x4 tractor with a Cummins ISX 500 15 0 liter
41. commended once every three months The gas analyzer NDIR subsystem used in the gas analyzers is very stable and tends not to drift significantly from their span calibrations Data from several laboratories using various vehicles and fuels suggests that when the Montana System is operated simultaneously with the laboratory system the difference is typically less than 10 for aggregate mass NO and CO The accuracy of HC and CO measurements depends on the fuel used and on the emission levels Vojtisek Lom and Allsop 2001 3 1 7 System Setup and Operation The time to install the instrument on the study trucks is typically two hours Figure 3 illustrates several aspects of the installation of the PEMS using the example of truck 5715 In Figure 3 the portable instrument is shown including its placement inside the vehicle and the connections for DC power engine data and exhaust sampling hoses Figure 4 shows the use of a cigarette lighter port to draw the required DC power necessary to power the PEMS main unit the placement of the GPS antenna on the roof and setup of the exhaust sampling lines from the truck exhaust pipe to the passenger cabin Figure 5 displays the routing of sampling hoses to the instrument via the passenger window the MAP sensor and the engine RPM sensor 40 Figure 20 Installation of the Portable Emissions Measurement System PEMS Main Unit in a Truck 215 5715 a the portable unit on a passenger seat fr
42. dynamometers tunnel studies remote sensing and on board measurement Available data regarding heavy duty vehicle emissions is typically from engine dynamometer measurements These data are reported in units of g bhp hr which are not directly relevant to in use emissions estimation Furthermore many engine dynamometer test cycles are based upon steady state modal tests that are not likely to be representative of real world emissions Although there are also transient engine dynamometer tests it is not likely that any particular standardized test cycle will be representative of operation of a particular type of vehicle and real world duty cycle Chassis dynamometer tests provide emissions data in units that are more amenable to the development of emission inventories such as grams of pollutant emitted per mile of vehicle travel This emission factor can be multiplied by estimates or measurements of vehicle miles traveled to arrive at an inventory These tests are expensive and the number of heavy duty dynamometer facilities is limited The applicability of chassis dynamometer test results to real world emissions is limited by the potential lack of representativeness of standard test cycles Tunnel studies are based upon measurements for a specific link of roadway and thus are not representative of an entire duty cycle Tunnel studies are limited in their ability to discriminate among specific vehicle types Remote sensing can be used to me
43. e additive 76 For fuel economy miles per gallon or fuel consumption rate g sec or g mile comparisons could be made for Trucks 6415 and 6667 based on estimates from the PEMS data Comparisons could be made based on the actual amount of fuel added to the truck after a test for Trucks 5715 and 6667 Based on the PEMS data for Trucks 6415 and 6667 there is no observable evidence of a beneficial effect of the additive in terms of reducing fuel consumption With respect to emission rates there is some variability in results among the three trucks For example based on fuel based emission factors one of the trucks had lower NOx emissions with the additive while the other two had slightly higher NO emissions Two of the three trucks had slightly lower PM emission rates The results for HC and CO are generally inconclusive Diesel engines typically have low emission rates of HC and CO because they operate with excess air which promotes efficient combustion Many of the measurements of HC and CO concentrations were below the detection limit of the gas analyzers Thus the average emission rates are subject to substantial uncertainty which leads to inconclusive comparisons Despite some of the challenges encountered particularly with Truck 5715 the data that were used for analysis underwent a rigorous quality assurance screening process Therefore the data that were used as the basis for comparisons are deemed to be valid Overall the re
44. e negative values that are not statistically different from zero or a small positive value Diesel vehicles typically produce HC emissions less than gasoline vehicles do Durbin et al 2000 Thus it 1s frequently the case that HC emission measurements are very low and not substantially different from zero Negative values of emissions estimates were assumed to be zero and were replaced with a numerical value of zero There were no negative values observed in these tests that were significantly different from zero Air Leakage Air Leakage quality procedure is used to eliminate some of the data which affected by the problem of ambient air infiltration into the exhaust gas sampling stream This infiltration could occur anywhere in the exhaust gas flow path between vehicle exhaust pipe and gas analyzers Air Leakage is decided based on value of Air to Fuel Ratio AFR which is the 48 mass basis ratio of intake air to fuel consumption A data second having AFR greater than the threshold value of AFR and one or more gas concentration lower than PEMS s precision level is eliminated Invalid Data Sometimes the PEMS would not record valid concentration data for the gases and or valid RPM value In such instances the value for the parameter gas concentration or RPM 1s recorded as zero and the corresponding column for validity 1s marked as NO by the PEMS These seconds of data are marked as invalid by the QA procedure This error i
45. e solved easily by rebooting the system in the field After rebooting the computer begins logging a new data file automatically Thus when this is noticed in the field this error can be addressed Loss of engine data is also obvious from the data file since the missing data are evident and any calculations of emission rates are clearly invalid The following types of engine errors are included in the quality assurance procedure 1 Unusual Engine RPM Engine RPM typically varies from not less than 600 RPM during idling to about 3 000 RPM during most kinds of vehicle operation As a conservative estimate the bounds for possible engine RPM were set as greater than or equal to 600 RPM and less than or equal to 10 000 RPM Qiao et al 2005 Thus if engine RPM is less than 600 or greater than 10 000 RPM those data need to be removed for the further data analysis However this problem did not occur in any of the data collected in our previous work for NCDOT This error occurred only briefly during one test 46 2 Engine RPM Freezing Freezing refers to situations in which a value that is expected to change dynamically on a second by second basis remains constant over an unacceptably or implausibly long period of time Engine RPM tends to fluctuate on a second by second basis even if the engine is running at approximately constant RPM Therefore we performed a check to identify situations in which engine RPM remained constant for more
46. ed sensors The temperature and speed signal data is collected by 38 the sensor array box and converted from an analog to a digital RS 232 serial signal which is transmitted to the PEMS main unit The sensor array box was placed in the driver cabin close to the PEMS main unit The temperature and speed sensors which were in the engine compartment are connected to the sensor array box using appropriate cables 3 1 5 Operating Software The Montana System includes a laptop computer that is used to collect and synchronize data obtained from the engine scanner gas analyzers and GPS system Data from all three of these sources are reported on a second by second basis The computer is controlled either by touching the screen or plugging in a keyboard Upon startup the computer queries the user regarding information about the test vehicle fuel used test characteristics weather conditions and operating information Most of this information is for identification purposes However the fuel type and composition engine displacement sample delivery delays unit configuration intake air sensor configuration and volumetric efficiency are critical inputs that affect the accuracy of the reported emission rates The details of the definition and significance of each of these are detailed in the Operation Manual of the instrument CATI 2003 The software provides a continuous display of data during normal operation including gas analyzer data engi
47. eeeeees 36 S 2 Enone Speed SCNSOF ssc sheccdacinsseantyilacsaeiaanusueadedadniceenivide uae 37 3 1 3 Intake Air Temperature Sensor ccccccceccccccccceceeeeeeeeeeeceeeeeaas 38 SLA eUSOr Ait Ay BOX srcane a teenie 38 Did OPETAUNG SOWIE arreir ioien nan aT as elena E ETE 39 3 1 6 Validation and Calibration cccccccccccccccceceeeeeeeeeeeeeeeeeseseeeees 39 3 1 7 System Setup and Operation scscccceeeeeesecccccssnennanecsseeeees 40 3 2 Preparation for Field Data Collection cc cceccccccccccccseeseeeecceeeeaaeeeees 42 3 3 Field Data Collecuon Procedure seenen vain seine edie Revieatvaiemoehte 43 3 4 Quality Assurance and Quality Control ccccssssssssessseeeeeeeeeeeeeees 46 4 0 Pe SUMS att tec tamh A E AE hone 49 4 1 scheduling of Field Data Collection cccccceceeeeeccecetsenenaasasseeeees 50 4 2 Vehicle Characteristics and Test Conditions cec cece ceecceeccessceeeceee scene 50 4 3 Sites ame Truck ROUES mesnine Oates a ened alee ce tee teases 54 AA Real World Duty C ycl ssossoisa a 56 4 5 Estimation of Fuel Use and Emission Rates for Baseline Test with B20 58 4 6 Quality Ass rahce RESUS oeseri a Ta 60 4 7 Results of Fuel use and Emission Rates for Each Tested Truck 60 4 8 Results of Fuel use and Emission Rates for Each Tested Truck 62 5 0 OTIC WEES LOIN a eE onic TE O EEE 76 REFERENCES ea T TAOTE T T E cian he TAE 78 APP
48. eighting the modal rates using a common set of three duty cycles The three duty cycles represent arterials highways and an overall mix of arterials and highways The trucks had only modest loads during each test which were delivered from Raleigh to specific destinations along the route The loads varied from approximately 3 000 to 8 000 Ibs compared to an unloaded weight of approximately 33 000 to 36 000 Ibs depending on the truck Since the loads are modest compared to the overall weight of the truck and since the loads and mileage associated with delivering the loads were similar for tests with and without additive these are not likely to affect the comparisons There was a problem with Truck 5715 in that the exhaust pipe ruptured partway into the test with additive leading to a discovery that it likely had a crack that grew in size during the course of operations up until that time and dating probably to before the baseline test The crack meant that ambient air was drawn into the exhaust pipe which lead to biasing the concentration measurements downward This does not affect estimation of fuel based emission rates which are based on molar ratios of pollutants in the exhaust and fuel properties but it does bias estimates of mass per time and mass per distance fuel use and emission rates for this truck Therefore only fuel based emission rates for Truck 5715 were used as a basis for comparing performance with the additive to without th
49. el results As an example the specifications for the GPSMAP 76CSx are GPS latitude and longitude accuracy lt 10 m DGPS accuracy Differential GPS 3 5 m Velocity accuracy 0 05 m sec steady state Altimeter accuracy 10 feet Garmin 2009 Road grade is estimated as AE t l RG 100 Equation 11 t t t l Where RG road grade at time t in AF t 1 change in elevation from time t 1 to time t m Dit1 distance travelled from time t 1 to time t m 12 Field tests show GPS elevation data and X Y coordinate data have significant precision related noise and using Equation 10 at a 1Hz resolution can result in spurious estimates of very large road grades such as 70 in some cases Real road grades are typically between 10 A moving window averaging technique could be applied to elevation data to smooth out some of the noise but it does not solve the issue of occurrence of impossible range road grades Moreover the same averaging technique cannot be applied to X Y coordinate data if the vehicle is moving Further Light Detection and Ranging LIDAR data may not be available for a given geographic area test route Road grade from Geographic Information System GIS may not be reliable depending on how such data were obtained In view of these limitations it 1s important to have a method to get good road grade estimates using on board GPS elevation data 13 4 Methodology 4 1 Sensitivi
50. elation for vehicle speed from OBD and GPS after synchronization 24 50 50 Effect of Synchronization on VSP and MAP Modal Model Results Sensitivity analysis of VSP and MAP modal model results with respect to synchronization was conducted using real world in use emissions measurement of a 1997 make 2 2 liter 130HP Honda Accord with 228 000 miles on the odometer Starting with second by second data approx 12550 seconds the datasets for PEMS OBD and GPS are synchronized using the visual and PCC technique explained previously This is the base case and subsequent cases are created by shifting either OBD or GPS data by 5 or 10 seconds VSP based results for fuel use and pollutant mass emission rates are generated for both OBD and GPS data shifts while MAP based results for fuel use and pollutant mass emission rates are generated for only OBD data shifts since GPS data does not influence MAP based results VSP based fuel use VSP based NO mass emissions rate and MAP based NO mass emissions rate are significantly affected when OBD data is out of sync with PEMS and GPS data Results for VSP and MAP based fuel use For OBD data out of sync by 5 seconds fuel use rate increased by 12 6 for VSP mode 1 and 6 1 5 5 and 6 1 for VSP modes 2 3 and 4 respectively For the same case fuel use rate decreased by 3 and 4 5 for VSP modes 11 and 13 A shift of 10 seconds to OBD data resulted in fuel use numbers for VSP modes 1 2
51. en problems are identified in the field then attempts are made to resolve the problems in the field For example if a leak or overheating problem is detected during data collection then the problem is fixed and testing resumes 47 Data recorded during the period when a leak or overheating event occurred are not included in any further analyses However no gas analyzer errors were detected in these tests Zeroing Procedure For data quality control and assurance purpose each gas analyzer bench is zeroed alternatively every 15 minutes While zeroing the gas analyzer will intake ambient air instead of tailpipe emissions After zeroing is finished a solenoid valve changes the intake from ambient air to the tailpipe There is a period of transition when this occurs In particular the oxygen sensor needs several seconds to respond the switching of gases since there is a large change in oxygen concentration when this switch occurs To allow adequate time for a complete purging of the previous gas source from the system a time delay of 10 seconds is assumed Thus for 10 seconds before zeroing begins the time period of zeroing and 10 seconds after zeroing ends data for the bench involved in zeroing are excluded from calculations of emission rates and the emission rates are estimated based only upon the other bench Negative Emissions Values Because of random measurement errors on occasion some of the measured concentrations will hav
52. ensitivity Analysis Base Case and Typical Range for Selected Input Parameters MAP NOx HC RPM IAT C kPa ppm ppm CO CO O2 15 0 liter HDD Truck 54 265 134 9 0 045 I2 12 8 30 80 98 300 10 200 5 45 0 0 08 0 5 9 0 10 19 2 2 liter LDG Car 46 83 290 32 0 059 14 11 0 23 37 60 15 90 10 400 2 50 0 0 3 8 14 5 0 02 1 0 14 4 2 Synchronization Of the two data sets that need to be synchronized one is labeled master and the other is labeled slave If PEMS data is one of the two datasets involved in synchronization then it is considered as master and the OBD or GPS datasets are considered slave If the OBD dataset is being synchronized with GPS then OBD is master and GPS is slave The choice of master and slave is a matter of convention and a different choice would not affect the final result as long as a consistent scheme of selecting master and slave is made Frequently used master slave synchronization pairs are given in Table 2 Table 2 Frequently used dataset parameter pairs to synchronize datasets Master Parameter Source Slave Parameter Source CO gt PEMS RPM OBD CO PEMS RPM OBD NO PEMS RPM OBD Vehicle Speed PEMS Vehicle Speed OBD Vehicle Speed PEMS Vehicle Speed GPS Vehicle Speed OBD Vehicle Speed GPS These pairs of parameters are used for synchronization because their time series shows correlated trends For example a large change in engine RPM over a short peri
53. ented in Tables A 1 and A 2 The estimated mass of fuel use and emissions for Truck 5715 are biased low because of perforation of the exhaust pipe by rust that was detected partway through the test with Fuel additive The numbers reported for this truck are for informational purposes only Because the amount of dilution likely differed when comparing the baseline test and the test with additive ratios of the mass per time rates for with additive to without additive are not shown for this truck B20FA 26 2 0 68 NO W OO ON p Table 9 Fuel Based Emission Factors based on Representative Real World Duty Cycles for the Combination Truck 215 5715 Fuel Representative Real World Dut Highwa Arterial Overall 113 113 112 B20 6 44 6 98 8 1 0 21 0 22 110 110 109 B20FA 6 10 6 89 13 6 157 13 8 0 26 0 27 0 26 0 97 0 95 2 16 1 22 0 97 0 95 2 14 L22 B20FA B20 O UJ ON gt Ia IO N ka D NO ON 20 Biodiesel 80 Petroleum Diesel 20 Biodiesel 80 Petroleum Diesel with Fuel additive The concentrations of HC and CO are lower for some MAP ranges than the detection limits of PEMS as documented in Tables A 1 and A 2 d The fuel based emission factors for NO CO and HC are not affected by the diluation of exhaust with ambient air because of perforation of the exhaust pipe with rust Hence ratios of these emission rates with additive to those without additive are shown Ho
54. er for some MAP ranges than the detection limits of PEMS as documented in Tables A 3 and A 4 ON 9 Table 14 Time Based Average Fuel Use and Emission Factors based on Representative Real World Duty Cycles for the Combination Truck 215 6667 Fuel Representative Real World Dut Highwa Arterial 20 Biodiesel 80 Petroleum Diesel 20 Biodiesel 80 Petroleum Diesel with Fuel additive The concentrations of HC and CO are lower for some MAP ranges than the B20 34 9 0 08 11 2 B20FA 34 7 0 07 1 00 1 05 1 66 8 70 1 00 0 98 B20FA B20 m KO ON leloa detection limits of PEMS as documented in Tables A 5 and A 6 Table 15 Fuel Based Emission Factors based on Representative Real World Duty Cycles for the Combination Truck 215 6667 Representative Real World Dut Arterial 17 B20 4 j 0 02 18 B20FA HC o gall 6 7 6 7 0 02 B20FA B20 paa 20 Biodiesel 80 Petroleum Diesel P 20 Biodiesel 80 Petroleum Diesel with Fuel additive The concentrations of HC and CO are lower for some MAP ranges than the detection limits of PEMS as documented in Tables A 5 and A 6 71 Table 16 Travel Distance Based Average Fuel Use and Emission Factors based on Representative Real World Duty Cycles for the Combination Truck 215 6667 NOx g mile 33 60 35 B20 HCIe mile 07 1 08 PMJg mile 000 001 000 B20FA CO g mile 07 2 08 PMJg m
55. erall Cumulative Frequency 0 0 2 0 4 0 6 0 8 l Normalized MAP The Normalization of MAP is between 96 kPa and 306 kPa Seconds Min MAP Max MAP Avg MAP Highway 5397 306 18410 96 306 203 Figure 27 Cumulative Frequency of Normalized Manifold Absolute Pressure for Each Duty Cycle 4 5 Estimation of Fuel Use and Emission Rates for Baseline Test with B20 The average fuel use and emission rates as well as total fuel use and emissions were estimated for the typical duty cycle for each vehicle in order to have a consistent basis for comparison Appendix A provides detail on time based modal fuel use and emission rates for B20 with Fuel additive and B20 without additive for each of the three tested vehicles For each vehicle the range of variability in engine manifold absolute pressure MAP was normalized 58 based on the minimum and maximum observed values of MAP and divided into 10 ranges These ranges are referred to as modes An average fuel use or emission rates were estimated for each mode As the value of MAP increases the average fuel use or emission rate for the corresponding MAP based mode increases monotonically An example can be seen for Truck No 215 6415 in Figure A 3 of Appendix A The frequency distribution of MAP for the typical duty cycle shown in Figure 10 was used to estimate the proportion of time spent in each of the 10 MAP based modes Weighted average fuel use and emission rates were e
56. es engine speed intake air density and the ideal gas law under constant engine volumetric efficiency and engine parameters measured by OBD or engine sensor array When multiplied with molecular weight of air it gives intake air mass flow rate P V engine MAP a engine 30N Mair M p e Moncine Equation 3 R T 273 15 P Fa GT ae a Equation 4 engine Where Cengine engine compression ratio 9 5 for light duty vehicle Ns number of strokes of engine 2 or 4 Pwuap a adjusted manifold absolute pressure kpa Pmap manifold absolute pressure MAP kpa PB barometric pressure kpa R universal gas constant 8 314 Jmol 1K 1 Sengine engine speed RPM rpm Vengine engine displacement L Tintake intake air temperature IAT C Nengine engine volumetric efficiency 0 85 for light duty vehicle Mair MairM Wair Equation 5 The dry basis molecular exhaust flow rate Me is computed using combustor mass balance equations The simplified equation for exhaust flow uses intake air molecular flow rate and the dry basis mole fractions of measured species in the exhaust 2V55 M Yo2 in air Equation e 2V Vogi T Vison Yno 0 5mMy yc ay Veci t Voot NY pc 0 5x z Pollutant mass emission rates m and fuel use rate my can be derived from dry basis mole fractions of pollutants and exhaust flow rate as per m M y MW Equation 7 m M Yoo Yoo NYyc MW ei Equation 8 w
57. eventies in Jamshedpur a town in eastern India He acquired a Bachelor of Engineering in Instrumentation Engineering in 2001 from Sant Longowal Institute of Engineering and Technology Thereafter for the next seven years worked with a young engineering startup named Quantum Age Tech Solutions QAT where he learnt how to build automated computer controlled test and measurement systems At QAT he led development teams that a designed and developed the pre processor and post processor for a tire noise simulation software and b development of a tire extruded rubber offline profilometer During the same time he also successfully developed business in overseas markets for the tire noise simulation software and negotiated exclusive India representation for Boltcalc a bolted joint analysis software from a UK based company In 2008 Gurdas chose to pursue his long standing desire to learn about and make a positive contribution to the environmental challenges facing the present generation He joined North Carolina State University s department of Civil Construction and Environmental Engineering as a Master s student and chose to continue towards a PhD in the same area He spends much of his time conducting measurements of vehicle emissions and writing software to conduct quality assurance on the data from these measurements In his spare time Gurdas likes to get behind the lens and capture the world as he sees it He also writes a blog whe
58. here Mi the mass per time for specie i i NO HC CO CO and fuel MW the molecular weight for specie i NO HC CO CO and fuel Air to Fuel ratio AFR is calculated as the ratio of intake air mass flow rate over fuel mass flow rate AFR We Equation 9 When intake air mass flow rate is available from OBD it is preferred over the value calculated as per speed density method When exhaust molecular flow rate is available again the measured value is substituted in place of the computed value A step by step derivation of key equations is given in Appendix A The mass emission rates and fuel use rate are compared using a VSP Vehicle Specific Power modal model or MAP modal model A VSP based model was first introduced in Jim nez Palacios s Ph D thesis Jime nez Palacios 1999 VSP for LDGV is calculated as VSP v 1 1a 9 81 sin arctan 0 132 0 000302v3 Equation 10 Where VSP Vehicle Specific Power kw ton m s V vehicle speed m s a vehicle acceleration m s r road grade Both VSP Jimenez Palacios 1999 and MAP are good indicators of engine load which in turn is a good measure of fuel consumption VSP provides a method to categorize and explain the variability in fuel use and tailpipe emissions Frey et a 2002a VSP accounts for power demand rolling resistance road grade and aerodynamic drag and can be estimated based upon second by second speed from OBD or GPS acceleration
59. ifold Absolute Pressure B20 20 Biodiesel 80 Petroleum Diesel B20FA B20 with Fuel additive Figure A 3 Average Fuel Use and Emission Rates of Each Pollutant on a Per Time Basis for Engine Based Modes for Combination Truck 215 6415 89 HC mg sec Opacity mg sec CO mg sec GB20 W B20FA 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 l Normalized Manifold Absolute Pressure mB20 m B20FA 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 I Normalized Manifold Absolute Pressure 140 120 100 80 60 40 20 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 Normalized Manifold Absolute Pressure 90 100 GB20 W B20FA T 75 T 4 50 Z 25 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 Normalized Manifold Absolute Pressure BB20 mB20FA E C 2 5 5 z A 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 Normalized Manifold Absolute Pressure B20 20 Biodiesel 80 Petroleum Diesel B20FA B20 with Fuel additive Figure A 4 Emission Rates of Each Pollutant on a Per Fuel Basis for Engine Based Modes for 215 6415 Combination Truck 9 HC g gallon CO g gallon 120 100 80 60 40 20 0 1 0 1 0 2 0 2 0 3 0 4 0 5 0 6 0 7 Normalized Manifold Absolute Pressure 0 3 0 4 0 5 0 6 0 7 Normalized Manifold Absolute Pressure 92 GB20 BB20FA 0 8 0 9 l BB20 m B20FA 0 8 0 9 I BB20 WB20FA a S 2 of z D es 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 Nor
60. ile 0o00 001 000 B20FA B20 098 096 097 20 Biodiesel 80 Petroleum Diesel P 20 Biodiesel 80 Petroleum Diesel with Fuel additive The concentrations of HC and CO are lower for some MAP ranges than the detection limits of PEMS as documented in Tables A 5 and A 6 2 For example as shown in Table 9 for Vehicle 215 6415 the time based average fuel use rate for B20 without additive varies from 7 0 g sec for travel on arterial duty cycle to 10 6 g sec for travel on highway duty cycle with an average of 10 5 g sec for overall duty cycle The variability in fuel use rate among different components of the duty cycle indicates the variation in engine load among these components While operating on B20 without additive the average time based NO emission rate for Vehicle 215 6415 varies from 68 mg sec for travel on arterial duty cycle to 99 mg sec for travel on highway duty cycle representing an increase of about 50 with an average of 99 mg sec for overall duty cycle The ratios of the fuel use and emission rates for the highway arterial and overall duty cycles are given in the bottom section of Table 9 based on comparison of fuel with additive to fuel without additive The values shown here for the overall duty cycle are also reported in Table 3 In general there were similar results for the ratio with to without additive for fuel use and emissions when comparing the duty cycles The fuel based emission
61. ine crankshaft The engine speed sensor has a strong magnet to attach easily on metal materials The reflective tape must be installed on a pulley that is connected to the crankshaft The placement of the reflective tape and the optical sensor for a Cummins ISX 500 engine is shown in Figure 2 Some of the key factors in placement of the sensor include 1 avoid proximity to the engine cooling fan and other moving components 2 place the sensor in a location where the magnet can securely affix the sensor to a surface and 3 place oe the sensor so that its cable can reach the sensor array box which is located in the driver cabin The signal from the RPM sensor is transmitted by cable to a sensor array box which in turn transmits the signal by a second cable to the main unit of the Montana system Path of Bley Figure 19 Placement of Optical Engine RPM Sensor and Reflective Tape on ISX 500 Engine Truck 215 6415 3 1 3 Intake Air Temperature Sensor The engine intake air sensor is a thermocouple that is installed in the intake air flow path Installation of the intake air temperature sensor is somewhat easy compared to the engine speed and MAP sensors Using duct tape or a plastic tie one can fix the intake air temperature sensor near the intake air flow where the MAP port is located 3 1 4 Sensor Array Box The sensor array box provides signal conditioning and data acquisition for the intake air temperature and engine spe
62. instances of missing seconds for every 10 15 vehicles tested Both PEMS and GPS record data with integer timestamps OBD Scan tools used to read OBD ports often display three problems 1 recording frequency is sensitive to number of parameters being recorded 2 speed of laptop computer running the scan tool and 3 Recording is at variable frequency approximately equal to 1 Hz which produces data rows with timestamps of 3 4 4 4 5 6 7 8 and so on In one example recording speeds of approximately Hz are achieved when recording up to 6 parameters but as more parameters are added the recording speed falls to 0 5 Hz or lower Some OBD devices report a separate time stamp for each data point for each parameter or a group of parameters In one example an ECU data logger Kvaser Memorator for a Plug in Hybrid Electric Vehicle PHEV produced a file with 270 000 rows of data representing test duration of 20 045 seconds As an example the log file had some of the parameters recorded with timestamp 1175 638 another group of parameters were recorded in the following row with the same timestamp as previous row that is 1175 638 then other groups recorded with timestamps of 1175 639 1175 647 and 1175 647 Moreover missing seconds are possible due to temporary loss of communication within an instrument or between the recording instrument and a data source for example OBD scan tool recording data from the OBD port Thus there is a
63. ion of exhaust gases NOx HC CO CO2 and O2 from each of the two identical gas benches two benches operating in parallel to ensure when one of them is zeroing the other is measuring Particulate Matter PM concentration OBD Timestamp Engine RPM MAP IAT Vehicle Speed GPS Timestamp GPS X and Y coordinates and Elevation Thus a data file with 30000 seconds of measurement constitutes more than 600 000 measurement values 2 3 Synchronization The mass emission rate and fuel use rate for each second 1s calculated based on second by second data recorded by PEMS OBD and GPS The measured parameters from PEMS OBD and GPS that go into calculations should be for the same instance of measurement This can be a challenge because the internal clocks of these instruments that write timestamps in the data file are independent and not synchronized with each other A brief overview of the key equations and calculation steps is given here to illustrate how data from PEMS OBD and GPS are used PEMS measurements of the exhaust gas composition are used to get dry basis mole fractions of pollutants j Ear d Equation 1 10 Yi Equation 2 100 Where Yippm the concentration of measured specie i in ppm i NO and HC Yi the concentration of measured specie i in i CO CO2 and O2 Intake air molecular flow rate Mair 18 estimated as per speed density method Vojtisek Lom 1998 This method us
64. ith 18 Test Condition B20 B20 and Fuel additive gt June 09 2005 Test Date Truck Truck Weighi measured measured Cargo Continued on next page 53 Table 4 Continued Chassis 215 6667 License PK 8854 Model Combination 2007 Intemational of 6 92001 GVW 4 000 1B saxes wins O O OOS Test Condition B20 B20 and Fuel additive gt June 16 2005 Test Date Truck Truck Weight measured measured Cargo 4 3 Sites and Truck Routes Selection of sites and routes for on board data collection was determined by the normal work requirements of NCDOT According to the NCDOT work schedule 1440 140 US64 US70 US421 and US301 were traveled more than other roads 1440 is the beltline of Raleigh Cary area 1440 was driven for every truck route to leave or return to NCDOT Division 5 maintenance yard at Raleigh 140 and US421 were driven to visit the NCDOT at North Wilkesboro and Winston Salem 140 US117 and US421were traveled to visit the NCDOT at Castle Hayne Burgaw and Clinton Table 5 Summary of Data Collection Routes 6415 B20FA North Wilkesboro 140 US421 26936 6667 B20 North Wilkesboro 140 US421 26512 6667 B20FA North Wilkesboro 140 US421 29254 All run start and stop at Raleigh Based on odometer reading Includes time spent in loading and unloading 54 Figure 9 displays a graphical summary of the two routes that were included in the field data collection effort fo
65. ive Real World Duty Cycles for the Combination Truck 215 6415 wo ccssssssseeesseeeeeeeeeeeeeeeeaaaaas 67 Table 12 Fuel Based Emission Factors based on Representative Real World Duty Cycles for the COMODO THICK ZN SOAS a hater eran a a 68 Table 13 Travel Distance Based Average Fuel Use and Emission Factors based on Representative Real World Duty Cycles for the Combination Truck 215 6415 69 Table 14 Time Based Average Fuel Use and Emission Factors based on Representative Real World Duty Cycles for the Combination Truck 215 6667 ccccccccccccsssssessesseseeeeeeeeeeeeeeeaaaaas 70 Table 15 Fuel Based Emission Factors based on Representative Real World Duty Cycles for the Combination Truck 2137000 aonr A R N 71 Table 16 Travel Distance Based Average Fuel Use and Emission Factors based on Representative Real World Duty Cycles for the Combination Truck 215 6667 0 T2 Vil Table A 1 Average Concentrations for Each Pollutant Based on 215 5715 Combination Truck b es Le 6 Ue fe gal B2 eee merpecene ere rary ten renee h aretrenr entry mary ee rere naire entry try eer te Steen ae ee eter ee 97 Table A 2 Average Concentrations for Each Pollutant Based on 215 5715 Combination Truck Fueled Tor B20 with Fue l ACCC ve riire este soeet ese et oad ae eevee eee aed ee ee 97 Table A 3 Average Concentrations for Each Pollutant Based on 215 6415 Combination Truck Pueled for B20 tsissics cas ineaca ees Ae ead eee a ee ee 98 Table
66. ization 23 Figure 8 Vehicle Speed from OBD and Vehicle Speed from GPS after synchronization 23 Figure 9 Coefficient of correlation for vehicle speed from OBD and GPS before SY CMEOMIZALIOM E A ea ices aoasaie atari anes 24 Figure 10 Coefficient of correlation for vehicle speed from OBD and GPS after synchronization sda crete EE A A eee A EE A E 24 Figure 11 Effect of shifting OBD Data on VSP based Fuel Use Rate eeeeeeeeeeeeererrrrrrrrrrn 26 Figure 12 Effect of shifting OBD data on VSP based NO Emissions Rate n s 2T Figure 13 Test data with error IAT readings corrected by the IAT algorithm 0000000 28 Figure 14 Comparison of road grades for RGstep values of 0 1 mile 0 05 mile and 0 25 mile 29 Figure 15 CDF for roadgrades calculated using RGstep values of 0 1 0 05 and 0 25 mile 29 Figure 16 Effect of RGstep criteria on VSP based NOx Emissions Rate neeese 3 Figure 17 Effect of RGgtep criteria on VSP based HC Emissions Rate o e 31 Figure A 1 Average Fuel Use and Emission Rates of Each Pollutant on a Per Time Basis for Engine Based Modes for Combination Truck 215 5715 ccccccccccccceeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeees 85 Figure A 2 Emission Rates of Each Pollutant on a Per Fuel Basis for Engine Based Modes for Combination Truck 215457 1a sicsiesscesataatetietneceausbetclbaieausatiade tiletnedsasbetnau neti uinadasuat anal A 87 Figure A 3 Average Fuel Use and Emission Rates of Each
67. le 12 there is no difference in fuel use and CO2 emission rate a 5 percent increase in NO emission rate and a 3 percent decrease in PM emission rate based on the overall duty cycle Differences in HC and CO are not significant because many measurements were below the detection limits for these pollutants The ratios of emissions with versus without additive based on the fuel based emission factors differ somewhat from those obtained from the time or distance based emission factors The fuel based NOx emission rate was 3 percent lower for Truck 5715 with the additive versus without whereas it was 8 and 5 percent higher for Trucks 6415 and 6667 respectively The PM emission rate was lower for the latter two trucks by 2 to 10 percent but higher for Truck 5714 by 22 percent The results for HC and CO are inconclusive for reasons previously stated The fuel use for the three trucks was assessed in two ways One was based on recording the actual amount of fuel required to refuel the truck after completing a test The other was an estimate based on the second by second PEMS data The estimate 1s made by considering the exhaust composition which is used to infer the air to fuel ratio and the mass throughput of air to the engine which is calculated based on engine RPM MAP and IAT Detailed results of these comparisons are given in the Appendix The PEMS estimate of fuel consumption agrees well with the actual amount of fuel that was added to the t
68. le and for the overall duty cycle on a mass per time basis and a mass per gallon of fuel basis Mileage based emission factors mass per mile of vehicle travel are provided for the components of the duty cycle and for the overall duty cycle Fuel usage rates are reported on both a per time and per mile of vehicle travel basis 59 4 6 Quality Assurance Results The detailed results of quality assurance for each of the three tested vehicles and for each of the two tests of each vehicle are given in Table A 8 in Appendix A The average rate of loss of data among all six tests was 11 percent However most of the loss of data is attributed to episodes of unusually high air to fuel ratios Unusual RPM Unusual IAT Negative Concentration and Invalid Data contributed to the remaining percent An average of approximately 26 000 seconds of data was collected for each vehicle thus on average approximately 23 200 seconds of data were used for estimation of fuel use and emission rates for each vehicle As noted in Section 4 2 it was discovered during the test of Truck 5715 with additive that the exhaust pipe ruptured as a result of perforation from rust The perforation had likely existed during the baseline test and continually increased leading to a gradual increase in infiltration of ambient air into the exhaust pipe over the course of the baseline test and during the comparison test with additive up until the time that the rupture was discovered
69. ll factors that cause load on the engine such as vehicle speed acceleration road grade and load Thus the frequency distribution of MAP for a component of a duty cycle is an empirical representation of the real world conditions that affected the engine As shown in Figure 10 there are empirical frequency distributions of MAP for each component of the duty cycle as well as for the weighted combination of all components The latter represents a complete Overall duty cycle Overall Cycle represents the total round trip driven by Truck 6415 B20 run The time spent idling when the truck was being loaded and or unloaded at a NCDOT unit is not included in the Overall cycle Idling time at an intersection is included Highway and Arterial are subsets of the Overall cycle Highway is the highway only driving from Winston Salem to Raleigh Arterial is the arterial only non highway driving from Winston Salem to Raleigh The Overall duty cycle is dominated by highway driving which comprises 85 percent of the total seconds with the remaining 15 percent coming from arterial driving 56 Table 6 Summary of Duty Cycle Time Fraction for MAP Modes Fron oF tie a 6 fas 008 0320 os foo oos fonos ws ooo foo 0 060 Total Distance miles 9747 669 30323 Sum of Fraction of Time for all MAP Modes may not add up to 1 000 because of rounding Crror 57 oo Highway Arterial Ov
70. lly inconclusive The last part of Table 3 shows the ratio of fuel use and emissions for B20 with Fuel additive versus baseline B20 On average total fuel use and total CO7 emission increased by 5 percent with the use of the Fuel additive The PM emissions decreased by 2 percent The NO emissions increased by 12 percent The results for HC and CO are not conclusive as discussed above Table 7 Total Fuel Use and Emissions based on Overall Representative Real World Duty Cycle for Three B 20 Fueled Combination Trucks with and without Fuel additive Overall Representative Real World Duty Cycle Average Fuel Use Kg 139 194 205 __ 153 201 205 5160 2150 1120 Fuel Use NOx 20 Biodiesel 80 Petroleum Diesel 20 Biodiesel 80 Petroleum Diesel with Fuel additive The concentrations of HC and CO are lower for some MAP ranges than the detection limits of PEMS as documented in Tables A 1 to A 6 of the Appendix 1 The estimated mass of fuel use and emissions for Truck 5715 are biased low because of perforation of the exhaust pipe by rust that was detected partway through the test with Fuel additive The numbers reported for this truck are for informational purposes only They are not included in the average results shown in the last 6l column Because the amount of dilution likely differed when comparing the baseline test and the test with additive ratios for with additive to without additive
71. malized Manifold Absolute Pressure 80 BB20 BB20FA S 60 a oF N 40 Q 20 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 Normalized Manifold Absolute Pressure 150 GB20 BB20FA a a en 100 E Z 50 0 0 8 0 9 l 0 1 0 2 0 3 0 4 0 5 0 6 0 7 Normalized Manifold Absolute Pressure B20 20 Biodiesel 80 Petroleum Diesel B20FA B20 with Fuel additive Figure A 5 Average Fuel Use and Emission Rates of Each Pollutant on a Per Time Basis for Engine Based Modes for Combination Truck 215 6667 93 HC mg sec Opacity mg sec CO mg sec 0 12 G B20 W B20FA 0 08 0 04 0 00 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 l Normalized Manifold Absolute Pressure mB20 m B20FA 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 Normalized Manifold Absolute Pressure 40 GB20 m B20FA 30 20 10 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 Normalized Manifold Absolute Pressure 94 mB20 WB20FA T eS Cy 4 Z 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 Normalized Manifold Absolute Pressure 0 06 B20 mB20FA 0 04 5 5 Z 0 02 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 Normalized Manifold Absolute Pressure B20 20 Biodiesel 80 Petroleum Diesel B20FA B20 with Fuel additive Figure A 6 Emission Rates of Each Pollutant on a Per Fuel Basis for Engine Based Modes for Combination Truck 215 6667 95 HC g gallon CO g gallon 0 1 0 1 0 2 0 2 0 3 0 4 0 5 0 6
72. n the high cost of such tests it 1s important that the data be quality assured and results be available shortly after the test is completed to ensure a decision can be made about the next optimization step Finally a standardized ready to use computer based quality assurance method allows researchers to focus on data analysis and interpretation and also to train new researchers with respect to QA procedure 3 Data Quality And Error Identification Data quality issues and errors typically observed in second by second data from PEMS are identical and described in the following paragraphs Data Sampling Frequency When measuring emissions for real world driving it 1s important to have data at a resolution that can capture driving transients such as acceleration on a ramp effects of stop and go traffic and change in road grade These influence engine power demand and in turn emissions from the vehicle With current technology the instruments deployed for in use measurements PEMS GPS and OBD are capable of recording at approximately 1 Hz A first step in developing a combined database from PEMS GPS and OBD data is to make sure all data are converted to the same reporting frequency A description of PEMS GPS and OBD data reporting frequency is given here PEMS and GPS Both PEMS and GPS generally provide second by second 1 Hz data output with very few instances of missing seconds On an average both PEMS and GPS reported 1 2
73. ne scanner data GPS data and calculated quantities including the emission rate in units of mass per time The following parameters are typically available on screen on a second by second basis engine rpm MAP concentrations of the measured pollutants exhaust flow fuel consumption and mass flow rates of the measured pollutants 3 1 6 Validation and Calibration The Montana System gas analyzer utilizes a two point calibration system that includes zero calibration and span calibration Zero calibration is performed using ambient air at frequent intervals every 5 15 minutes at power up every 30 minutes once fully warmed up Although zero air stored in bottles or generated using an external zero air generator can be used it is believed that the ambient air pollutant levels are negligible compared to those found in undiluted exhaust therefore 39 ambient air is viewed as sufficient for most conditions For zero calibration purposes it is assumed that ambient air contains 20 9 vol oxygen and negligible NO HC or CO CO levels in ambient air are approximately 300 400 ppm which are negligible compared to the typical levels of CO in exhaust gases Span calibration is performed using a BAR 90 low concentration calibration gas mixture which has a known gas composition The calibration gas includes a mixture of known concentrations of CO2 CO NO and hydrocarbons with the balance being N2 Span gas calibration is re
74. need to convert data to a 1 Hz basis in order to combine with PEMS and GPS data Data Synchronization After the data from PEMS GPS and OBD is converted to 1 Hz it needs to be synchronized or time aligned for calculation of mass based emission rates fuel use rate and VSP modal results Misaligned timestamps can be caused because 1 instruments each have their own independent internal clocks and 2 a physical process 1s varying gas travel time in vehicle exhaust pipe and measurement system gas sampling line Time misalignment due to varying gas travel time in exhaust and sampling line has been discussed by Weilenmann et al 2003 Konstantas and Stamatelos 2004 Hawley et al 2004 Ropkins et al 2007 Frey et al 2008 A brief discussion on other errors related to second by second s b s or 1Hz data from PEMS OBD and engine sensor array 1s given below For a detailed discussion of these errors refer Frey et al 2008 Frey and Kim 2005 Frey et al 2003 Frey et al 2002b Engine Speed RPM Error The RPM lower limit and upper limit for a vehicle model or type is generally known and is used to evaluate the validity of RPM measurement A RPM reading below the idling RPM can be used as lower limit and the upper limit can be estimated from manufacturer specifications or from the maximum on the dashboard tachometer if available For 4 stroke gasoline and diesel on road vehicles a valid RPM range of 500 rpm to 4000 rpm is typical
75. ning system A temporarily mounted sensor array is used to measure Manifold Absolute Pressure MAP intake air temperature and engine RPM in order to estimate air and fuel use A GPS system measures vehicle position The on board computer synchronizes the incoming emissions engine and GPS data Intake airflow exhaust flow and mass emissions are estimated using a method reported by Vojtisek Lom and Cobb 1997 The gases and pollutants measured include O2 HC CO CO2 NO and PM using the following detection methods e HC CO and CO using non dispersive infrared NDIR The accuracy for CO and CO are excellent The accuracy of the HC measurement depends on type of fuel used 35 e NO measured using electrochemical cell On most vehicles with Tier 2 or older engines NO is comprised of approximately 95 volume percent NO e PM is measured using light scattering with measurement ranging from ambient levels to low double digits opacity The Montana System is designed to measure emissions during the actual use of the vehicle or equipment in its regular daily operation The complete system comes in two weatherproof plastic cases one of which contains the monitoring system itself and the other of which contains sample inlet and exhaust lines tie down straps AC adapter power and data cables various electronic engine sensor connectors and other parts The monitoring system weighs approximately 35 bs The system typically runs off
76. od of time such as during sudden acceleration or deceleration 1s often accompanied by noticeable change having same trend in emissions concentration of CO2 CO and NOx To assist in synchronization a throttle snap is built into the test procedure where the driver floors the gas pedal and thus spikes the RPM and exhaust emissions while the vehicle 1s parked This 1s done at the start and end of a test route or when the vehicle is ready to start driving after a long period of idling such as loading unloading for combination trucks The master parameter from the master database and the slave parameter from the slave database are plotted on a common time axis and sections of the test with high rate of change in the master and or slave parameter are marked out Timestamp of the slave parameter is adjusted until 15 the instance of start of rise or fall in the master parameter coincides with the start of rise or fall of the slave parameter Other sections of test data having similar characteristic steep rise or fall are checked to make sure that the time adjustment is applicable throughout the dataset As a confirmation of the visual technique of synchronization any appropriate pair of master slave time series data arrays when correctly synchronized will return a high Pearson Coefficient of Correlation PCC Kubelt and Bonnel 2007 The PCC denoted as r for a master slave data array pair is given as dx Hy E E Eg
77. ode 0 1 and increased the emissions rate by 16 6 for MAP mode 0 3 For a 10 second shift to OBD data NO emissions rate 26 3 and 49 3 for MAP modes 0 2 and 0 3 and decreased by 24 8 38 and 29 1 for MAP modes 0 8 0 9 and 1 A shift of 10 seconds to GPS data increased NO emissions rate by 32 1 and 26 6 for VSP modes 1 and 4 26 EOBD no offset OBD 5 sec offset OBD 10sec offset NO mg sec l 2 3 4 5 6 7 8 9 10 11 12 13 14 Vehicle Specific Power Mode Figure 12 Effect of shifting OBD data on VSP based NO Emissions Rate 5 3 Intake Air Temperature In a field test of a 15 liter HDDV combination truck the sensor array connected to the PEMS experienced communication problems that resulted in data having spurious IAT values The updated IAT algorithm was able to recognize a total of 605 seconds of erroneous IAT data out of total test duration of 29254 seconds out of which 560 seconds were correctly interpolated The previous algorithm had marked 1031 seconds as error many of which were false positives and did not use interpolation to save any seconds from being deleted Figure 13 shows a section of the test with error AT values and subsequently interpolated IAT values 27 64 TAT BeforeQA 62 S IAT AfterQA 60 A 58 gee soe f m T s 54 52 IAT C 50 23200 23210 23220 23230 23240 23250 Timestamp sec Figure 13 Test data with error IAT readings corrected by the IAT algorithm 5 4
78. om a QA criteria strength perspective if the RPM value is missing for more than one second it is advised to delete the second rather than impute the missing RPM value On the other hand if a couple of IAT values are missing or error they can be safely imputed using linear interpolation 18 12 00 2 10 00 8 00 E 6 00 e 4 00 g 2 00 0 00 E 2 00 O 4 00 5 6 00 8 00 Y FP 10 00 12 00 10 0 Percent Change in Selected Input 10 Figure 1 Sensitivity of g s Fuel Use to Input Parameters for Heavy Duty Diesel Vehicle 12 00 2 8 00 aB g 4 00 i o 0 00 O 4 00 aB 8 00 lt gt A 12 00 10 0 Percent Change in Selected Input 10 Figure 2 Sensitivity of g s Fuel Use to Input Parameters for Light Duty Gasoline Vehicle 19 5 2 Synchronization For LDGV the recommended pairs for synchronizing PEMS and OBD data are CO RPM and NOx RPM both of which produce prominent and sharp PCC peak Generally PCC values are in the vicinity of 0 3 and 0 6 for CO RPM pair and NO RPM pair respectively For HDDV the recommended pair is CO2 RPM which produces PCC values in the vicinity of 0 5 Other options are O2 RPM and NO RPM which typically produce 0 4 and 0 4 PCC values respectively For any vehicle type when using the same parameter type pairs such as vehicle speed from OBD and vehicle speed from GPS or NO from gas analyzer 1 and 2 PCC values are typically close
79. on using a NDIR analyzer HC concentration is measured by a FID analyzer NO concentration 1s measured by a Chemiluminescence Detector CLD exhaust flow measured by Pitot flow meter ECU data monitored using Dearborn Instruments module and has standard inputs to measure GPS signals longitude latitude altitude and velocity exhaust temperature exhaust pressure ambient temperature atmospheric pressure and ambient humidity Akard et al 2005 Horiba 2010 The OEM 2100 Montana and OEM 2100AX Axion from CATI have two identical parallel operation 5 gas analyzers to measure exhaust gas concentrations The gas analyzers measure hydrocarbons HCs carbon monoxide CO and carbon dioxide CO2 using nondispersive infrared NDIR nitric oxide NO and oxygen O2 using electrochemical sensors and particulate matter PM concentrations using light scattering Frey and Kim 2009 CATI 2007 CATI 2008 Vojtisek Lom and Allsop 2001 This dual analyzer in parallel arrangement is provided because the gas analyzers need to zero periodically during a test When a gas analyzer is zeroing it is using ambient air as a reference to recalibrate itself Both benches zero at pre set intervals such as every 10 minutes but never together When both analyzers are measuring the PEMS computes the average of readings from both analyzers to estimate mass emission rate and fuel use rate When one analyzer is zeroing the PEMS uses readings from the non ze
80. ons rate A technique based on visual comparison plus use of Pearson s Coefficient of Correlation is demonstrated to be effective in synchronizing independent data streams The second part addresses the question of how to assess claims about effect of alternative fuels for example B20 and fuel additives on vehicle fuel use and emissions rate A solution using quality assurance procedures field measurement techniques and driving cycle bases modal fuel use and emissions rates is presented Three diesel trucks are tested with baseline fuel and subsequently alternative fuel It is shown that the use of the fuel additive under test did not produce significantly different results compared to the baseline Methods for Quality Assurance of Portable Emissions Measurement System Data and Methods for Field Comparison of Alternative Fuels by Gurdas Singh Sandhu A thesis submitted to the Graduate Faculty of North Carolina State University in partial fulfillment of the requirements for the degree of Master of Science Environmental Engineering Raleigh North Carolina 2011 APPROVED BY Dr Joseph DeCarolis Dr John Baugh Dr Christopher Frey Chair of Advisory Committee DEDICATION To my parents Jaswant Kaur Sandhu and Harnek Singh Sandhu who taught me to be curious and To the chairperson of my advisory committee Dr Christopher Frey who encouraged me to learn li BIOGRAPHY Gurdas Singh Sandhu was born in the late s
81. ont view b the portable unit on a passenger seat rear view c routing exhaust hoses from PEMS through the window Figure 21 Installation of the Portable Emissions Measurement system PEMS Power Cable GPS Receiver and Exhaust Sampling Lines in a Truck 215 5715 a accessing power from the vehicle cigarette lighter b GPS receiver on the roof of the test vehicle c sampling exhaust gases using a probe secured with a hose clamp r Figure 22 Installation of the Portable Emissions Measurement System PEMS Sampling Hoses and Engine Sensors in a Truck 215 5715 a routing sampling hoses through the window secured with ties front view b Installation of MAP sensor c Installation of RPM sensor 4 After completing all installation steps the instrument needs to warm up for approximately 45 minutes This time period is recommended in order to ensure consistency of measurements made by the instrument CATI 2003 During testing periodic checks of the system status are recommended For example the security of all connections with the vehicle should be evaluated This can be done by determining whether engine data is updated on the instrument display in an appropriate manner whether the gas concentrations are reasonable and whether the instrument is receiving power If any of the data relating to gas concentration and or engine parameter is frozen or missing then it will be necessary to reboot the
82. ontinuously update which can be corrected by rebooting the gas analyzer However these problems did not occur during the testing Decommissioning occurs after the end of the test period During decommissioning the NCSU research assistant discontinued data collection copied data that have been collected powered down the Montana system and removed the exhaust sample lines power cable data cable and GPS receiver and cable The use of the PEMS did not involve any modification to the vehicle 44 PRE INSTALLATION Check if filter in gas sampling line needs to be replaced Install sampling Gas and PM probes into vehicle exhaust and connect other end with sampling bowls to Montana Install exhaust 3 lines and zeroing 1 line and connect to Montana Install Montana GPS unit on roof Connect to Montana data port and DC power port on 3 way splitter Install Garmin GPS 3 antenna on roof and connect to respective handset Install RPM sensor and connect wire to sensor array box Install Temperature sensor and connect wire to sensor array box Install MAP sensor and connect wire to Montana Connect sensor array box to Montana Connect 3 way power splitter to vehicle DC power port cigarette lighter port Connect Montana power to 3 way splitter Check MAP RPM Temperature values Check Gas concentrations and PM level Check GPS fix might not work indoors TEST DAY Warm up Montana for at least 45 minutes Turn on Garmin GPS 3
83. phail N M Colyar J D 2003 On Road Measurement of Vehicle Tailpipe Emissions Using a Portable Instrument J Air amp Waste Manage Assoc 53 992 1002 Frey H C Kim K 2005 Operational Evaluation of Emissions and Fuel Use of B20 Versus Diesel Fueled Dump Trucks Research Project No 2004 18 FHWA NC 2005 07 Prepared by the Department of Civil Construction and Environmental Engineering North Carolina State University Raleigh NC for North Carolina Department of Transportation Raleigh NC 2005 Frey H C Rasdorf W Kim K Pang S Lewis P Abolhassani S 2008 Real world duty cycles and utilization for construction equipment in North Carolina Research Project No FHWA NC 2006 55 Prepared by North Carolina State University for North Carolina Department of Transportation Research and Development Group Raleigh NC 79 Frey H C Kim K 2009 In use measurement of the activity fuel use and emissions of eight cement mixer trucks operated on each of petroleum diesel and soy based B20 biodiesel Transportation Research Part D 14 2009 585 592 Garmin 2009 GPSMAP 76CSx owner s manual May 2009 Rev E http static garmincdn com pumac GPSMAP76CSx_OwnersManual pdf Accessed December 20 2010 Gierczak C A Korniski T J Wallington T J Butler J W 2006 Laboratory Evaluation of the SEMTECH G Portable Emissions Measurement System PEMS For Gasoline Fueled Vehicles SAE Technical Paper No
84. ported below 25 ppm then the reading is marked as an error and not used to calculate mass emission rates Inter Analyzer Discrepancy IAD Error In PEMS having two identical gas analyzers the absolute value of the difference in the instantaneous readings of the two analyzers for a given pollutant is the IAD The maximum acceptable difference MAD between the readings of the two gas analyzers for a given pollutant 1s equal to twice the precision of the pollutant sensor This is because the acceptable measured concentration values from the two sensors are maximally separated when one sensor is reading pollutant concentration as True Value Precision and other sensor is reading True Value Precision A difference in the reading of the two gas analyzers greater than the MAD is reported as an IAD error When IAD lt MAD the average of the reading of two gas analyzers is used When IAD gt MAD and both analyzers report values greater than detection limit the number of consecutive seconds of measurement with IAD gt MAD 1s counted If count is less than or equal to 15 then average of two analyzers is used else the analyzer data is deleted for the said seconds When IAD gt MAD and one analyzer is above detection limit and other 1s below detection limit the value of the analyzer above detection limit is used When IAD gt MAD and both analyzers are below detection limit then an average is taken of both analyzers Air Leakage Any le
85. r the three trucks that were tested on B20 with or without the Fuel additive There are two routes in the Figure 9 In the morning all three trucks start at the NCDOT Division 5 maintenance yard which is marked as Raleigh with a red dot in Figure 9 This yard is located at Blue Ridge and Trinity Roads in Raleigh NC The duty cycles of these vehicles typically included travel to NCDOT locations in Raleigh North Wilkesboro Winston Salem Castle Hayne Burgaw and Clinton all of which are in North Carolina For truck 215 5715 B20 and B20FA runs the travel loop was Raleigh Castle Hayne Burgaw Clinton Raleigh For trucks 215 6415 and 215 6667 B20 and B20FA runs the travel loop was Raleigh North Wilkesboro Winston Salem Raleigh Re a ge Jaw Ze iF i XE us North Wilkesboro E k a North Garolina tf kaS Eve alt 4442 mi Figure 26 Map of the Geographic Area of In Use Field Measurements 55 4 4 Real World Duty Cycle The trucks were tested during real world duty cycles In order to make comparisons between tests for each truck a typical duty cycle was selected The duty cycle and its components are quantified in terms of frequency distributions of manifold absolute pressure MAP of the engine MAP has been found in previous work to be highly correlated with fuel use and emissions of diesel engines and is measured by the sensor array used with the Montana system MAP takes into account a
86. re he shares those experiences that deeply influence him Gurdas enjoys driving especially if the road is unknown and there is music to accompany Gurdas finds humor to be healing and love to be eternal In 2010 he met and subsequently got engaged to a wonderful woman He hopes to soon start a family and continue to learn ever more from new responsibilities il ACKNOWLEDGEMENTS At the very outset I express my sincere thanks to Dr Christopher Frey for his kind guidance and brilliant intellectual leadership and for patiently bearing with my many shortcomings I then thank my committee members Dr John Baugh for insights on algorithmic efficiency and the occasional haiku and Dr Joseph DeCarolis for sharing his expertise on energy modeling and motivating me on days when solutions were not coming forth Next I remember friends and colleagues who with their love and understanding made this journey satisfying and pleasurable It may not be possible to name all of them here but you know who you are I then pay gratitude to my parents who even from thousands of miles away continue to be my pillars of love and truth What would I be without you Finally I acknowledge the assistance of clerical staff lab managers and other faculty at North Carolina State University who in their own unique ways helped me 1V TABLE OF CONTENTS LST TOFTABLES roenan toe eee Vil ME OE EG OR sachin dscns Sacra ee rc cece E A A T IX Chapter 1 Methods for
87. re the test and then refueled to same level after the test Calculated from Modal Model fuel g sec and total seconds for each mode gt Actual Measured Actual 100 Difference of Odometer reading at start and end of test 7 Calculated from sec by sec speed data from GPS 80 00 70 00 4 60 00 4 50 00 Measured Fuel Use gal 40 00 30 00 q m T l T 30 00 40 00 50 00 60 00 70 00 80 00 Actual Fuel Use gal Figure A 7 Scatter plot of Actual Fuel Use vs Measured Fuel Use 100 Table A 8 Rate of Loss of Data Because of Data Quality Errors Fuel Amount of Data Lost for Specific Type of eh Error QA ere al sec CE RA Ra KR Hk aad KAKA KA dl bt sl Bal al Sl a dhl sc i LA NB lee wl Baas el i Bl ac al a iad k Overall Total Seconds 155 991 eee a 15 690 1 031 ae ae ee 138 950 Percentage of raw data eel ey 10 06 Mao a 10 92 89 08 B20 20 Biodiesel 80 Petroleum Diesel B20FA B20 with Fuel additive gt Total Raw Data Definition of Errors 1 Missing Manifold Absolute Pressure MAP 2 Unusual Engine Speed engine RPM 3 Analyzer Freezing 4 Inter analyzer Discrepancy IAD 5 Air Leakage 6 7 8 Unusual Intake Air Temperature IAT Negative Emission Value Invalid Data as Flagged by the PEMS software i Average Error Rate Quality Assured Data 101 5715 B20
88. roing analyzer to calculate mass emission rate and fuel use rate Depending on vehicle type and test configuration measurements also involve a an On Board Diagnostic OBD scan tool or engine sensor array and b a Global Position System GPS unit The OBD scan tool is a hardware and software system that connects to the OBD II port found on most LDGVs manufactured since 1996 on one end and a recording computer on the other end The OBD scan tool is used to read Electronic Control Unit ECU data such as engine RPM Manifold Absolute Pressure MAP Intake Air Temperature IAT mass air flow rate fuel use rate and vehicle speed When testing a vehicle that does not have an OBD port or a compatible OBD scan tool is not available an engine sensor array is used to record RPM and IAT and a pressure sensor is used to measure MAP The sensor array box receives analog signals from RPM and IAT sensors and converts them to digital signals that can be read by the PEMS The MAP pressure sensor is directly connection to the PEMS Both Montana and Axion have a GPS sensor connection and record GPS X and Y latitude and longitude coordinates Additionally stand alone GPS units can be used to get more accurate X and Y coordinates and measure additional parameters such as altitude based on barometric pressure which can be used to derive road grade For each second of data recorded there are at least 20 parameters including PEMS Timestamp concentrat
89. rs based on the real world typical duty cycles of Figure 10 The distance based rates are shown for informational purposes These rates are biased low because of dilution of the exhaust with ambient air as a result of perforation of the exhaust pipe by rust Therefore ratios of the distance based rates with additive to without additive are not shown 62 e Modal Emission Factors Time based modal fuel use and emission factors for B20 without and with Fuel additive are given in Figure A 1 of Appendix A Fuel based emission factors for B20 without and with Fuel additive are given in Figure A 2 of Appendix A The average exhaust gas concentrations for NO HC CO COs and O2 for each MAP based mode are given in Tables A 1 and A 2 of Appendix A for B20 without and with Fuel additive respectively Vehicle 215 6415 e Table 9 Time based average fuel use and emission factors based on the real world typical duty cycles of Figure 10 e Table 10 Fuel based emission factors based on the real world typical duty cycles of Figure 10 e Table 11 Travel distance based average fuel use and emission factors based on the real world typical duty cycles of Figure 10 e Modal Emission Factors Time based modal fuel use and emission factors for fuel without and with additive are given in Figure A 3 of Appendix A Fuel based emission factors for the B20 without and with Fuel additive are given in Figure A 4 of Appendix A The average exhaust gas concen
90. s sparse and since the basic data is missing the QA procedure deletes the said second of data Loss of Power to Instrument A loss of power to the instrument resulted in a complete loss of data collection during the time period when power was not available However the system saves data up to the point at which the power loss occurs A typical cause of power loss for manual transmission vehicles is stalling of the engine due to a problem shifting Such problems typically occur when going from idle into first gear or for the lower gears After a loss of power the instrument needs to be rebooted which takes approximately five to ten minutes During the power loss and rebooting no data can be collected NCSU has developed a series of Macros in Visual Basic in conjunction with MS Excel Raw data from the Montana system is processed via these macros to identify data quality problems Where possible such problems are corrected If correction is not possible then the errant data are omitted from the final database used for analysis 4 0 Results The results include the field data collection schedule vehicle characteristics and test conditions quality assurance and detailed characterization of each vehicle 49 4 1 Scheduling of Field Data Collection Field data collection occurred during a period from June 9 2009 to August 11 2009 as summarized in Table 1 The baseline tests with B20 for each vehicle were conducted in quick suc
91. s to lead to substantial differences in results when making comparisons of fuel consumption with versus without the additive Similarly we compared the distribution of time for vehicle speed as given in Figures A 11 to A 13 of the Appendix For Truck 5715 the average speed for each test ranged from 39 to 42 mph with a standard deviation of 27 to 28 mph For Truck 6415 the average speed was 43 to 45 mph with a standard deviation of 28 to 30 mph For Truck 6667 the average speed was 39 to 44 mph with a standard deviation of 27 to 31 mph Thus the operating conditions of the tests with additive were very similar to the operating conditions of the test without additive 75 5 0 Conclusions The purpose of this study is to compare fuel use and emissions with an additive to without an additive for B20 fueled tractor trailer trucks Data were collected for real world duty cycles for each of three trucks Each truck operated on the same route for a day of testing without additive and a later day of testing with additive The activity patterns of each test for a given truck were very similar based on comparisons of frequency distributions of engine load quantified based on MAP and of vehicle speed However in order to reduce the effect of inherent variability in operating conditions when comparing the two tests comparisons were made by first creating modal rates of fuel use and emissions with respect to ten ranges of increasing MAP and then w
92. sis However the dilution affects the mass per time and mass per distance emission rates since these are calculated based on the air to fuel ratio estimated from the exhaust composition The dilution of the exhaust gas with ambient air leads to an increase in the apparent air to fuel ratio This in turn affects the calculation of g sec and g mile emission rates Thus although these rates are reported they are footnoted as not being reliable because the amount of dilution is not known Furthermore the estimated fuel flow rate g sec for this truck calculated from PEMS data is 52 low because of the dilution In sum the reliable results for truck 5715 are for the g gallon results for NO CO and HC These numbers are not affected by dilution of the exhaust Table 4 Data Collection Field Log Chassis iD 215 5715 License PM 1645 Type Combination Trailer Length Year 1999 Make International or i 6 Model 2574 6X4 Fofas SAxkeswihi SS Test Condition B20 B20 and Fuel additive BS june 10 2009 Test Date Truck Truck Weight measured measured Cargo Tail pipe rupture estimated to have occurred about 30 miles prior to NCDOT unit at Castle Hayne travelling from Raleigh The rupture was fixed at the Castle Hayne stop and test data after that point which is about 50 of total test data for this run can be considered to be free of tail pipe dilution due to rupture GVW 500 hp 2100 of Axles 5 Axles w
93. stimated based on the typical duty cycle The weighted average rate is multiplied by the total time spent in the duty cycle to arrive at an estimate of total fuel use and emissions for the duty cycle For key components of the duty cycle an average mileage based fuel use and emission rate is estimated by dividing the total fuel use and emissions for the component or total cycle by the corresponding distance traveled for the component or the total cycle respectively In addition fuel based emission factors can easily be estimated These emission factors are on a gram of pollutant emitted per gallon of fuel consumed basis The fuel based emission factors are estimated based on a carbon mass balance for the fuel and the pollutants The carbon in the fuel is emitted as CO2 CO hydrocarbons and particulate matter Typically more than 99 percent of the carbon in the fuel is emitted as CO2 The amount of CO and hydrocarbons emitted by diesel engines tends to be small because these engines operate with a fuel lean mix of air and fuel Hence there is sufficient oxygen available in the air and fuel mixture to promote a high proportion of complete oxidation of carbon in the fuel to CO2 The amount of carbon in particulate matter 1s very small and does not significantly affect the carbon balance for purposes of estimating fuel based emission factors For each vehicle detailed results are provided for emission factors for each component of the duty cyc
94. sults do not imply any supportable conclusion for either beneficial or adverse effects of the tested additive 11 REFERENCES Abolhasani S Frey H C Kim K Rasdorf W Lewis P Pang S 2008 Real World In Use Activity Fuel Use and Emissions for Nonroad Construction Vehicles A Case Study for Excavators J Air amp Waste Manage Assoc 58 1033 1046 Ajtay D Weilenmann M Soltic P 2005 Towards accurate instantaneous emission models Atmospheric Environment Volume 39 Issue 13 p 2443 2449 Akard M Nakamura H Aoki S Kihara N Adachi M 2005 Performance Results and Design Considerations for a New In Use Testing Instrument SAE Technical Paper No 2005 01 3606 Society of Automotive Engineers Warrendale PA Baldauf R W Fulper C Gabele P Tierney G Somers J Warilal J 2001 The Kansas City Light Duty Vehicle Emission Study US EPA report http www epa gov ttn chief conference ei13 mobile baldauf_pres pdf Accessed August 5 2010 CATI 2007 OEM 2100 Montana System Operation Manual Clean Air Technologies International Inc Buffalo New York USA Version 2 1 October 2007 CATI 2008 OEM 2100AX Axion User s Manual Clean Air Technologies International Inc Buffalo New York USA Version 2 0 2008 Collins J F Shepherd P Durbin T D Lents J Norbeck J Barth M 2007 Measurements of In Use Emissions from Modern Vehicles Using an On Board Measurement
95. system and the actual amount of fuel consumed as determined by the amount of fuel that was put into the fuel tank after each day of testing A summary is given of the total sample size of the data sets for each tested vehicle and fuel There were approximately 7 500 to 24 000 seconds of data for most of the tests A summary is given of the overall rate of loss of data because of data quality issues On average 10 of the data were lost or excluded because of data quality issues RPM error occurred for the test of truck 215 6415 with B20 in the rainy day Except this RPM error on average only 6 0 of data were lost or excluded because of data quality issues Without considering this RPM error the most frequent causes of data loss were RPM error and analyzer freezing 84 15 BB20 WB20FA a a 2 2 10 F D Sa 5 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 Normalized Manifold Absolute Pressure 50 BB20 BB20FA 40 Q A A a BO N 5 20 10 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 Normalized Manifold Absolute Pressure 500 GB20 BB20FA w 400 Q 2 oN 300 Z 200 100 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 l Normalized Manifold Absolute Pressure B20 20 Biodiesel 80 Petroleum Diesel B20FA B20 with Fuel additive Figure A 1 Average Fuel Use and Emission Rates of Each Pollutant on a Per Time Basis for Engine Based Modes for Combination Truck 215 5715 85 T P t on E 5 a O
96. tep 0 25 mile VSP based results for NO and HC emissions rate is shown in Figure 16 and Figure 17 In general road grades calculated using the RGstep scheme provide numbers that are within known extremes An RGstep value within a reasonable range 0 05 mile to 0 25 mile does not significantly affect the VSP based results This may change if the test route involves larger percentage of hills in which case a smaller RGstep value may produce more accurate results 30 BRGStep 0 1 mile RGStep_0 05 mile amp RGStep_ 0 25 mile NO mg sec l 2 3 4 5 6 7 8 9 10 11 12 13 14 Vehicle Specific Power Mode Figure 16 Effect of RGstep criteria on VSP based NO Emissions Rate BRGStep 0 1 mile 5 RGStep_ 0 05 mile EE RGStep_ 0 25 mile HC mg sec l 2 3 4 5 6 7 8 9 10 11 12 13 14 Vehicle Specific Power Mode Figure 17 Effect of RGstep criteria on VSP based HC Emissions Rate 31 6 Conclusions Second by second emissions rate calculated using the speed density method 1s highly sensitive to MAP and RPM values The use of proper installation calibration and verification of these two parameters is important to the accuracy of final mass per time fuel use and emissions rate Synchronization of multiple data stream from independently functioning instruments 1s another important step in quality assurance Pearson s Correlation Coefficient can be used to approximate the correct synchronization setting However it is not always a
97. than three seconds This problem occurs only in situations where the engine scanner became physically disconnected from the data logging computer This type of error is rare and did not occur during these tests Gas Analyzer Errors The Montana system has two gas analyzers which are referred to as benches Most of the time both benches are in use Occasionally one bench is taken off line for zeroing Therefore most of the time the emissions measurements from each of the two benches can be compared to evaluate the consistency between the two If both benches are producing consistent measurements then the measurements from both are averaged to arrive at a single estimate on a second by second basis of the emissions of each pollutant When the relative error in the emissions measurement between both benches is within five percent and if no other errors are detected then an average value is calculated based upon both of the benches However if the relative error exceeds five percent then further assessment of data quality 1s indicated A discrepancy in measurements might be due to any of the following a a leakage in the sample line leading to one bench b overheating of one of the benches or c problems with the sampling pump for one of the benches leading to inadequate flow If one of these problems is identified for one of the benches then only data obtained from the other bench was used for emissions estimation Wh
98. to 0 9 to 1 0 The PEMS and OBD databases for a real world in use emissions testing of a Honda Accord 1997 model 2 2 liter engine 130HP is conducted using NO from PEMS as master parameter and RPM from OBD as slave parameter Figure 3 and Figure 4 show before and after synchronization time series plots Figure 5 and Figure 6 show PCC plots before and after synchronization Here RPM timestamp is adjusted by 5 seconds for the two datasets to be synchronized Correct synchronization results in PCC peak occurring at 0 mark on the x axis which signifies the slave timestamp need not be further adjusted 600 3700 NOx PEMS 500 RPM_ OBD 3200 _ 400 2700 5 300 2200 z x z S 200 1700 1200 100 S 700 53270 53280 53290 53300 53310 53320 53330 53340 53350 Timestamp sec Figure 3 NO from PEMS and RPM from OBD before synchronization 20 600 3700 NOx PEMS 500 RPM OBD 3200 _ 400 2700 2 4 300 2200 z 200 1700 100 1200 0 4 700 53270 53280 53290 53300 53310 53320 53330 53340 53350 Timestamp sec Figure 4 NO from PEMS and RPM from OBD after synchronization nae Correlation Coefficient Pearson s Correlation Coefficient Slave Parameter Timestamp Shift s Figure 5 Correlation between NO PEMS and RPM OBD before synchronization 24 RPM 0 7 ne Correlation Coefficient 0 6 0 5 0 4 0 3 9 RETS CSA D2 OY 0 2 aG OCE DOO
99. trations for NO HC CO CO and O2 for each MAP based mode are given in Tables A 3 and A 4 of Appendix A for B20 without and with Fuel additive respectively Vehicle 215 6667 e Table 12 Time based average fuel use and emission factors based on the real world typical duty cycles of Figure 10 e Table 13 Fuel based emission factors based on the real world typical duty cycles of Figure 10 63 e Table 14 Travel distance based average fuel use and emission factors based on the real world typical duty cycles of Figure 10 e Modal Emission Factors Time based modal fuel use and emission factors for fuel without and with additive are given in Figure A 5 of Appendix A Fuel based emission factors for the B20 without and with Fuel additive are given in Figure A 6 of Appendix A The average exhaust gas concentrations for NO HC CO CO and O for each MAP based mode are given in Tables A 5 and A 6 of Appendix A for B20 without and with Fuel additive respectively Table 8 Time Based Average Fuel Use and Emission Factors based on Representative Real World Duty Cycles for the Combination Truck 215 5715 Arterial 7 69 270 HC mg CO gt g 15 o i ae B20 15 15 15 23 3 2D 0 50 Fuel Use g s 8 48 PM mg s 20 Biodiesel 80 Petroleum Diesel 20 Biodiesel 80 Petroleum Diesel with Fuel additive The concentrations of HC and CO are lower for some MAP ranges than the detection limits of PEMS as docum
100. ty Analysis Knowledge of sensitivity of outputs to change in inputs is important to the QA method because input parameters that can change rapidly and have significant impact on results will need to undergo tighter QA criteria while parameters that have less significant impact on end result can do with more relaxed QA criteria The QA criteria strength is important because tighter criteria will remove an error or missing second of measurement more often than relaxed criteria Sensitivity analysis was conducted for a 15 liter heavy duty diesel HDD combination truck and a 2 2 liter light duty gasoline LDG car to measure the percent change in fuel use g s pollutant mass emissions rate for NOx HC CO CO2 Oz each g s intake air mass flow rate g s and dry exhaust mass flow rate g s for a 10 change in exhaust gas concentrations NOx HC CO CO2 O2 and engine data RPM IAT MAP A typical second of data base case was picked from actual real world in use emissions measurement database and the parameter under test was varied by 10 with subsequent recalculation of output results Table 1 gives the base case and typical range of values for the input parameters Since the resulting fuel use and mass emission rates for each second are based on empirical calculations that use input data for only that second the sensitivity of outputs to inputs calculated using one second can be safely assumed to be the typical sensitivity Table 1 S
101. ue is missing for 4 or more seconds and MAP 1s being used to calculate intake molecular flow rate as per speed density method then the timestamps are marked as error and removed from the error free QAed database Zeroing Error For the period when a gas analyzer is zeroing plus the 10 seconds immediately before and after the zeroing period the mass emission rates are calculated using emission concentration readings from the non zeroing gas analyzer This is done because for the bench that is zeroing the preceding and following seconds contain a mix of exhaust sample and ambient air 10 Gas Analyzer Freezing Error A gas analyzer is marked with freezing error if the reported concentration measurement for each of the measured pollutants from the gas analyzer is same over two consecutive seconds but one or more of the engine parameters RPM IAT MAP for the two seconds Negative Concentration Error The PEMS can sometimes report negative pollutant concentrations This usually happens when the real concentration of the pollutant is low and not statistically significantly different from zero Sensor precision is used to infer what negative values will be considered as zero For example the precision of the CATI NO sensor is 25 ppm Since the lowest feasible value for concentration is 0 ppm the NO sensor may read that as 25 ppm If NO for any gas analyzer is reported between 0 and 25 ppm it is replaced with a value of 0 ppm If NO is re
102. ve Real World Dut Highwa Arterial 10 6 20 Biodiesel 80 Petroleum Diesel 20 Biodiesel 80 Petroleum Diesel with Fuel additive The concentrations of HC and CO are lower for some MAP ranges than the detection limits of PEMS as documented in Tables A 3 and A 4 B20 32 9 1 66 11 6 118 B20FA 35 8 1 65 1 09 1 18 1 03 1 20 1 09 B20FA B20 09 gt OO N Table 12 Fuel Based Emission Factors based on Representative Real World Duty Cycles for the Combination Truck 215 6415 Fuel Representative Real World Dut Highwa Arterial Overall 1 08 0 95 1 10 0 91 20 Biodiesel 80 Petroleum Diesel P 20 Biodiesel 80 Petroleum Diesel with Fuel additive The concentrations of HC and CO are lower for some MAP ranges than the B20FA B20 0 94 1 09 _ Highway _ Overall 31 B20 3 12 0 50 36 B20FA 3 13 0 46 108 0 95 094 aor 109 091 090 detection limits of PEMS as documented in Tables A 3 and A 4 68 Table 13 Travel Distance Based Average Fuel Use and Emission Factors based on Representative Real World Duty Cycles for the Combination Truck 215 6415 NOx g mile 5 5 100 60 B20 _ PM g mile __ _0 09__ 0 16 010 B20FA PM e mile __ _0 09_ 0 16 010 B20FA B20 0 99 0 99 0 99 20 Biodiesel 80 Petroleum Diesel P 20 Biodiesel 80 Petroleum Diesel with Fuel additive The concentrations of HC and CO are low
103. wever the PM emission rate is calculated in a different manner from those of the gases and these results are affected by dilution of the exhaust with ambient air Therefore ratios are not shown for PM 65 Table 10 Travel Distance Based Average Fuel Use and Emission Factors based on Representative Real World Duty Cycles for the Combination Truck 215 57 15 Fuel Representative Real World Dut Overall 4 4 459 Fuel g mile B20 426 747 16 0 0 92 0 92 1420 0 03 69 Fuel g mile 469 817 505 17 0 0 97 2 16 CO g mi 747 B20FA 7 1560 0 04 20 Biodiesel 80 Petroleum Diesel 20 Biodiesel 80 Petroleum Diesel with Fuel additive The concentrations of HC and CO are lower for some MAP ranges than the detection limits of PEMS as documented in Tables A 1 and A 2 1 The fuel based emission factors for NO CO and HC are not affected by the diluation of exhaust with ambient air because of perforation of the exhaust pipe with rust Hence ratios of these emission rates with additive to those without additive are shown However the PM emission rate is calculated in a different manner from those of the gases and these results are affected by dilution of the exhaust with ambient air Therefore ratios are not shown for PM ON 6 Table 11 Time Based Average Fuel Use and Emission Factors based on Representative Real World Duty Cycles for the Combination Truck 215 6415 Fuel Representati

ABSTRACT SANDHU, GURDAS SINGH. Methods for

Contents

Download Pdf Manuals

Related Search

Related Contents