UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Particulate matter emissions from a Cummins ISX 400 engine operating on high-pressure direct injection… Baribeau, Anne-Marie 2001

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2001-0312.pdf [ 10.25MB ]
Metadata
JSON: 831-1.0080964.json
JSON-LD: 831-1.0080964-ld.json
RDF/XML (Pretty): 831-1.0080964-rdf.xml
RDF/JSON: 831-1.0080964-rdf.json
Turtle: 831-1.0080964-turtle.txt
N-Triples: 831-1.0080964-rdf-ntriples.txt
Original Record: 831-1.0080964-source.json
Full Text
831-1.0080964-fulltext.txt
Citation
831-1.0080964.ris

Full Text

PARTICULATE MATTER EMISSIONS FROM A CUMMINS ISX 400 ENGINE OPERATING ON HIGH-PRESSURE DIRECT INJECTION OF NATURAL GAS WITH DIESEL PILOT by A N N E - M A R I E B A R I B E A U B . Sc . (Hons.) . Q u e e n ' s Universi ty , K i n g s t o n , Canada, 1997 A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F A P P L I E D S C I E N C E i n T H E F A C U L T Y O F G R A D U A T E S T U D I E S Department o f M e c h a n i c a l Engineering W e accept this thesis as conforming to the required standard T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A June 2001 © A n n e - M a r i e Baribeau, 2001 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of British Columbia Vancouver, Canada Department DE-6 (2/88) Abstract High-pressure direct injection (HPDI) of natural gas coupled with diesel pilot injection is a promising method by which emissions of particulate matter (PM) and nitrogen oxides (NOx) can be reduced. In light of increasingly stringent U . S . E P A emissions regulations on heavy-duty diesel engines, there is great need to develop combustion strategies that reduce targeted pollutants; for heavy-duty engines these are namely P M and N O x . A gravimetric technique has been used in this investigation to measure the P M emissions from a Cummins ISX 400 engine converted to H P D I of natural gas with diesel pilot. The engine experiments have been run using the A V L 8-mode steady-state test protocol, which simulates the U . S . E P A heavy-duty transient test procedure. The repeatability of particulate matter emissions measurements from H P D I fuelling has been established, as well as the effect of varying injection pressure, timing and pilot diesel fuel quantity on P M emissions. In addition, the composition of the H P D I exhaust P M was analysed for fuel-like and oil-like volatile organic fraction ( V O F ) components using a direct capillary injection of filter-borne P M into a gas chromatograph/flame ionization detector (GC/FID) technique. The effect of retarding injection timing on P M emissions was not consistent across the range of engine operating conditions characterized by modes 1,4,5, and 8 of the A V L test. It was suspected that high cylinder pressures resulting from advanced fuel injection timing or unwanted fuel interaction with cylinder wall and piston surfaces was impeding good mixing between injected natural gas and air for the current engine geometry. Results also showed that increased injection pressure was a potential strategy for reducing P M emissions, but for high load test points and the current engine geometry, a suspected air/fuel mixing problem may have produced higher P M emissions than expected. The difference in the P M emissions for these three injected pilot fuel quantities was not statistically significant, so no conclusions were drawn regarding the effect of increasing the pilot fuel quantity on P M emissions. Finally, the V O F of total P M mass emissions from H P D I combustion was between 34-58% based on analyses using modes 1,4,5,8 of the A V L 8-mode test. Table of Contents Abstract ii Table of Contents iii List of Tables vii List of Figures ix Preface xv Acknowledgments xvi 1.0 Introduction 1.1 Overview 1 1.2 Motivation for Research 1 1.3 Research Objectives and Scope 2 1.4 Methodology 2 2.0 Literature Review 2.1 Overview 4 2.2 Characteristics of Diesel Particulate Matter 5 2.2.1 Particulate Matter Volatile Organic Fraction 5 2.2.2 Particle Size Distribution 6 2.3 Particle Formation 7 2.4 Engine Operating Parameters Affecting Diesel Particulate Matter Emissions 2.4.1 Injection Pressure 8 2.4.2 Injection Timing 9 2.4.3 Lubricating O i l 10 2.4.4 Unburned Fuel 11 2.4.5 Engine Speed and Load 11 2.4.6 Wall Impingement/Air-Fuel Mixing 12 2.4.7 Fuel Properties 12 2.5 Dilution Systems for Measurement of Particulate Matter Emissions 2.5.1 Overview 13 2.5.2 Full-Flow Constant-Volume Sampling ( C V S ) System ( E P A Standard) 14 2.5.3 Partial Flow (Mini-) Dilution Systems 15 2.5.4 Mass Flow Measurement ( M F M ) Based System 15 i i i 2.5.5 Exhaust Gas Analysis ( E G A ) Based System 16 2.5.6 Sierra BG-1 Micro-Dilution System 17 2.6 Influence of Dilution System Operating Parameters on Particulate Filter Loadings, and P M Composition/Size Distribution Characteristics 2.6.1 Dilution Ratio 18 2.6.2 Temperature of Diluted Exhaust A t the Filter ("Filter Temperature") 20 2.7 Variability in Particulate Matter Mass Emissions 21 3.0 Experimental Procedure 3.1 Overview 22 3.2 Apparatus: Cummins ISX 400 Engine 22 3.3 Engine Test Protocol 24 3.4 Apparatus: Sierra BG-1 Micro-Dilution System 25 3.5 Pre-Sampling Procedures: Particulate Matter Emissions Testing 3.5.1 Filter Conditioning and Weighing 25 3.5.2 Sierra BG-1 Sampler Calibration 3.5.2.1 System Leak Check 26 3.5.2.2 Flowmeter Calibration 26 3.5.3 Engine Stabilization 26 3.6 Test Sampling Procedures: Particulate Matter Emissions Testing 27 3.7 Post-Sampling Procedures: Particulate Matter Emissions Testing 27 3.8 Gaseous Pollutant Monitoring 27 3.9 Procedure for Determining the Unburned Pilot Fuel-Like and Lubricating Oil -Like Components of the Particulate Volatile Organic Fraction ( V O F ) 28 4.0 Results and Discussion 4.1 Overview: HPDI Testing 30 4.2 Repeatability of P M Measurements (Study A ) 31 4.3 Same-Day Repeatability of A V L 8-Mode Test (Study B) 32 4.4 Injection Timing Study 33 4.5 Pressure Study 37 i v 4.5.1 Effect of Higher Injection Pressure on P M Emissions: Mode 1 (Low Speed, Low Load) and 5 (High Speed, L o w Load) of A V L 8-Mode Test 37 4.5.2 Effect of Higher Injection Pressure on P M Emissions: Mode 4 (Low Speed, High Load) and 8 (High Speed, High Load) of A V L 8-Mode Test 39 4.6 P M Composition: Unburned Pilot Fuel-Like and Lubricating Oil -Like Components of Particulate Matter Volatile Organic Fraction ( V O F ) 42 4.7 Pilot Fuel Quantity Study: Mode 2 of A V L 8-Mode Test 44 5.0 Conclusions 46 6.0 Recommendations 48 7.0 Nomenclature 50 8.0 References 52 Appendices Appendix A . A Critical Review of the E P A 2007 Mass-Based Particulate Matter Emissions Legislation Target of 0.01 g/bhp.hr for Heavy-Duty Engines 55 References 61 Glossary of Terms 63 Appendix B . Filter Handling Procedures B . l Filter Pre-Conditioning 64 B.2 P M Sampling Procedures B.2.1 Filter Handling 64 B.2.2 P M Sampling 64 B.2.3 Post-PM Sampling Filter Handling 64 B.2.4 Post-PM Weighing of Filters 65 B.3 Effect of Filter Conditioning Parameters on Particulate Matter Mass and Composition B.3.1 Relative Humidity in Conditioning Chamber 65 B.3.2 Temperature in Conditioning Chamber 65 Appendix C . Sierra BG-1 Micro-Dilution Tunnel Calibration and Operation Procedures C l Pre-Test Procedures 66 v C.2 Procedures During Testing 67 C.3 Post-Test Procedures 67 C.4 BG-1 Calibration File: Verification of Flow Coefficients 68 C . 5 P M Test Data Sheet 69 Appendix D . Overview of P M Emissions Testing and Analyses D . O Overview of Tests and Analyses Performed 70 D . A Repeatability Study A : Emissions and Fuel Flow Tables and Plots 71 D . B Repeatability Study B : Emissions and Fuel Flow Tables and Plots 75 D . T Timing Study: Emissions and Fuel Flow Tables and Plots 78 D.P Pressure Study: Emissions and Fuel Flow Tables and Plots 86 D . P W Pilot Fuel Quantity Study: Emissions and Fuel Flow Tables and Plots 93 Appendix E . Direct Capillary Gas Injection Method of Analysing Particulate Matter Components to Determine Lubricating Oil -Like and Unburned Diesel Fuel-Like Fractions in the Volatile Organic Hydrocarbon Fraction of Exhaust P M E . 1 Calculation of Fuel and Oil-Derived Hydrocarbons in Engine Particulate Matter 96 E . 1.1 Physical Process of Particulate Organic Fraction Volatization 96 E . 1.2 Interpretation of Chromatographic Spectra 96 v i List of Tables Table 3.2.1 Cummins ISX 400 Engine Configuration Data 23 3.3.2 A V L 8-Mode Steady-State Test Engine Operating Conditions 24 4.1.1 H P D I Test Matrix 30 4.2 Coefficients of Variation for P M Emissions Data: Modes 1, 4, 5, 8 of A V L 8-Mode Test and Comparative Data from Equivalent ISO 11-Mode Test Points 31 4.7 Diesel Pilot Fraction of Total Fuel Flow and Natural Gas ( C N G ) Flow for 3 Different Pilot Pulse Width Settings 45 Table C.5 P M Testing Data Entry Sheet 69 D.O Table of P M Testing and Analyses 70 DA(i ) . Mode 1 Emissions (g, kg/bhp.hr) : Repeatability Study A 71 DA(iv) . Mode 4 Emissions (g, kg/bhp.hr) : Repeatability Study A 71 DA(v) . Mode 5 Emissions (g, kg/bhp.hr): Repeatability Study A 71 DA(vi i i ) . Mode 8 Emissions (g, kg/bhp.hr) : Repeatability Study A 71 D A I . Mode 1 Emissions and Fuel Flows (g, kg/hr): Repeatability Study A 72 D A 4 . Mode 4 Emissions and Fuel Flows (g, kg/hr): Repeatability Study A 72 D A 5 . Mode 5 Emissions and Fuel Flows (g, kg/hr): Repeatability Study A 73 D A 8 . Mode 8 Emissions and Fuel Flows (g, kg/hr): Repeatability Study A 73 DB(i) . Repeatability Study B : A V L 8-Mode Test #1: Modal Emissions Data (g, kg/bhp.hr) 75 DB(ii) . Repeatability Study B : A V L 8-Mode Test #2: Modal Emissions Data (g, kg/bhp.hr) 75 D B 1. Repeatability Study B : A V L 8-Mode Test # 1: Modal Emissions & Fuel Flow Data (g, kg/hr) 76 D B 2 . Repeatability Study B : A V L 8-Mode Test #2: Modal Emissions & Fuel Flow Data (g, kg/hr) 76 Table D T I . Timing Study: Modes 1, 4 of A V L 8-Mode Test Emissions Data (g, kg/bhp.hr) 78 v i i Table D T 2 . Timing Study: Modes 5, 8 of A V L 8-Mode Test Emissions Data (g, kg/bhp.hr) 78 D T 3 . Timing Study: Modes 1, 4 of A V L 8-Mode Test Emissions & Fuel Flow Data (g, kg/hr) 79 D T 4 . Timing Study: Modes 5, 8 of A V L 8-Mode Test Emissions & Fuel Flow Data (g, kg/hr) 79 D P I . Pressure Study: Emissions and Fuel Flow Data for Modes 1, 4, 5, 8 of A V L 8- Mode Test (g, kg/hr) 86 DP2. Pressure Study: Emissions Data for Modes 1, 4, 5, 8 of A V L 8-Mode Test (g, kg/bhp.hr) 86 D P W 1 . Emissions For Various Pi 1 ot Fuel Pulse Width (g, kg/bhp-hr) 93 D P W 2 . Emissions and Fuel Flows for Various Pilot Fuel Pulse Widths (g, kg/hr) 93 Table E l . P M Filter Codes and Corresponding P M Test Point Description 98 E2. Raw Data from Particulate V O F Analysis: Primary Filters from Repeatability Study B Test Points: Modes 1,4,5,8 of A V L 8-Mode Test 99 E3. Raw Data from Particulate V O F Analysis: Backup Filters from Repeatability Study B Test Points: Modes 1,4,5,8 of A V L 8-Mode Test 99 v i i i L i s t o f F i g u r e s Figure 2.2 Particulate Matter Component Breakdown Pie Chart 5 Figure 2.2.2 Structural Characteristics of Particulate Matter: Nuclei and Accumulation Mode Particles 7 Figure 2.5.5 Exhaust Gas Analysis System: Relevant C 0 2 Measurements 16 Figure 2.5.6 Sierra BG-1 Micro-Dilution System 17 Figure 3.9a Typical Chromatographic Spectrum for Particulate V O F 28 Figure 3.9b Typical Chromatographic Spectrum for Prepared Diesel Pilot Fuel Sample 28 Figure 3.9c Typical Chromatographic Spectrum for Lubricating O i l Sample 28 Figure 4.4a. Effect of Injection Timing on P M , N O x , and C O Emissions: Mode 1 of A V L 8-Mode Test 34 Figure 4.4b. Effect of Injection Timing on P M , N O x , and C O Emissions: Mode 4 of A V L 8-Mode Test 34 Figure 4.4c. Cylinder Pressure Traces for Advances, Intermediate, and Retarded Injection Timing Scenarios: Mode 4 of A V L 8-Mode Test 35 Figure 4.4d. Effect of Injection Timing on P M , N O x , and C O Emissions: Mode 5 of A V L 8-Mode Test 36 Figure 4.4e. Effect of Injection Timing on P M , N O x , and C O Emissions: Mode 8 of A V L 8-Mode Test 37 Figure 4.5a. Effect of Gas Rail Pressure on P M Emissions: Modes 1,5 of A V L 8-Mode Test 38 Figure 4.5b. Differential Heat Release Plot for 13 and 25 M P a Injection Pressure Cases: Mode 5 of A V L 8-Mode Test 39 Figure 4.5.2a. Effect of Gas Rail Pressure on P M Emissions: Modes 4,8 of A V L 8-Mode Test 39 Figure 4.5.2b. Differential Heat Release Plot for 23 and 25 M P a Injection Pressure Cases: Mode 8 of A V L 8-Mode Test 40 ix Figure 4.5d. Carbon Monoxide Emissions L o g : 17 and 25 M P a Injection Pressure Cases, Mode 4 of A V L 8-Mode Test 41 Figure 4.5e. Carbon Monoxide Emissions L o g : 23 and 25 M P a Injection Pressure Cases, Mode 8 of A V L 8-Mode Test 41 Figure 4.6d. P M Component Emission Rates (g/hr): Repeatability Study, Modes 1,4,5,8 of A V L 8-Mode Test #1 43 Figure 4.6e P M Component Emission Rates (g/hr): Repeatability Study, Modes 1,4,5,8 of A V L 8-Mode Test #2 43 Figure 4.7a. Effect of Varying Pilot Pulse Width on P M Emissions: Mode 2 of A V L 8-Mode Test 45 Figure D A I . Fuel Flow and P M Emissions Data for 5 Repeated Mode 1 Test Points 74 Figure D A 4 . Fuel Flow and P M Emissions Data for 5 Repeated Mode 4 Test Points 74 Figure D A 5 . Fuel Flow and P M Emissions Data for 5 Repeated Mode 5 Test Points 74 Figure D A 8 . Fuel Flow and P M Emissions Data for 5 Repeated Mode 8 Test Points 74 Figure D B 1. Fuel Flow and P M Emissions Data for A V L 8-Mode Test # 1 77 Figure D B 2 . Fuel Flow and P M Emissions Data for A V L 8-Mode Test #2 77 Figure D T I i. Cylinder Pressure Trace for 3 Injection Timing Settings: Mode 1 of A V L 8-Mode Test 80 Figure D T l i i . Differential Heat Release for 3 Injection Timing Settings: Mode 1 of A V L 8-Mode Test 80 Figure D T I iii . Integrated Heat Release for 3 Injection Timing Settings: Mode 1 o f A V L 8-Mode Test 80 Figure D T l i v . C O Emissions vs. Time for 3 Injection Timing Settings: Mode 1 of A V L 8-Mode Test 80 Figure D T 4 i . Cylinder Pressure Trace for 3 Injection Timing Settings: Mode 4 of A V L 8-Mode Test " 81 x Figure D T 4 i i . Differential Heat Release for 3 Injection Timing Settings: Mode 4 of A V L 8-Mode Test 81 Figure D T 4 i i i . Integrated Heat Release for 3 Injection Timing Settings: Mode 4 of A V L 8-Mode Test 81 Figure D T 4 i v . C O Emissions vs. Time for 3 Injection Timing Settings: Mode 4 of A V L 8-Mode Test 81 Figure D T 5 i . Cylinder Pressure Trace for 3 Injection Timing Settings: Mode 5 of A V L 8-Mode Test 82 Figure D T 5 i i . Differential Heat Release for 3 Injection Timing Settings: Mode 5 of A V L 8-Mode Test 82 Figure D T 5 i i i . Integrated Heat Release for 3 Injection Timing Settings: Mode 5 of A V L 8-Mode Test 82 Figure DT5iv . C O Emissions vs. Time for 3 Injection Timing Settings: Mode 5 of A V L 8-Mode Test 82 Figure D T 8 i . Cylinder Pressure Trace for 3 Injection Timing Settings: Mode 8 of A V L 8-Mode Test 83 Figure D T 8 i i . Differential Heat Release for 3 Injection Timing Settings: Mode 8 of A V L 8-Mode Test 83 Figure D T 8 i i i . Integrated Heat Release for 3 Injection Timing Settings: Mode 8 of A V L 8-Mode Test 83 Figure D T 8 i v . C O Emissions vs. Time for 3 Injection Timing Settings: Mode 8 of A V L 8-Mode Test 83 Figure D T I v. P M , N O x , C O Plots for 3 Injection Timing Settings: Mode 1 of A V L 8-Mode Test 84 Figure DT4v . P M , N O x , C O Plots for 3 Injection Timing Settings: Mode 4 of A V L 8-Mode Test 84 Figure D T 5 v . P M , N O x , C O Plots for 3 Injection Timing Settings: Mode 5 of A V L 8-Mode Test 84 Figure DT8v. P M , N O x , C O Plots for 3 Injection Timing Settings: Mode 8 of A V L 8-Mode Test 84 x i Figure D T l v i . Timing Study: Fuel Flow and P M Emissions Data for Mode 1 of A V L 8-Mode Test 85 Figure D T 4 v i . Timing Study: Fuel Flow and P M Emissions Data for Mode 4 of A V L 8-Mode Test 85 Figure D T 5 v i . Timing Study: Fuel Flow and P M Emissions Data for Mode 5 of A V L 8-Mode Test 85 Figure D T 8 v i . Timing Study: Fuel Flow and P M Emissions Data for Mode 8 of A V L 8-Mode Test 85 Figure D P . Pressure Study: Fuel Flow and P M Emissions Data for Modes 1,4,5,8 of A V L 8-Mode Test 87 Figure D P la . Cylinder Pressure Trace for Gas Injection Pressure Settings of 10 and 25 M P a : Mode 1 of A V L 8-Mode Test 88 Figure D P l b . Differential Heat Release for Gas Injection Pressure Settings of 10 and 25 M P a : Mode 1 of A V L 8-Mode Test 88 Figure D P l c . Integrated Heat Release for Gas Injection Pressure Settings of 10 and 25 M P a : Mode 1 of A V L 8-Mode Test 88 Figure DPI d. C O Emissions vs. Time for Gas Injection Pressure Settings of 10 and 25 M P a : Mode 1 of A V L 8-Mode Test 88 Figure DP4a. Cylinder Pressure Trace for Gas Injection Pressure Settings of 17 and 25 M P a : Mode 4 of A V L 8-Mode Test 89 Figure DP4b. Differential Heat Release for Gas Injection Pressure Settings of 17 and 25 M P a : Mode 4 of A V L 8-Mode Test 89 Figure DP4c. Integrated Heat Release for Gas Injection Pressure Settings of 17 and 25 M P a : Mode 4 of A V L 8-Mode Test 89 Figure DP4d. C O Emissions vs. Time for Gas Injection Pressure Settings of 17 and 25 M P a : Mode 4 of A V L 8-Mode Test 89 Figure DP5a. Cylinder Pressure Trace for Gas Injection Pressure Settings of 13 and 25 M P a : Mode 5 of A V L 8-Mode Test 90 Figure DP5b. Differential Heat Release for Gas Injection Pressure Settings of 13 and 25 M P a : Mode 5 of A V L 8-Mode Test 90 xii Figure DP5c. Integrated Heat Release for Gas Injection Pressure Settings of 13 and 25 M P a : Mode 5 of A V L 8-Mode Test 90 Figure DP5d. C O Emissions vs. Time for Gas Injection Pressure Settings of 13 and 25 M P a : Mode 5 of A V L 8-Mode Test 90 Figure DP8a. Cylinder Pressure Trace for Gas Injection Pressure Settings of 23 and 25 M P a : Mode 8 of A V L 8-Mode Test 91 Figure DP8b. Differential Heat Release for Gas Injection Pressure Settings of 23 and 25 M P a : Mode 8 of A V L 8-Mode Test 91 Figure DP8c. Integrated Heat Release for Gas Injection Pressure Settings of 23 and 25 M P a : Mode 8 of A V L 8-Mode Test 91 Figure DP8d. C O Emissions vs. Time for Gas Injection Pressure Settings of 23 and 25 M P a : Mode 8 of A V L 8-Mode Test 91 Figure D P l e . P M , N O x , C O Plots for Mode 1: Injection Pressure Settings of 10 and 25 M P a 92 FigureDP4e. P M , N O x , C O Plots for Mode 4: Injection Pressure Settings of 17 and 25 M P a 92 Figure DP5e. P M , N O x , C O Plots for Mode 5: Injection Pressure Settings of 13 and 25 M P a 92 Figure DP8e. P M , N O x , C O Plots for Mode 8: Injection Pressure Settings of 23 and 25 M P a 92 Figure D P W a . Cylinder Pressure Trace for 3 Pilot Pulse Width Settings: Mode 2 of A V L 8-Mode Test 94 Figure D P W b . Differential Heat Release for 3 Pilot Pulse Width Settings: Mode 2 of A V L 8-Mode Test 94 Figure D P W c . P M , N O x , C O Plots for 3 Pilot Pulse Width Settings: Mode 2 of A V L 8-Mode Test 95 Figure D P W d . Fuel Flow and P M Emissions Data for 3 Pilot Pulse Width Settings: Mode 2 of A V L 8-Mode Test 95 Figure E l . Chromatographic Spectrum of Particulate V O F on Filter EM-263 100 x i i i Figure E2. Chromatographic Spectrum of Particulate V O F on Filter EM-264 101 Figure E3. Chromatographic Spectrum of Particulate V O F on Filter EM-269 102 Figure E4. Chromatographic Spectrum of Particulate V O F on Filter EM-270 103 Figure E5. Chromatographic Spectrum of Particulate V O F on Filter EM-271 104 Figure E6. Chromatographic Spectrum of Particulate V O F on Filter EM-272 105 Figure E7. Chromatographic Spectrum of Particulate V O F on Filter EM-277 106 Figure E8. Chromatographic Spectrum of Particulate V O F on Filter EM-278 107 Figure E9 . Chromatograph ic Spectrum of Particulate V O F on Filter EM-281 108 Figure E10. Chromatographic Spectrum of Particulate V O F on Filter EM-282 109 Figure E l l . Chromatographic Spectrum of Particulate V O F on Filter EM-289 110 Figure E12. Chromatographic Spectrum of Particulate V O F on Filter EM-290 111 Figure E13. Chromatographic (repeat) Spectrum of Particulate V O F on Filter EM-290 112 Figure E14. Chromatographic Spectrum of Particulate V O F on Filter EM-291 113 Figure E15 Chromatographic Spectrum of Particulate V O F on Filter EM-292 114 Figure E16 Chromatographic Spectrum of Particulate V O F on Filter EM-297 115 Figure E17 Chromatographic Spectrum of Particulate V O F on Filter EM-298 116 Figure E18. Chromatographic (travel blank) Spectrum of Particulate V O F on Filter EM-261 117 Figure E19. Chromatographic Spectrum of Westport Diesel Fuel Sample 118 Figure E20. Chromatographic Spectrum of Westport Engine O i l Sample 119 Preface I would like to thank Westport Innovations, N S E R C and Drs. Philip H i l l , Steven Rogak and Patric Ouellette for the opportunity to investigate ways in which particulate matter emissions can be reduced using H P D I as an alternative fuelling strategy for heavy-duty engine applications. This work has allowed me to develop my technical engineering and research skills, as well as satisfy a strong personal desire to make a positive individual contribution to the environment. Finally, to my parents, Jean and Ruth Baribeau, and my brother, Andre. The benefit of their unconditional support, encouragement and respect has made it possible for me to undertake intensive activities such as this M . A . Sc. project. x v Acknowledgments The completion of this work would not have been possible without the invaluable collaboration of the following people and organizations: Adam Craddock Alan Steeves Barbara Murray BB Ursu Campbell Perry Carl Jemma Christine Black Costi Nedelcu Dr. Guenter Eigendorf Dr. Jim McTaggart-Cowan Dr. Michael Brauer Dr. P.G. Hill Dr. Patric Ouellette Dr. Steven Rogak Dr. Tony Kozak George Dancu Gerry Rohling Guenter Blatt James Harrington Jan Marsden Jeff Thompson John Smitherman John Todd Josef Sircelj Kevin Oversby Matthew Olin Mary Mager Mati Raudsepp National Sciences and Engineering Research Council of Canada (NSERC) Pamela Nair Paul Grant Radu Oprea Ricardo Consulting Engineers Robin Poworoznik Ryan Thompson Sandeep Munshi Scott MacDonald Sean Bygrave Shu Oshika Stewart Whitfield UBC Department of Chemistry UBC Department of Forest Sciences UBC Department of Mechanical Engineering UBC Department of Materials and Metallurgical Engineering UBC School of Occupational Hygiene Vanessa Trkulja Victor Leung Westport Innovations Inc. Westport Trucking Team Westport IT Department x v i 1.0 Introduction 1.1 Overview Pressure to reduce particulate matter (PM) emissions from heavy-duty vehicles has come from the recognition that this minority class of the vehicle population contributes a disproportionately large percentage of P M to ambient air. For example, the E P A emissions trends report ( U . S . E P A , 2000a) reported that diesel engines were estimated to contribute 77% of national mobile P M emissions under 2.5 microns in diameter (PM 2.5) in 1998. Particulate matter emissions from diesel engines contain black carbon (soot), volatile organic species, sulfates, nitrates, fuel and lubricant oil additives, and wear metals from engine operation. The soot results from incomplete oxidation of fuel into carbon dioxide ( C O 2 ) ; Driscoll et al. (1997) found that rats exposed to soot aerosols exhibited an inflammatory lung response, cell-derived oxidative damage and an increase in cell mutational response, suggesting a link to lung cancer. The volatile organic species from partially burned fuel and lubricating oil products and unburned fuel and oil has also caused mutagenic cell activity in rats based on Findings by Huisingh et al. (1978). The relative importance of particulate matter size and composition remains an open question, as discussed in Appendix A . A number of strategies sharing the common goal of reducing engine tailpipe P M and nitrogen oxide (NOx) emissions have been under development for several years. Many involve adding expensive and problematic exhaust aftertreatment systems to existing engines. B y contrast, the high-pressure direct injection (HPDI) of natural gas with diesel pilot results in cleaner in-cylinder combustion that reduces the need to mitigate emissions downstream of the combustion chamber. In addition, Hodgins et al. (1996) showed that the efficiency o f the diesel engine is maintained, an achievement that spark ignition and pilot ignition engines reliant on fumigation of natural gas have been unable to match. 1.2 Motivation for Research Given the detrimental effects listed above relating to diesel P M , there is significant value in measuring P M emissions from H P D I engines to confirm their lower emission rates 1 relative to P M emissions from diesel-powered engines. When testing to determine an engine's P M emission rate, it is important to obtain samples that represent the P M that is formed when engine exhaust is diluted with ambient air. There is significant challenge associated with accomplishing this seemingly easy goal. It has been shown by Lapuerta et al. (1999), Abdul-Khalek et al. (1999, 2000) and MacDonald et al. (1980) that many factors such as: residence time of particles in the dilution tunnel, dilution ratio, relative humidity and sample temperature at the filter ("filter temperature") can affect both particulate filter loadings as well as composition and size distribution characteristics. 1.3 Research Objectives and Scope The objectives of this investigation are to examine the following: 1. The repeatability of H P D I P M emission measurements 2. The effect of injection timing and pressure on H P D I P M emissions 3. The relationship between H P D I P M mass emissions and injected pilot fuel quantity 4. The composition of H P D I particulate matter P M emissions will be correlated with pressure trace, heat release, C O , N O x and fuel flow data where such relationships help to explain trends in the experimental data. The scope of this work does not include a size distribution study of H P D I exhaust P M . Distribution information will hopefully be obtained in near future projects for the H P D I particulate matter emissions. 1.4 Methodology A l l P M measurements for the studies mentioned in 1-3 from section 1.3 were carried out using a gravimetric technique. A Sierra BG-1 Micro-Dilution system was used to draw a fraction of the exhaust from a Cummins ISX 400 engine, dilute it, and collect the resulting P M on a set of filters subsequently weighed to determine the mass of P M collected. The tests were carried out at a fixed dilution ratio, test duration, and diluted sample flowrate. The test protocol used for ISX 400 engine operation in studies 1-3 was the A V L 8-mode steady-state test protocol. Finally, the procedure used to identify the pilot fuel-like and lubricating oil-like components of the P M was direct capillary injection gas chromatography (GC) of filter-2 borne P M . The P M collected on filters during testing was analysed using a G C and flame ionization detector (FID) to determine the amounts of unburned pilot fuel-like and lubricating oil-like components in the P M . 3 Literature Review 2.1 Overview The particles in diesel exhaust are of special concern because, due to their respirable size, they can penetrate deep into human lungs. The composition of diesel particulate matter includes many species that are known for their adverse health effects, including several carcinogens. For this reason, the U . S . E P A continues to tighten legislation governing particulate matter emissions from heavy-duty (HD) diesel engines. A discussion of the health-based research behind proposed 0.01 g/bhp.hr P M emission rate standards is given in Appendix A . Diesel combustion involves a diffusion flame in which fuel-rich gas pockets often form. Combustion in these pockets produces high amounts of P M , namely from unoxidized soot and unburned fuel resulting from insufficient air available for complete combustion. Previous studies have shown that engine operating parameters such as injection pressure and timing can influence P M emissions from diesel-fuelled engines. In addition, both the presence of lubricating oil in the engine cylinders and poor combustion efficiency yielding significant amounts of unburned fuel during combustion can contribute to exhaust P M . Measured P M emissions are also a function of the conditions under which the P M is collected. Test parameters such as dilution ratio and sample temperature at the filter have been found to significantly affect the outcome of collected P M mass. The aforementioned variables affect both the size distribution characteristics and particle number concentrations, which ultimately affect the mass of P M collected on the filters. The contents of this literature review address the following PM-related topics: • characteristics of Diesel P M • particle formation • engine operating parameters affecting P M emissions • dilution systems for measuring P M emissions • dilution system operating parameters affecting P M emissions • variability in P M mass emissions 2.2 Characteristics of Diesel Particulate Matter Diesel PM is a complex mixture of solid and liquid material. Specifically, the components include soot, partially burned hydrocarbons, sulfate and nitrate ions, unburned fuel and lubricating o i l and trace metals such as sulfur, calcium, iron, silicon, chromium zinc, phosphorous, and calcium originating from the diesel fuel and lubricating oi l additive packages. Entire studies by authors such as Schauer et al . (1999) have been dedicated to determining the nature o f the hydrocarbons found in diesel exhaust PM. Figure 2.2.3 shows the composition of particulate matter from a modern diesel engine. Figure 2.2 Particulate Matter Component Breakdown Particulate Component Breakdown: EURO III Engine, EN590/00 Fuel, ESC Test Cycle (Wedekind et al., 2000) sulphate ions nitrate ions 5% \ 2 % water 7% f^m fuel-like / volatile organics jsoot and 16% i f ^ f f l ^ / other J 57% oil-like volatile organics 13% The composition of the particulate matter at any given time is a function of engine conditions and can be influenced by the PM collection system itself. 2.2.1 Particulate Volatile Organic Fraction (VOF) The V O F is the portion of the exhaust PM that is thermally desorbed when the PM is heated to a temperature o f 325°C. Wedekind et al . (2000) determined the V O F using this thermal desorption technique to be 29% (by mass) of the test cycle PM emissions from engine operation on the European Stationary Cycle (ESC) test. Another method to determine the organic fraction of the PM is solvent extraction. The particulate matter is continuously cycled with a solvent such as dichloromethane, and the extracted organic 5 fraction subsequently referred to as the soluble organic fraction (SOF). Meyer et al. (1980) found that the contribution of the S O F to total P M emissions ranged between 8-36% (by mass), using a similar extraction technique. The purpose of the S O F and V O F separation procedures is two-fold. First, the quantity of organic material in the particulate matter can be determined, and this organic carbon quantity subtracted from the total P M mass to determine the remaining quantity of elemental carbon. Second, the organic fraction can be further analysed using gas chromatography/flame ionization detection (GC/FID) to determine the quantities of unburned pilot fuel-like and lubricating oil-like components in the organic fraction. A n explanation of the procedure and assumptions on which this analysis is based is given in Appendix E of this report. The G C / F I D results aid in understanding the origins of the particulate matter. The V O F itself is not a regulated component of exhaust particulate matter emissions, but is known to contain potentially carcinogenic compounds such as polycyclic aromatic hydrocarbons (PAH) . The purpose of the V O F analysis carried out in this investigation was to determine the sources of V O F resembling unburned pilot fuel and lubricating oil . 2.2.2 Particle Size Distribution Particles initially formed during combustion are referred to as primary particles. Larger particles formed from dynamic processes such as nucleation, condensation, adsorption and agglomeration involving these primary particles are referred to as secondary particles. The size distribution of diesel P M is often bi-modal, which means that some of the particles are in the 7-40 nm diameter range, or nuclei mode as further described in section 2.3, and some of the larger particles are found in the 40-1000 nm range, or accumulation mode. Nuclei mode particles tend to consist of a nucleus of soot or hydrated sulfuric acid ( H 2 S 0 4 ) with one or many layers of condensed hydrocarbon species adsorbed to their surfaces, whereas accumulation mode particles consist of agglomerations of existing nuclei mode particles, as shown in Figure 2.2.2 on the following page. 6 Figure 2.2.2 Structural Characteristics of Particulate Matter : Nuclei and Accumulat ion M o d e Particles (source: DieselNet) Primary Ca rbon Spherule H 2 S O 4 . H 2 O Part icle Nuclei Mode Particles 0.007 - 0.04 urn diameter Accumulation Mode Particle 0.04 - 1u.m diameter Abdul-Khalek et al. (1998) found that the number of particles in the nuclei mode in diesel exhaust contributed to more than 50% of the total particle count, whereas most of the PM mass originated from the accumulation mode particulate fraction. The size distribution of diluted exhaust PM is known to be a function of the exhaust dilution process parameters such as: dilution temperature, dilution ratio, relative humidity and residence time in system mixing chamber. 2.3 Particle Formation Clark et al. (1998) described sooty particulate matter as consisting of spherules of elemental carbon agglomerated together into complex and irregular forms. These clustered and chain-like soot particles are typically less than 1 micron in length or diameter. In direct-injection diesel engines, soot results from partial pyrolysis of hydrocarbon in fuel-rich zones such as the core of the fuel spray. Partial pyrolysis means that some of the fuel is not fully oxidized to combustion products, but instead decomposes without oxygen at high temperatures. One result of this is that elemental carbon from the combustion fuel is not completely oxidized to carbon dioxide and soot is produced. Heywood (1988) attributed the soot fraction that contributes significantly to PM mass emissions to a 10% fraction of unoxidized elemental carbon left over from the combustion process. 7 Nanoparticles are a classification of particles with principal dimension less than 50 nm. Kayes and Hochgreb (1998) proposed that nanoparticles usually form when a species in the exhaust gas forms an initial nucleus (homogenous nucleation), and subsequently grow in size when volatile species are absorbed and adsorbed to them depending on local temperature and species concentrations. The resulting species are said to be in the nuclei mode. Abdul-Khalek et al. (1999) and Lapuerta et al. (1999) showed that the exact growth rate of particles depends on diluted exhaust mixture conditions such as temperature, dilution ratio, relative humidity and residence time of the dilution system mixing chamber. Usually the only volatile exhaust species present in high enough concentrations to produce significant particle growth were sulfuric acid and soluble organic hydrocarbons, according to Abdul-Khalek et al. (2000). 2.4 Factors Affecting Diesel Particulate Matter Emissions 2.4.1 Inject ion Pressure Williams et al. (1989) studied the effects of injection pressure on P M emissions, and found that lower injection pressures and subsequent overfuelling caused a 26% increase in P M emissions. Lower injection pressures cause excess P M formation because poor fuel atomization decreases the surface area between fuel and oxidizer, resulting in decreased combustion efficiency. Fuel demand at lower injection pressures increases because as the combustion efficiency decreases, more fuel must be added to maintain a constant engine load. Kato et al. (1989) showed that ultra-high injection pressure of 220 M P a decreased P M emissions by 80% when compared to P M emissions for injection pressure of 100 M P a . The results obtained also supported the phenomenon of reduced soot concentration resulting from better atomization of fuel spray. The authors also found that high injection pressure improved air entrainment into the spray, and resulted in faster combustion and a high soot oxidation rate. Instantaneous focused-shadow pictures of the spray verified the latter phenomenon. Needham et al. (1990) also investigated the effects of high injection pressures on P M emissions, and, at full load, found that increased injection pressure helped to improve air/fuel mixing rates, thus controlling soot emissions and lowering the particulate carbon 8 fraction. In order to mitigate N O x emissions at this high injection pressure, it was necessary to retard injection timing. Ikegami et al. (1990) suggested that high injection pressures, causing finer fuel droplets and better air/fuel mixing, diminished the condensation of high boiling point fuel components on the combustion chamber walls. The result was a decrease in particulate S O F originating from this high boiling point fraction, leading to a reduction in overall P M emissions. Further investigation of high injection pressures strategies in diesel engines by Pierpont et al. (1995) attributed resulting decreases in P M emissions to better fuel atomization and air/fuel mixing via extended spray penetration, air entrainment and wall impingement. However, the favourable reduction in P M emissions was coupled with a rise in N O x emissions, requiring that timing be retarded and subsequently resulting in an increase in brake specific fuel consumption (BSFC) . From the numerous sources referred to in this section, it is easily seen that P M reductions can be achieved using high-pressure injection strategies, but at the expense of increased fuel consumption and higher parasitic losses. 2.4.2 Inject ion T i m i n g It is well known that a trade-off exists between optimizing engine operation for optimal P M and N O x emissions. When injection timing is retarded to control N O x emissions, P M tends to increase (Heywood, 1988). P M likely increases due to lower temperatures in the cylinder (resulting from retarded timing), which lowers the soot oxidation rate. For example, Smith et al. (1998) studied the effect of varying injection timing on P M emissions for 2 engine conditions: 2000 rpm and 2 bar brake mean effective pressure ( B M E P ) , and 2750 rpm, 5 bar B M E P . The authors found that for the 2000 rpm/2 bar B M E P condition, retarding the injection timing increased P M emissions and reduced emissions of nitrogen oxides (NO x ) . A t 2750 rpm and 5 bar B M E P , retarding injection timing did not affect P M emissions, but did reduce brake specific N O x emissions. Unfortunately the reduction in N O x emissions achieved by retarding injection timing comes with a price of increased brake specific fuel consumption. 9 Gomes and Yates (1992) also found that at low load (2 bar B M E P ) , P M emissions increased rapidly as start of fuel injection was retarded from T D C (0° crank angle), with most of the P M emissions from the S O F . A t 500 kPa B M E P , there was no consistent trend between retarding injection timing and increasing P M emissions, similar to the findings of Smith et al. (1998). 2 .4 .3 L u b r i c a t i n g O i l A s mentioned in section 2.2.1, unburned oil-like hydrocarbons can often be found in the volatile or soluble organic fraction ( V O F or SOF) of exhaust P M . A steady-state engine operation study by Mayer et al. (1980) using radioactive oil tracers found that 1.5 to 25 % (by mass) of the total P M mass emissions originated from lubricating oil, with the low figure corresponding to the low speed/high load condition and the high figure corresponding to the high speed/low load operating condition. This result makes practical sense, because low speed, high load yields a higher fuel/air ratio than that corresponding to the high speed, low load condition. High-speed engine operation also gives rise to higher oil consumption, and more opportunity for resulting oil-like P M components. In the same study, lubricating oil contributed to the total S O F of the exhaust P M by between 16 and 80 % (by mass), with the low figure corresponding to low speed/high load and the high figure corresponding to high speed/medium load. . The particulate S O F increased with increasing engine speed, but had no clear dependency on load. B y using a radioactive tracer technique, it was determined that all of the engine oil contribution to P M manifested itself in the S O F . The authors suggested that the oil probably contributed to exhaust P M after most of the combustion reactions forming the insoluble (sooty) fraction of the P M had been completed. The two major sources of lube oil P M emissions were lube oil passing the valve stem seals and lube oil from the cylinder wall, with the latter being the dominant source. Two studies of the lube oil contribution to particulate S O F were performed by Andrews et al. (1991) on both a Petter A A I IDI engine and a turbocharged Perkins 4-236 D l engine at maximum torque. The technique used for this analysis separated lubricating oil from unburned fuel in a high-pressure liquid chromatograph ( H P L C ) with transfer of the unburned fuel fraction to a separate gas chromatograph (GC) for analysis. The percent 10 lube oil contribution to particulate S O F from the Petter engine testing was 25-80%. Testing on the Perkins engine attributed 86% of the particulate S O F emissions to lubricating oil . The reader is cautioned that reported contribution of unburned fuel vs. lubricating oil to particulate S O F is highly dependent on the method used to extract and analyse the S O F itself. 2.4.4 Unburned Fuel in Exhaust PM Williams et al. (1989) explained that survival of unburned or partially oxidized fuel in exhaust P M is known to result in high amounts of polycyclic aromatic hydrocarbons (PAH), which are known carcinogens. Chemical speciation of the exhaust P M and knowledge of the chemical composition of liquid diesel fuel show that the P A H in this unburned fuel is the only realistic "explanation" for large concentrations of P A H in the exhaust P M . Fuel injected late in the combustion cycle may be subjected to low-temperature zones that are below the fuel-ignition temperature, thus resulting in poor fuel combustion efficiency and high residual amounts of unburned fuel in the cylinder. It is suspected that the unburned fuel can bypass the high-temperature combustion zones in the cylinder and, once emitted, condense onto primary exhaust particles. Two mechanisms by which unburned fuel can survive in-cylinder temperature and pressure conditions are: late fuel injection and flame quenching due to walls, high speed flows and rapid drops in temperature. The relative abundance of unburned fuel in exhaust P M is a function of engine speed and load. The authors found that the bulk of the particulate S O F , including the P A H in particular, is from unburned fuel using the Andrews et al. (1991) technique described in section 2.4.3. These results seem to contradict those described in 2.4.3 regarding the lubricating oil contribution to P M mass; however, the reader is cautioned that reported contribution of unburned fuel vs. lubricating oil to particulate S O F is highly dependent on the method used to extract and analyse the S O F itself. 2.4.5 Effect of Engine Speed and Load on PM Emissions Gomes and Yates (1992) found that at low load (200 kPa B M E P ) , high pressure injection reduced P M emissions by 80%>, whereas for the same injection pressure, high load (500 kPa B M E P ) operation only reduced P M emissions by 50%. In addition, at low load, the S O F was 80-90% of total P M but declined to between 20-30% for high load operation. The authors suggested that the lower temperatures characteristic of low-load operation 11 favoured the condensation and adsorption of volatile species (the SOF) , resulting in higher adsorbed S O F mass in total P M emissions for low-load engine operation. A t low load, the S O F mass emissions decreased with increasing speed, and soot emissions increased with increasing speed. Low S O F emissions can be explained by higher exhaust gas temperatures occurring at higher speeds, which inhibit the adsorption of S O F species onto solid carrier particles. Higher soot at high speeds can be explained by a reduction in time available for soot oxidation at these higher speeds. A t high load, the S O F did not change markedly with increasing engine speed, but the soot emissions increased significantly with engine speed. Again, this was explained by a reduction in the time available for soot oxidation at these higher speeds. 2.4.6 Wall Impingement / Air-Fuel Mixing Considerations Rao et al. (1992) used high-speed photographic studies of the combustion to shed light on the P M generation process in a Ford F S D 425 engine. The authors found that at low load, the swirling air carried fuel vapour away from spray boundaries resulting in a lean mixture beyond the limit of combustion, and ensuing low cylinder temperatures, high amount of unburned fuel and hence high P M emissions. B y contrast, at high load, most of the injected fuel impinged on the cylinder wall, and heavy adjacent spray overlap created fuel-rich zones and sooty clouds in these areas. 2.4.7 Fuel Properties Fleisch et al. (1995) tested the performance and emissions of a Navistar 7.3-L V - 8 T444E direct-injection turbocharged diesel engine operating on dimethyl ether ( D M E ) . For the A V L 8-mode composite test, P M emissions were 0.033 g/bhp.hr using an A V L mini-dilution tunnel. This is a much lower P M emission rate than a conventionally-fuelled diesel engine. Further chemical analysis of the filter-borne P M showed that all of the P M originated from lubricating oil . Sorenson and Mikkelsen (1995) tested the emissions of an air-cooled, single-cylinder, 4-stroke normally aspirated direct injection CI engine also operating on D M E . The authors did not test P M emissions using a dilution system but used a Bosch smoke meter to 1 2 measure soot emissions. The Bosch smoke number (BSN) was approximately 0.1 for operation on D M E for all indicated engine power settings ranging from 0-4 k W . Over the same power range, diesel fuelling of the engine produced B S N ranging from 1 (lowest power) to 4 (highest power). The authors discussed the essentially soot-free D M E exhaust by explaining that D M E molecules have essentially no soot precursors. Engines have more recently been tested using Fischer-Tropsch (F-T) synthetic diesel fuel with low aromatic and sulfur content. Norton et al. (1998) tested a 1994 Navistar 7.3-L V-8 T444E direct-injection turbocharged CI engine for both the hot-start portion of the F T P transient test using a test cell dynamometer and the 5-mile route on a chassis dynamometer. The former test showed a 14% reduction in P M emissions in comparing results using F - T fuel vs. conventional No.2 diesel fuel. The latter test yielded a 24% reduction in P M emissions using F - T fuel vs. California diesel fuel. 2.5 Dilution Systems for Measurement of PM Emissions 2.5.1 O v e r v i e w Traditionally, P M measurements are carried out using a system that dilutes either the full exhaust flow or some fraction of it. The main considerations in designing a P M sample are dilution procedures and sample detection characteristics such as filter media, filter analysis and aerosol instrumentation. Systems that involve dilution of the entire exhaust flow are called full-flow, or constant volume sampling ( C V S ) systems. These systems require large amounts of space and capital to build. A s a result, other partial-flow systems that dilute only a fraction of the exhaust flow have been developed, referred to as mini- or micro-dilution systems. Since engine emissions reporting must comply with E P A standards, data obtained using partial-flow systems is required to correlate well with P M data obtained using the full-flow C V S system. The diluted exhaust flowing through the system is collected on filters that are weighed and subjected to chemical analyses to determine both the composition of the P M as well as the source from which it originated. The purpose of the dilution process is to simulate the reactions that occur between engine exhaust and ambient air, so that the P M collected on filters is representative that found in ambient air. More recent methods of measurement have been proposed, such as flame ionization detection, to provide real-13 time measurements of P M during transient tests. The results in this study were obtained using a Sierra B G - 1 micro-dilution system. When using a dilution system to collect P M , a variety of critical test parameters must be carefully chosen. These parameters include: temperature of the diluted exhaust mixture at the filters, dilution ratio, relative humidity, residence time and filter conditioning procedures. Unsupported selection of these test parameters affects P M mass loading and size distribution characteristics by varying the conditions under which the different particle formation reactions occur. 2.5.2 Flow Constant Volume Sampling (CVS) Dilution System (EPA) In 1972, the United States Environmental Protection Agency (US E P A ) defined a full flow C V S system for P M measurement of H D diesel engines. The system dilutes the entire exhaust stream coming from the engine with clean air in a tunnel. A fraction of this diluted exhaust flow is then drawn through a set of Teflon-coated glass fibre filters. A s described by the constant volume sampling name given to the system, the total mass flow through the tunnel is constant through all phases of engine testing. The mass flow is chosen such that the 5 2 ° C restriction on diluted exhaust temperature at the filters is adhered to under even the hottest engine driving phase. The dilution ratio in the C V S system is calculated as follows: DR = G E D F / G E X H Where GEDF is the diluted exhaust flowrate (held constant) and G E XH the exhaust flowrate. Since the system has a constant volume of diluted exhaust going through it, the dilution ratio is smaller for hot engine operating conditions and larger in idle mode, when the diluted exhaust temperature approaches the dilution air temperature. Since GEDF is constant, rearranging the equation above yields the " C V S " condition as follows: DR G E X H = G E D F = constant The C V S system is designed primarily to act as a physico-chemical reactor. The dilution and cooling that take place when clean air mixes with exhaust simulates the physical and chemical reactions that take place between post-tailpipe exhaust and ambient air. This way, the P M collected on filters is more representative of P M found in the atmosphere than P M trapped at the exhaust valve port in the engine. In addition to fulfilling the role 14 described above, the dilution system simultaneously serves as a measuring device for P M . Unfortunately, full-flow C V S systems are very bulky and costly, so much of the P M testing work currently performed involves the use of mini- and micro-dilution systems as described below. 2.5.3 Partial Flow (Mini-) Dilution Systems (PDS) These systems sample and dilute only a fraction of the total engine exhaust, which makes them smaller and more economical than the C V S system described above. However, their viability as an alternative to C V S systems must be proven; E E C directive 91/542 requires that equation D R ( G E X H ) = G E D F = constant must be fulfilled for a P D S within 7%. This requirement is based on the strong dilution ratio influence on the temperature and the exhaust gas component concentrations of the diluted exhaust stream. These quantities are integral to P M formation. 2.5.4 Mass Flow Measurement (MFM) Based System A s the name implies, mass flow measurement systems use mass flow quantities for the exhaust gas, dilution air, and diluted exhaust to determine the dilution ratio. Schindler and Engeljehringer (1989) defined the amount of exhaust sampled (G p ) in the tunnel and the dilution ratio (DR) as follows: G p = GT O T - G D I L DR = G T O T / G p Where GTOT is the flowrate of diluted exhaust through the system, and GDIL the flowrate of dilution air into the tunnel. The accuracy of mass flow meters for an M F M system is often specified to better than 1%. Errors in mass flow measurements can result in errors of 15% in the sample flow for moderate to high dilution ratios. The advantage of using M F M over an exhaust gas analysis ( E G A ; discussed in 2.5.5 below) system is that calculation of the dilution ratio is faster using M F M , due to faster response time of flow measurement devices. E G A implies the use of N D I R detection of exhaust CO2, with a relatively slow associated instrument response time. The time lag 15 associated with obtaining a reading from the CO2 analyzer between successive engine operating points is too long to satisfy transient testing sequences. 2.5.5 Exhaust Gas Analysis (EGA) Based System Dilution systems that use E G A rely on carbon dioxide concentration [CO2] measurements to indicate the degree to which the exhaust gases have been diluted. Macdonald et al. (1980) used the relationships below to determine the dilution ratio using E G A : The flow of CO2 through the system is represented in the diagram below: Figure 2.5.5 Exhaust Gas Analysis System: Relevant CO2 Measurements [C02]a X 1]a • t [C02]e X Tie where n represents molar flow, [CO2] represents the carbon dioxide concentration, and a, e, and d represent the dilution air, engine exhaust, and total flow in the dilution system, respectively. The dilution ratio (DR) itself is defined as r\d/ r\e, or molar flow of diluted exhaust divided by the molar flow of exhaust gas. Dilution system [C02]d X TJd B y conservation of mass, the CO2 entering the system must equal the C 0 2 exiting the system. Therefore equation (1) can be written: [C02]a X T l a + [C02]e X Tie = [C02]d X T l d Assuming no chemical reactions occur between the exhaust gases and the dilution air, equation (2) applies to the system: Tla + Tie = Id Combining equations (1) and (2) and solving for the dilution ratio gives equation (3): _ [co2l,wel-[co2lwel DR = Tld/ T l e [co2lwel-[co2l. 16 For concentrations measured using a non-dispersive infra-red (NDIR) C 0 2 analyzer, the dry concentration quantities must be converted to a wet basis before the true dilution ratio can be calculated, as seen in the equation below: [C02]e, w e t = [C0 2] e,d r y X (1 + [HzOle)1 2.5.6 Sierra BG-1 Micro-Dilution System (MFM) Since the late 1980s, Graze (1993) and Caterpillar Inc. have been developing a micro-dilution system to measure engine P M . Development of this variety of particulate matter measurement system was undertaken for a variety of reasons: • Need for a less costly system than the E P A recommended full-flow C V S system, with correlation to the C V S system as per ISO 8178 standards • Need for a portable system not requiring an elaborate amount of space nor setup/operational expertise • Short sampling times with high repeatability • Reduced P M deposition, mainly associated with transfer tube carrying exhaust to main dilution tunnel Fractional exhaust sampler insensitive to engine size (problem with mini-dilution tunnels in the past that were designed for a narrow range of exhaust flowrates, full - f l o w system capabilities limited to < 500 hp engines) Figure 2.5.6. Sierra Micro-Dilution System J J Dilution A i r Flow • ^ g ^ " , " " " " O " O » n Q " " n ' T T y a K = > Sample Flow Q e x h + Q a Perforated Stainless Steel Tube Pressure Vessel 17 The system uses volumetric flow measurements to determine the dilution ratio (DR), using values for the total flow through the tunnel of mixed exhaust and dilution air, and the value of dilution air going into the tunnel. TotalFlow TotalFlow DK = = TotalFlow - DilutionAirflow ExhaustFlow The system includes a PC-controlled calibration sequence that avoids the effect of flowmeter relative calibration accuracies, which can yield as high as ± 22% error at a dilution ratio of 10:1. The mentioned calibration sequence limits sample flow deviations to the repeatability capabilities of the individual flowmeters (0.10-0.15%)), which amount to a worst case of ± 2-3% scatter. Another virtue of the Sierra system is that it is designed to reduce the possibility of P M deposition on tunnel walls. T w o features of the system address this problem: the first is that a relatively short, 5-inch transfer tube channels a fraction of the engine exhaust to the main dilution tunnel, compared to EPA-recommended transfer tube lengths of 25 feet. Caterpillar in-house testing showed that a significant P M deposition occurred in the 25 foot line, even when it was heated according to ISO 8178 specifications; clearly the length of the transfer tube had to be minimized. The second feature of the system is that the porous tube dilution section of the system yields a significant reduction in P M deposition on tunnel walls. This is due to the effects of both transpiration cooling and high dilution air velocity emanating from the pores, physically impinging on P M travelling towards the tunnel walls by means of thermo- or electrophoresis, diffusion, or turbulent deposition and prohibiting them from depositing upon the walls. 2.6 Influence of Dilution System Operating Parameters on Particulate Matter Mass Loadings and Composition 2.6.1 D i l u t i o n R a t i o In the measurement of P M emissions, engine exhaust is generally diluted with filtered ambient air, and the resulting diluted exhaust mixture subsequently passed through a set of filters. The purpose of the dilution is to simulate the reactions occurring between P M and ambient air so that the collected particles resemble those contributing to atmospheric 18 pollution and human morbidity. The extent to which the given exhaust flow is diluted with ambient air is referred to as the dilution ratio, expressed as follows: D i l u t i o n rat io ( D R ) = ( Q e x h + Q a i r ) / Q e x h where Q e x h and Q a i r are the volumetric flow rates of the exhaust sample and dilution air at standard temperature and pressure, respectively. Dilution of exhaust affects both the temperature of the exhaust gases and the concentration of particulate species in the exhaust, subsequently affecting nucleation and condensation processes. Thus changing the dilution ratio can affect the total P M emission rate and size distribution characteristics of the exhaust P M in question. However, the E P A does not stipulate a specific dilution ratio as part of the P M testing procedure. Thus it is very important that researchers report the dilution ratio used for testing. Three studies described below found the similar result that particulate matter sampling at high dilution ratios causes a decrease in P M mass collected during testing. Lapuerta et al (1999) found that increasing the D R from 5 and 25 caused a 12% reduction in the measured P M mass. The filter temperature was held constant at 3 5 ° C for this trial. Further investigation of the effect of changing D R on particulate matter composition showed that increasing the D R from 2 to 8 decreased the measured P M mass by 38%). In this case, it was suggested that dilution was responsible for decreasing the vapour pressure of light hydrocarbons, as well as decreasing their saturation pressure by virtue of cooling. The result of these effects was a reduced probability of condensation and adsorption of soluble organic species by soot particles. Abdul-Khalek et al. (1999) established that the relationship between high dilution ratio and lower P M fdter loadings due to lower ultrafine particle generation was also attributed to reduced vapour phase concentration of all species present in the exhaust P M at high dilution ratios. These lower concentrations were found to weaken the nucleation of new particles and hence the driving force for growth. Some early research by MacDonald et al. (1980) investigating the effects of varying dilution ratio on P M emissions found that increasing the dilution ratio from 5 to 100 (for a fixed mixture temperature at the filters of 5 2 ° C ) decreased the P M mass by 26%. 19 Within this D R range, a substantial amount of the reduction took place between dilution ratios of 5 and 30. Abdul-Khalek et al. (2000) also found that a decreased particle diameter growth fate at high dilution ratios occurred for all dilution temperatures. This results from dilution of the exhaust, which reduces the concentration of all species in the exhaust. It should be noted that most practical dilution systems are nearly adiabatic, so increases in D R always accompany a decrease in temperature. 2.6.2 Temperature of Diluted Exhaust At the Filter ("Filter Temperature") The temperature of the diluted exhaust mixture serves as one indication of the degree to which the exhaust stream has been diluted. This temperature influences the characteristics of the particulate matter collected on the measurement system fdters and subsequently the P M emission rate calculated from this parameter. High filter temperatures will decrease the volatile organic fraction of the P M and lower temperatures will increase this fraction of the P M . MacDonald et al. (1980) also suggested that sample temperature at the fdter did affect the mass of P M accumulated; an increase in this temperature from 3 5 ° C to 100 °C (for DR=10.8) decreased the P M mass loading by 13%. This was due to a reduction in the soluble organic fraction of the P M from 35% at 3 5 ° C to 20% at 1 0 0 ° C . Abdul-Khalek et al. (1999) also showed that higher sample temperature at the filter resulted in lower P M fdter loadings due to lower ultrafine particle generation. Higher temperature was found to slow particle nucleation rate and subsequent growth due to the increase in vapour pressure of volatile species. Particles in the accumulation mode were largely unaffected by changes in dilution temperature. Accumulation mode particles are formed by agglomeration of carbonaceous, nuclei mode particles, and typically have diameters between 0.04 - 1 urn with a maximum concentration at about 0.1 um. This class of particle typically consists of solid particles, whose physical characteristics are not affected by small temperature changes. 20 Variability in Particulate Matter Mass Emissions A number of factors affect the nature and mass of P M emitted for a given test point. Kittelson (2001, personal communication) suggested that one very significant factor was engine stabilization time between test modes. Kittleson investigated variability in particle number counts with Abdul-Khalek et al. (1998) and found that high variability in P M emissions for a given test point were a result of engine stabilization times that were too short. This conclusion arose from coefficient of variation ( C O V ) analysis of particle number counts for repeated test points. The highest C O V was associated with mode 6 (100% load, 1600 rpm) of the ISO test, which was executed following mode 5 (10% load, 2600 rpm). The shift from this low-load, high-speed condition to maximum load, medium speed test point seemed to have influenced the P M number count. A longer stabilization time of 15 minutes (compared to the 5-minute original stabilization time) was subsequently proposed so as to adequately condition the entire system for the test point in question. A longer stabilization period prior to P M sampling may allow cylinder wall temperatures and exhaust gas characteristics to stabilize for the selected test point. The same authors suggested that high variability in P M emissions for a given test point were a result of test execution sequence. This conclusion arose from the same coefficient of variation ( C O V ) analysis described above. Kittelson also suggested that another significant source of error was related to flow control in the Sierra BG-1 Micro-Dilution system (personal communication, 2001). The error might occur in taking the difference between two relatively large flows to get the sample flow. Typically the total flow rate of the system is set at 180 L P M , for a dilution ratio of 12. This results in an incoming exhaust flow rate of 15 L P M and a resulting dilution air flow rate of 165 L P M . MacDonald et al. (1980) also found that day-to-day diesel engine P M variations were significantly large relative to the effects of changing D R or diluted exhaust temperature on P M loading. This finding emphasized the importance of running comparative P M loading trials on the same day in order to mitigate this source of variability. 21 Experimental Procedure 3.1 Overview In order to measure exhaust particulate matter, the engine in question must be coupled with a system that dilutes the exhaust stream with filtered ambient air. The function of this dilution process is to simulate reactions that occur between engine exhaust P M and ambient air, so that P M emissions measured from test engines represent those present in "real life". For this study, the engine in question was a Cummins ISX 400 engine, and the dilution system was the Sierra BG-1 Micro-Dilution system. Both are described in more detail in the following sections. The gravimetric method of collecting and measuring exhaust particulate matter emissions was used to collect P M from a sample of the exhaust flow. This method consists of 3 principal stages. The first involves both pre-test conditioning of the filters used to collect the P M mass, as well as calibration of the P M sampling apparatus and engine warm-up. The second stage involves collection of P M on filter media for selected engine and dilution tunnel operating conditions. In light of the dependence of P M emissions on chosen dilution ratio and filter temperature, all tests were run at a set dilution ratio of 12 and the temperature at the sample filter was monitored during testing to ensure that it did not surpass 5 2 ° C at any time. The indicated dilution ratio of 12 used for testing was chosen because Macdonald et al. (1980, 1984) and Kayes and Hochgreb (1998) suggested that the maximum P M mass will be measured at D R between 10 and 30, and that P M mass is least sensitive to dilution ratio at around DR=15. A slightly smaller D R of 12 was thus chosen to minimize P M sampling time required for adequate mass accumulation on the filters. The final post-test procedure involves conditioning and weighing of the P M filters as well as off-site analysis to determine the contribution of lubricating oil and unburned diesel pilot fuel to the particulate V O F . 3.2 Apparatus - Cummins ISX 400 Engine The engine used in this work is a Cummins ISX 400 heavy-duty truck engine. It is a direct injection type and has a displacement of 15 L . Engine configuration data are shown in table 3.2.1 on the following page: 22 Table 3.2.1 Cummins ISX 400 Engine Configuration Data Diesel fuelling HPDI Modifications Engine Type Cummins ISX 400 Bore 137 mm Stroke 169 mm Number of Cylinders 6 Displacement 14.95 L Combustion Chamber Type Toroidal Bowl Compression Ratio 19:1 Number of Valves/Cylinder 4 Injection Type Direct Injection Fuel Injection Pump Cummins CAPS Pump Injection Pressure (Typical) Up to 200 MPa CNG=19MPa Pilot = 78 MPa Fuel Injection Nozzle Cummins EUI Westport HPDI injectors Maximum Power Output 298 kW@ 1800 rpm Rated Torque 1965 Nm@ 1200 rpm The speed and load of the Cummins engine were controlled independently (fixed speed, varied load via throttle position adjustment) by a Digalog dynamometer and a fuel control system. A back pressure regulator on the exhaust is used to create a back pressure in the system that simulates the presence of a vehicle exhaust system. Filtered lubricating oil (Valvoline S A E 15 W-40) was delivered to the engine via an external pump. The oil temperature was maintained at 8 0 ° C , and the engine coolant temperature maintained at 100°C. Cylinder pressure measurements are obtained using Kistler 6071 water-cooled high-temperature pressure transducers mounted in each of cylinders 1 and 6, with the remaining 4 cylinders instrumented with P C B 112 B11 pressure sensors. The start of combustion (SOC) and end of combustion are determined from real-time A V L heat release analysis of cylinder pressure data. The Westport Intelligent Driver Module (IDM) controls the commanded start of injection (SOI) for diesel and gas injection. A B E I Model H25 incremental optical crank-angle encoder determined the engine crankshaft position. 23 A heated probe was mounted in the exhaust line to sample gaseous emissions. The Horiba 7100 D E G R emissions bench consisted of a chemiluminescent oxides of nitrogen (NOx) analyzer, a flame ionization total hydrocarbon (tHC) detector, both non-dispersive infrared carbon monoxide (CO) and carbon dioxide (CO2) detectors, and a paramagnetic 0 2 analyser. For every test point, emissions, engine speed, power, fuel consumption and cylinder temperature and pressure data were logged at 1 H z over a 3-minute period. Averaged values of selected data were used to calculate composite test cycle emission rates. 3.3 Engine Test Protocol The A V L 8-mode steady-state test protocol was used to obtain P M emissions data, as this procedure is used to predict F T P transient testing results. The engine operating loads and speeds for the A V L 8-mode test are listed below in table 3.3.2 (based on Cummins ISX 400 1965 N m torque curve): Table 3.3.2 AVL 8-Mode Steady-State Test Engine Operating Conditions Mode Load (%) Speed (rpm) Weighting Factor 1 0* 600 35.01 2 25 732 6.34 3 63 852 2.91 4 84 984 3.34 5 18 1800 8.40 6 40 1740 10.45 7 69 1740 10.21 8 95 1668 7.34 * residual torque on the engine dynamometer ~ 65 Nm 24 T o calculate composite brake specific emissions (BSE) over the entire A V L 8-mode test cycle, both the modal emission rate and modal power are multiplied by their respective weighting factors, and these modal values summed over all 8 modes of the test, as in the following equation: mod eS ^ EmissionRatexWF BSE mod el mod eS ^ BrakePowerxWF mod e\ where the emission rate is measured in g/hr, and the engine power in brake horsepower (bh P ) . 3.4 Apparatus - Sierra BG-1 Micro Dilution System A dedicated probe was used to sample the particulate matter emissions. The exhaust sample drawn by the probe was diluted in a Sierra BG-1 micro-dilution tunnel using filtered and dried air. A constant sample flow rate of 180 litres per minute ( L P M ) was maintained across two E M F A B T X 4 0 H I 2 0 - W W filters in series used to collect the exhaust particulate. The ratio of dilution air+exhaust to exhaust flowrate (dilution ratio) was kept constant at a value of 12. Maintaining this dilution ratio throughout all modes of the A V L 8-mode test ensured that the filter temperature did not exceed 5 2 ° C , as specified by the E P A . Sampling occurred over a 10-minute period, to ensure that enough P M was collected on the filters for subsequent chemical species analysis. For further information regarding filter conditioning and weighing procedures, please refer to Appendix B . 3.5 Pre-Sampllng Procedures: Particulate Matter Emissions Testing 3.5.1 Filter Conditioning and Weighing A t least 48 hours prior to the test, the blank P M filters are placed in a relative humidity (RH) and temperature-controlled chamber for at least 24 hours. The chamber is held at 4 5 ± 8 % R H and 2 0 ± 5 ° C . Once the filters have remained in the chamber for the required minimum 24 hours, they are ready to be weighed. Before weighing the actual filters, the micro-balance on which they are weighed is autocalibrated, tared and externally calibrated by weighing both a 200 and 500 mg weight 3 times. The standard deviations of the mass measurements for the 200 and 500 mg weights are 0.003 and 0.002 mg, or 25 0.0015% and 0.0004% of their typical masses, respectively. In addition, two conditioning chamber control filters (housed in Petri dishes and permanently stored in the chamber) are weighed. The standard deviations (and the standard deviation as a percentage of typical P M fdter loading mass) for these control blank filters #1 and #2 were 0.005 mg (0.84%) and 0.004 (0.67%), respectively. A l l calibration masses, control filter masses, R H and temperature data are recorded in a daily filter weighing control log file. Calibration masses and filters must be handled with tweezers at all times. Each blank P M filter is placed in its own labelled Petri dish and weighed 3 times, and the values recorded in a data file. A t any time, the balance operator can choose to weigh filters and calibration masses 4 or 5 times if the balance readings deviate by more than ± 0 . 0 1 0 p.g. 3.5.2 Sierra BG-1 Sampler Calibration 3.5.2.1 System Leak Check A leak test is performed whereby the system is pressurized to 20 psi and then the loss in system pressure is monitored over 3 minutes. The allowable pressure loss over this interval is 5 psi. If this condition is violated, all system connections are leak-checked and tightened if necessary, and the leak test subsequently re-executed. 3.5.2.2 Flowmeter Calibration The Sierra system has two key flowmeters: one to measure dilution air flow and one to measure dilution air + exhaust flow (hereafter referred to as "total flow"). The dilution air flowmeter is deemed the "master" flowmeter and the total flowmeter the "slave" in the calibration scheme. The master flowmeter flows according to a set of discrete flow points over a user-specified range, and the slave mirrors the flow measured by the master. This allows for minimal relative error between flow measurements during system operation, as the flow values determine the dilution ratio and its associated accuracy. 3.5.3 Engine Stabilization The engine is run at idle until the coolant temperature reaches 8 0 ° C . It is then brought to the desired speed/load condition. The engine operator subsequently waits until cylinder exhaust temperatures are acceptably uniform prior to collecting the P M exhaust samples. The time required for stabilization of these temperatures is generally 5 minutes. 26 3.6 Test Sampling Procedures: Particulate Matter Emissions Testing A t the beginning of each test, two blank filters are placed in the filter holder, the filter holder clamped shut, re-opened, and the set of filters removed and returned to their respective Petri dishes. These travel blanks are used to determine any gain/loss in filter mass associated with inevitable filter handling. In addition to this, two dummy filters are placed in the filter holder and the sampler activated to simulate regular operation for two minutes. The purpose of this exercise is to purge all system channels of residual contaminants that might bias results of the first test point of the day. Following this, once the engine has stabilized for the desired test point conditions, the dilution ratio is set to 12, sampling time to 10 minutes, and total flowrate 180 litres/minute ( L P M ) . T w o filters are placed in the filter holder, which is then placed in-line with the sampling system. The sampling is subsequently begun, and during the test, the pressure drop across the sample filters, temperature at the filters, and ambient test cell temperature monitored every three minutes over the test period. Once the test period has ended, final temperature, differential pressure across the filters, and final mass flow data are recorded, the filters removed from the filter holder and placed in their respective Petri dishes. 3.7 Post-Sampling Procedures: Particulate Matter Emissions Testing The Petri dishes containing PM-loaded filters are placed in the conditioning chamber and re-weighed 24 hours from this event in a way similar to that described in section 3.5.1. This final filter mass minus the initial filter mass from the initial weighing equals the mass of P M emitted over the test period. 3.8 Gaseous Pollutant Monitoring Emissions of nitrogen oxides (NOx), carbon monoxide (CO), total hydrocarbons (tHC), methane (CH4) and carbon monoxide ( C 0 2 ) are also measured using a Horiba 7100-D E G R emissions analyzer. Data for each test point were logged for 180 seconds of the 10-minute P M collection period. 27 3.9 Procedure for Determining the Unburned Pilot Fuel-Like and Lubricating Oil-Like Components of the Particulate Volatile Organic Fraction (VOF) The procedure described below was developed by Cuthbertson and Shore (1993) at Ricardo Consulting Engineers ( U K ) , and was used to analyse H P D I exhaust P M fdters to determine relative amounts of fuel - and oil-like particulate V O F components. The origins of the particulate V O F were determined by using gas chromatography/flame ionization detection (GC/FID) to generate chromatographic spectra of the diesel pilot fuel and lubricating oil used in the engine at the time of P M sampling. Initially, engine oil and diesel fuel samples were obtained following P M sampling. Then the pilot diesel fuel sample prepared prior to injecting it into the G C / F I D ; it was heated to 3 2 5 ° C , as Cartellieri and Tritthart (1984) observed that only the heavier components (i.e. least volatile components) of the diesel fuel were retained during filtration of exhaust. Then the individual oil and prepared fuel samples were injected into the G C / F I D one at a time in order to obtain a chromatographic spectrum of each sample's composition. A n example o f the spectrum obtained for the prepared fuel and oil are given in figures 3.9b and 3.9c below. Following this, a section of a PM-loaded filter was inserted into a pre-chamber to the G C / F I D , and heated to 3 2 5 ° C so that the volatile P M components could be devolved from the sample. This fraction was then run though the G C / F I D to obtain a chromatographic spectrum, similar to the one shown in Figure 3.9a below. Figures 3.9 a,b,c: Typical Chromatographic Spectra for Particulate VOF and Its Components (from Cuthbertson and Shore, 1988) 0 10 Minutes 20 28 The particulate V O F spectrum characteristically has two distinct humps, as seen in Figure 3.9a on the previous page. This spectrum was then compared to the two spectra for unburned fuel and oil, so that a semi-quantitative assessment of the fuel-like and oil-like V O F components could be obtained. The calculation of these quantities was based on the assumption that all hydrocarbons eluted in the G C before a selected cut-off time (determined visually from the retention time at which the trough between the two humps in Figure 3.9a occurred) were fuel-related and those eluting after the cut-off time were oil-related. The procedure used to determine the amounts of fuel-like and oil-like P M components is further described in Appendix E of this report. The total V O F was subtracted from the each modal P M mass to give a rough estimate of the soot, nitrate and sulphate ions, and other species present in the exhaust P M . 29 4.0 Results and Discussion 4.1 Overview: HPDI Test Matrix This section contains results from 5 different studies carried out to determine the composition, repeatability of, and effect of varying selected engine operating parameters on, HPDI P M emissions (Table 4.1.1). P M emissions will be correlated with pressure trace, heat release, C O , N O x and fuel flow data where such relationships help to explain trends in the experimental data. Table 4.1.1 HPDI Test Matrix Test Description Analyses Procedure HPDI "4 comers" repeatability (Study A) Exhaust gases, gravimetric filter analysis Modes 1,4,5,8 repeated 5x each Gas Rail Pressure (GRP) = 19 MPa Timing settings*: Mode 1: -2°BTDC Mode 4: -6°BTDC Mode 5: - 23° BTDC Mode 8: - 17° BTDC Repeatability Testing HPDI AVL 8-Mode Composite AVL 8-mode tests #1,2 (Study B) Exhaust gases, gravimetric filter analysis, VOF composition analysis (modes 1,4,5,8) GRP = 19 MPa Timing Study Exhaust gases, gravimetric filter analysis GRP = 19 MPa Modes 1,4,5,8 run at 4 different timings: Timing settings (most advanced to most retarded): Mode 1: -16°, -14° BTDC, +3° ATDC M o d e 4 : - l l 0 , - 8 0 , - l ° B T D C Mode 5:-36°,-30°, -16.5° BTDC Mode 8: -29°, -18°, -15° BTDC Pressure Study exhaust gases, gravimetric filter analysis Mode 1: P,ow=10 MPa, Timing = 1°BTDC Mode 4: P|0W=17 MPa, Timing = 6°BTDC Mode5:P,ow=13 MPa, Timing = 23°BTDC Mode 8: P|OW=23 MPa, Timing = 16°BTDC Phigh= 25 MPa for all modes Pilot Fuel Quantity Study exhaust gases, gravimetric filter analysis Mode 2 Pilot Pulse Width (PW) = 0.48, 0.65, 1.1 ms (fixed injection timing, injection pressure) *timing setting = start of injection (commanded) for pilot diesel fuel 30 4.2 Repeatability of PM Measurements (Study A) The repeatability of the results was described by calculating the "coefficient of variation" ( C O V ) of the P M emission rate. The C O V is the standard deviation divided by the mean of all repeated values, expressed as a percentage. Details of the test results are contained in Table 4.2. For this test, mode 1 was run continuously for approximately 1 hour, during which time P M was sampled 5 times, each sampling period lasting 10 minutes. The same procedure was followed for modes 4, 5, and 8. It was found that mode 4 had the highest C O V of 9% (worst repeatability) followed by: mode 8 (6%), mode 1 (5%), with mode 5 having the lowest C O V (4%) and hence best repeatability. The mode 1 and 5 C O V s were calculated for a sample size of n=5, whereas the mode 4 and 8 C O V s were based on a sample size n=4 (1 outlier removed based on visual scatterplot data analysis). Abdul-Khalek et al. (1998) calculated the C O V based on a sample size of n=3 for each mode of the ISO 11-mode test. These C O V data agree reasonably well with the results from Study A , as seen in a comparison of column (e) and (g) data (Table 4.2 below). Table 4.2. Coefficients of Variation for PM Emissions Data: Modes 1,4,5,8 of A V L 8-Mode Test and Comparative Data from Equivalent ISO 11-Mode Test Points AVL Equivalent Engine Engine (e) (f) (g) Test ISO Test Speed Load COV COV COV* Mode Mode AVL/(ISO) AVL/(ISO) (Repeatability (Repeatability (ISO (rpm) (%) Study A) Study B) Test, (%) (%) Abdul-Khalek et al.) (%) 1 11 600 (750) 0(0) 5 10 6 4 - 984 84 9 10 -5 9 1800(1600) 18(25) 4 20 4 8 6 1668(1600) 95 (100) 6 15 8 * COV of PM volume concentration data (volume concentration measured in pm /cm ) 31 4.3 Same-Day Repeatability of A VL 8-Mode Test (Study B) The A V L 8-mode test was repeated twice in the same day to determine variability of P M emissions and effect of test sequence on P M emissions. Modes 1-8 of A V L test #1 were run in ascending order in the morning, and the same test sequence repeated in the afternoon of the same day. The results suggested that the test sequence possibly affected the repeatability of the modal P M results, based on data shown in Table 4.2 for studies A and B . The C O V data in column (f) are higher than those in column (e). The results in (f) were obtained by running modes 1-8 of the A V L test in numerical sequence, introducing the possibility that the test sequence would affect modal P M results. This effect is most pronounced for the mode 5 C O V value, which jumps from 4% in (e) to 20% in (f)- One possible explanation for this is that mode 5 sampling follows mode 4 (high load/low speed) sampling, where mode 4 is known to produce high amounts of P M on a mass basis. It is possible that some of the particulate matter associated with mode 4 is sticking to the walls of the sampling system and is only re-entrained once mode 5 operation has begun. A n alternate explanation is that the mode 5 P M emissions may be most affected by the variation in P M emissions associated with running like test points at different times in the same day. The cycle P M emissions for A V L test #1 and test #2 run in study B were 0.048 and 0.056 g/bhp.hr, respectively, giving a C O V of 11.5% for the cycle P M emission rate, which uses a weighted sum of all modal emissions. By contrast, the results in (e) were obtained by running the engine at each of modes 1,4,5 and 8 and sampling for P M five times during each load/speed condition, essentially removing the variability introduced by running unlike test points back-to-back. This explains the general trend of lower C O V s for Study A repeatability results when compared to those obtained for Study B (mode 8 excepted). Other gaseous emissions had the following coefficients of variation based on the two A V L tests results from study B : N O x : 1.2%, C O 2 0 . 8 % , C H 4 : 3.9%, n m H C : 5.2%, C O : 3.0%. Variations in particulate and gaseous emissions from day-to-day (assessed by 32 comparing study A results for modes 1,4,5,8 to study B results for the same modes) are significantly greater than the variations within Study A or Study B (please refer to tables D . A and D . B of Appendix D) , which suggests that engine variability does affect the repeatability of P M emissions. Injection Tinting Study Smith et al. (1998) showed that advancing timing reduced P M emissions in diesel-fuelled CI engines. The results obtained for this study showed that this was not the case for all H P D I tests: in some cases, for a given set of injection parameters, advanced injection timing did not necessarily reduce P M emissions. The timing is referred to as commanded pilot start of injection (SOI) in the subsequent figures and in table 4.1.1, and differs from the actual beginning of combustion ( B O C ) due to mechanical and hydraulic delays associated with the fuel injection process. For the H P D I system, advancing injection timing generally increased cylinder temperature, but also affected injected gas dynamics and air/fuel mixing. For some timing scenarios, the cylinder temperature and resulting soot oxidation rate dominated the system's tendency to produce P M , and in other cases, poor gas dynamics and fuel/air mixing affected combustion efficiency, usually resulting in higher P M emissions. C O emissions were analysed alongside modal P M emissions trends, as C O may serve as an indication of poor in-cylinder air utilization resulting from poor mixing. N O x emissions were also plotted, as changing injection timing typically affects these emissions as well. The difference between emissions levels for the various injection timing cases was statistically significant unless otherwise noted in the text. Mode 1 (idle) results (calculated in g/bhp.hr based on residual torque of -65 N m associated with the engine dynamometer) showed that the P M emissions increased as timing was further retarded, a result matching that referred to above in the work of Smith et al (1998). The typical P M - N O x tradeoff also occurred; the N O x tended to decrease as timing was further retarded; this occurs because cylinder temperatures tend to decrease as timing is retarded off, a condition favouring lower N O x emissions and subsequently higher P M emissions. The C O emissions trend resembled the P M trend and increased as timing was retarded. 33 Figure 4.4a: Effect of Injection Timing on Particulate Matter, Nitrogen Oxide, and Carbon Monoxide Emissions: Mode 1 of A V L 8-Mode Test Timing Study: PM, NOx, and CO Emissions: Mode 1 (idle) of AVL 8-Mode Test CO — , 8 3 £ u - o Ul 30 25 20 15 10 5 0 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1 1 I s s 1 -20 -15 -10 -5 Pilot SOI (degrees) I*NOx ACO • P M I * the difference in PM and CO emissions for each of the -16° and-14" BTDC SOI cases was not statistically significant for either of the 2 emissions The mode 4 results were opposite those for mode 1, and also contradicted the trend described by Smith et al. (1998). P M emissions were seen to decrease as injection timing was retarded (Fig. 4.4b). Fig. 4.4b: Effect of Injection Timing on Particulate Matter, Nitrogen Oxide, and Carbon Monoxide Emissions: Mode 4 of AVL 8-Mode Test Timing Study: PM, NOx, and CO Emissions: Mode 4 (Low Speed, High Load) of AVL 8-Mode Test c o l | LLI - O X 3? o c o F 6 -12 -10 Pilot SOI (degrees) • NOx A CO • P M 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 a. _ •i ^ This was an unexpected but promising result, showing low P M emissions for a high-load, high fuel/air ratio, retarded timing test point. Another interesting result was that NOx emissions remained constant for all three timing settings. This may indicate a possible 34 combustion problem (NOx typically decreases as timing is retarded) or the possibility of using retarded timing at high-load operation to lower both PM and NOx emissions. This trend can be explained by examining pressure trace as a function of crank angle data (Fig.4.4c) and CO emissions. In Fig. 4.4c, the most advanced timing case corresponds to the highest peak cylinder pressure (and thus temperature), with the opposite being true for the most retarded timing case. It seems in this case that injection of the natural gas into a higher cylinder pressure leads to poorer mixing between gas and air, resulting in poorer combustion and air utilization and unburned fuel in the cylinder, which manifests itself as higher PM emissions along with unoxidized soot. CO and PM emissions were consistently highest over time for the most advanced timing case, which suggested poor mixing between injected fuel and air at this condition (Fig. 4.4b). Heywood (1988) suggested that high CO levels were a result of either fuel-lean or fuel-rich zones in the cylinder, which occur regularly in a diffusion flame. This phenomenon was not consistently observed at all modes of the A V L test, but provided an interesting result for brief discussion. F i g . 4.4c: C y l i n d e r P r e s s u r e T r a c e s f o r A d v a n c e d , Intermediate a n d R e t a r d e d Inject ion T i m i n g S c e n a r i o s : M o d e 4 o f A V L 8 - M o d e T e s t Cylinder Pressure Traces: Injection Timing Study -Mode 4 (Low Speed, High Load) 3 | r50n 1 -60 -40 -20 0 20 40 60 crank angle (degrees) —•— SOI=11 deg BTDC ~ — SOI=8 deg BTDC SOI=1 deg BTDC * SOI = pilot start of injection The mode 5 results in Fig. 4.4d show no clear trend between PM emissions and timing for all three timing cases; initially, a trend of increasing PM emissions is seen as timing is retarded in going from the most advanced (-36° BTDC) to the intermediate (-30° BTDC) timing, similar to the result from the work of Smith et al. (1998). However, the retarded injection result is anomalous, as it shows a decrease in PM emissions from the intermediate timing case. It is, however, higher than the PM emission rate for the most advanced timing, so it can be said that PM emissions do increase for this mode as timing is further retarded. The NOx trend in Fig 4.4d is typical, with NOx decreasing with 35 retarded timing. There is no match in the trend between C O results and P M results for all three timings, but the C O and P M for the most retarded timing case are higher than the respective C O and P M for the most advanced timing case. F i g . 4 .4d : E f f e c t o f Inject ion T i m i n g o n P a r t i c u l a t e M a t t e r , N i t r o g e n O x i d e , a n d C a r b o n M o n o x i d e E m i s s i o n s : M o d e 5 o f A V L 8 - M o d e T e s t Timing Study: PM, NOx, and C O Emissions: Mode 5 (High Speed, Low Load) of AVL 8-Mode Test c o ~ l l HI -Q X <5 S I - ra O OC O 14 12 10 8 6 4 2 0 II • 4 A • A -40 -30 -20 -10 0 Pilot SOI (degrees) iNOx A C O • P M 10 0.25 0.2 03 5 o | l i 0.1 0.05 £ 0 20 The P M emissions for mode 8 exhibited a trend similar to that described for mode 5 (Fig. 4.4e). Initially, a trend of increasing P M emissions is seen as timing is retarded in going from the most advanced (-29° B T D C ) to the intermediate ( -18° B T D C ) timing, similar to the result from the work of Smith et al. (1998). However, the retarded injection result is anomalous, as it shows a decrease in P M emissions from both the advanced and intermediate timing case. The P M / N O x tradeoff is also typical in going from the most advanced case to the intermediate timing case; as timing is retarded, P M increases and N O x decreases. Finally, there is no correlation between the C O and P M emissions trends as timing is further retarded. 36 Fig. 4.4e: Effect of Injection Timing on Particulate Matter, Nitrogen Oxide, and Carbon Monoxide Emissions: Mode 8 of AVL 8-Mode Test Timing Study: PM, NOx, and C O Emissions: Mode 8 (High Speed, High Load) of AVL 8-Mode Test . S r . , UJ 5 . X O ) O O * o 0.5 I -35 -30 -25 -20 -15 Pilot SOI (degrees) I NOx A CO * P M -10 0.06 0.05 B (0 E I 0 0 4 § I 0.03 8 •= 0.02 iS ~ 0.01 Q-Difference between NOx emissions for the -18" and-15" BTDC is not statistically significant 4.5 Pressure Study: Modes 1,4,5,8 of AVL 8-Mode Test The effect of injection pressure on P M emissions is investigated. The low and high pressure values discussed in this section refer to the gas rail pressure (GRP) of the natural gas. For the HPDI system, injection pressure of diesel and natural gas must be balanced and cannot be independently controlled; thus an increase in GRP for the natural gas also means an increase in the diesel pilot injection pressure. The natural gas rail pressure was varied for 2 cases: a low pressure case based on the minimum condition for sonic flow through the injector nozzle (value dependent on engine speed/load operation) and a maximum pressure value of 25 MPa independent of engine load and speed. 4.5.1 Effect of Higher Injection Pressure on PM Emissions: Mode 1 (Low Speed, Low Load) and 5 (High Speed, Low Load) of A V L 8-Mode Test It was found that, at modes 1 and 5 (Fig. 4.5a), higher injection pressure resulted in lower P M emissions, similar to results obtained by Yokota et al. (1991) using diesel fuel injection. The resulting decreases seen in P M emissions for modes 1 and 5 were statistically significant (for a 95% confidence interval) as they were outside of the 3-a error limits of the P M measurements, as seen in Fig. 4.5a. 37 Fig 4.5a: Effect of Gas Rail Pressure on PM Emissions: Modes 1,5: AVL 8-Mode Test Effect of G a s Rail P r e s s u r e o n Part iculate E m i s s i o n s : M o d e s 1 (Idle) a n d 5 (High S p e e d , L o w 1 Mode 5 P l o w pressure B n j g h pressure (25 MPa) Higher injection pressures result in better fuel atomization, in-cylinder penetration and air/fuel mixing of diesel pilot. Yokota et al. (1991) suggested that all of these factors contribute to higher soot oxidation and lower overall P M formation from diesel fuel sprays. A t model , the diesel pilot fraction of the total fuel flow was 3.6% and 25% for the low and high G R P cases, respectively. A t mode 5, the diesel pilot fraction of the total fuel flow was 10% and 16.5% for the low and high G R P cases, respectively. Higher diesel fuel quantity injected at a higher G R P for both modes 1 and 5 also resulted in lower P M emissions. This is a very interesting result, given that increasing diesel pilot fuel is suspected to increase P M emissions. In F i g 4.5b, a plot of differential heat release shows a higher combustion rate of natural gas (indicated in the steepness of the positive slope of the curve) leading to higher peak cylinder temperature for the mode 5 high-pressure case. Higher peak cylinder temperatures and heat release occurring earlier in the cycle are both ideal for soot oxidation, and may explain the lower P M emissions achieved with the high injection pressure strategy for this mode. 38 Fig 4.5b: Differential Heat Release Plot for Low and High Injection Pressure Cases: Mode 5 of AVL 8-Mode Test Differential Heat Release Profiles: Gas Rail Pressure Study Mode 5 ( High Speed, Low Load) I 86-1 r -crank angle (degrees) •P=13 MPa P=25 MPa The resulting decreases seen in P M emissions for modes 1 and 5 were statistically significant (for a 95% confidence interval) as they were outside of the 3-a error limits of the P M measurements, as seen in Fig. 4.5a. 4.5.2 Effect of Higher Injection Pressure on PM Emissions: Mode 4 (Low Speed, High Load) and 8 (High Speed, High Load) of A V L 8-Mode Test The difference in P M emissions due to increasing injection pressure was not statistically significant (based on statistical parameters calculated for modes 4 and 8, respectively, in Study A ) , as shown by the error bars in Fig. 4.5.2a. Thus it is not possible to claim that the P M emissions increased as suggested by the histogram. Fig. 4.5.2a. Effect of Gas Rail Pressure on PM Emissions: Modes 4,8: A V L 8-Mode Test Effect of Gas Rail Pressure on Particulate Emissions: Modes 4 (High Load, Low Speed) and 8 (High Load, High Speed) n Mode Plow pressure Bhigh pressure (25 MPa) 39 The 2 M P a pressure difference between high and low injection pressures for Mode 8 is possibly not large enough to yield a significant reduction in the P M emission rate, as is seen in the heat release curves for each of these mode 8 pressure cases (Fig. 4.5.2b). Unfortunately the physical limitations of the engine and injectors dictated the 23 M P a low G R P case and 25 M P a high G R P case, resulting in a small pressure difference between the two cases possibly not significant enough to affect mode 8 P M emissions. F i g . 4.5.2b: D i f f e r e n t i a l H e a t Release Plot f o r L o w a n d H i g h Inject ion Pressure C a s e s : M o d e 8 o f A V L 8 - M o d e T e s t Differential Heat Release Profiles: Gas Rail Pressure Study - Mode 8 (High Speed, High Load) CD 10 i l CO O) 0) C a. ?5 < s I 1 2 150 100 -60 -40 -20 20 40 60 crank angle (degrees) •P=23MPa •P=25MPa Though the injection pressure study results for modes 4 and 8 were found to be statistically insignificant, an attempt will be made to explain what might have caused an increase in P M emissions for higher injection pressures at these modes. Analysis of other indicator emissions (such as C O ) suggests that increasing G R P of natural gas may actually result in poorer air/fuel mixing and subsequently high C O emissions (as seen in Figs. 4.5d and 4.5e) and higher P M emissions. High C O trends are often accompanied by a high P M trend in the case of fuel-rich combustion (Heywood, 1988). Fig. 4.5d: Carbon Monoxide Emissions: Mode 4 of AVL 8-Mode Test 850 ? 650 Q. Q. 450 o o 250 50 Pressure Study: CO vs. Time Plots Mode 4: Low Speed, High Load 50 100 Time (s) 150 200 •P=17MPa •P=25MPa Fig. 4.5e: Carbon Monoxide Emissions Log for Mode 8 of A V L 8-Mode Test Pressure Study: C O vs. Time Plots Mode 8: High Speed, High Load 300 250 Q. 200 8 1 5 0 100 50 50 100 Time (s) 150 200 M8 P=23MPa •M8 P=25 MPa Poor air-natural gas fuel mixing is possibly a result of the high pressure natural gas spray prematurely reaching the cylinder wall during and following injection, not properly utilizing the air that is swirling in the piston bowl. Heywood (1988) suggested that such fuel-rich zones located radially along the flame deck are probable regions of high-CO (and probably P M ) production. More information on the relative fluid dynamics of the natural gas and air occurring in the regions of the piston bowl and cylinder head would help to confirm this theory. 41 The results of this investigation introduce an important concept: increasing injection pressure of natural gas in the H P D I combustion system may only result in lower P M emissions if the air and fuel mixes to produce conditions favouring efficient combustion. 4.6 PM Composition: Unburned Pilot Fuel-Like and Lubricating Oil-Like Components of Particulate Matter Volatile Organic Fraction (VOF) Exhaust particulate matter contains a number of different species, namely solid carbon (soot), volatile organic species, sulphate and nitrate ions, and metallic elements derived from lubricating oil/fuel additives or general engine wear. The scope of this composition study was limited to determining the V O F of the particulate matter and the relative amounts of unburned fuel-like and lubricating oil-like components in the V O F . The method use to perform this analysis is described in section 3.9 of this report. This method was used on filters for P M emissions from modes 1,4,5, and 8 of two identical A V L 8-mode tests carried out at different times on the same day. The tests will be referred to herein as A V L test #1 and A V L test #2. The split between unburned fuel and lubricating oil in the particulate V O F was affected by the engine speed/load operating condition. A s seen in Figures 4.6d and e on the following page, the lubricating oil contribution to the P M V O F was highest for the two high speed modes (modes 5 and 8) and much smaller for idle and mode 4 low- and medium-speed conditions. Meyer et al. (1980) found that the lubricating oil contribution to the organic fraction was greater for high-speed engine operating conditions, similar to the results obtained in this analysis. 42 Fig. 4.6d: Particulate Matter Component Emission Rates, (g/hr): Repeatability Study: Modes 1,4,5,8 of AVL 8-Mode Test #1 Particulate Matter Component Breakdown: Fuel-Like, Oil-Like Volatile Organic Fraction and Soot/Other Components Repeatability Study: AVL 8-Mode Test #1 1 4 5 8 AVL Test Mode |^fuel-like particulate VOF Moil-like particulate VOF Msoot and other PM components * phi - equivalence ratio Fig. 4.6e: Particulate Matter Component Emission Rates, (g/hr): Repeatability Study: Modes 1,4,5,8 of A V L 8-Mode Test #2 Particulate Matter Component Breakdown: Fuel-Like, Oil-Like Volatile Organic Fraction and Soot/Other Components Repeatability Study: AVL 8-Mode Test #2 T 3 5 E AVL Test Mode Hfuel-like particulate VOF Moil-like particulate VOF Msoot and other PM components The lubricating oil contribution to particulate V O F was higher for all modes in A V L test #2 (vertically striped portion in Figures 4.6d and e above) than in test #1. This may be a result of the engine oil running slightly "thicker" in the morning test (#1), whereas it may have thinned out in the afternoon test (#2) enough to cause it to migrate past valve guides and piston rings in to the cylinder, ultimately ending up in the exhaust P M . 43 A t mode 5, the unburned pilot diesel fuel contributed more to the particulate V O F than did the lubricating oil , possibly due to less stable combustion at this low load, high speed point. The poorer stability of combustion at this mode was verified by calculating the C O V of brake mean effective pressure ( B M E P ) for this mode. The mode 5 C O V of B M E P was 2.4%, which was high compared to the C O V of B M E P for modes 1,4, and 8, which were 0.7, 0.3 and 0.1%, respectively. Conversely, the unburned fuel portion of the particulate V O F was lower than the lubricating oil V O F component for mode 8. The low C O V B M E P for this high speed, high load point suggested more stable combustion, and thus lower amounts of unburned pilot fuel were present in the particulate V O F . In this study, the V O F portion of total P M mass emissions from the selected A V L test modes was between 34-58%). Wedekind et al. (2000) found that the V O F contribution to total P M emissions for a EURO-III certified HE) diesel engine was 29% for the entire European Stationary Cycle (ESC) test. The former result is not surprising as Kittelson (2001, personal communication) suggested that newer alternative fuel engine P M would have a higher organic fraction than P M from conventional Diesel engines. The results in Fig. 4.6d and e were obtained by determining the unburned fuel and lubricating oil V O F on each filter in a set comprised of a primary and a backup filter. The filters used in this study were those from repeatability study B . A n interesting aspect of this analysis was that the secondary filter often contained negligible amounts of lubricating oil V O F , but always contained a significant, if not greater amount of unburned fuel than its primary counterpart for the same test point. Visual examination of the primary filter showed a greyish stain from the collected particulate matter, whereas the secondary filter showed no stain whatsoever. Though there was no visual evidence of any P M loading on this secondary filter, both the gravimetric weighing and V O F analysis showed that there was significant particulate matter on this filter resembling unburned pilot fuel. 4.7 Pilot Fuel Quantity Study: Mode 2 of A VL 8-Mode Test In order to determine how varying the injected pilot diesel fuel quantity would affect P M mass emissions, the engine was run at mode 2 (25% load, 750 rpm - low speed) of the A V L 8-Mode test at 3 different pilot fuel injection durations. The variation in pilot fuel injection duration is referred to as pilot pulse width (PW) and is measured in 44 milliseconds. The chosen pilot P W values were 0.48, 0.65 and 1.1 ms, respectively, and the diesel pilot portion of the total fuel flow for each of these P W settings is given in -Table 4.7 below: Table 4.7 Diesel Pilot Fraction of Total Fuel Flow and Natural Gas (CNG) Flow for 3 Different Pilot PW Settings Pulse W i d t h (ms) Pilot Fuel in Total Fuel F low (%) C N G flow (kg/hr) 0.48 8.3 2.32 0.65 24.8 2.36 1.1 35.6 2.28 Increasing the pilot P W from 0.48 ms to 0.65 ms seemed to result in an increase in P M emissions, as per Fig. 4.7a below. However, the P M emission rate decreased when the pilot P W was further increased to 1.1 ms, in contrast to the former result. The maximum difference in P M emission rate from one pilot P W setting to another was 4%, which is less than typical C O V values calculated in Study A for modes l,4,5,or 8 (Table 4.2). This suggests that the difference in P M emission rates for the chosen pilot P W settings may not have been significant. Figure 4.7a: Effect of Varying Pilot Pulse Width on PM Emissions Pilot P u l s e Width S tudy Total (CNG+Diesel) Fue l F low, Diese l F low, P M E m i s s i o n s : M o d e 2 of A V L 8-Mode T e s t 45 5.0 Conclusions The repeatability of particulate matter emissions for H P D I fuelling has been established, as well as the effect of varying injection pressure, timing and pilot diesel fuel quantity on P M emissions. In addition, the unburned fuel-like and oil-like components of the P M have been quantified for selected test points chosen to cover the entire range of engine operating conditions. Testing for each of modes 1,4, 5 and 8 repeated 5 times each showed that P M emissions were within 10% of their respective means for all modes. Execution of P M testing for the A V L 8-mode test run once in the morning and once in the afternoon of the same day suggested that test sequence and engine warm-up time may affect P M emissions. The repeatability of P M emissions for the same test point run at two separate times of the day varied by 10-20% for modes 1,4,5 and 8. Cycle emissions for the entire A V L 8-mode test had a coefficient of variation of 15%. The suspected sources of variability were: P M deposition/re-entrainment occurring along the walls of the sampling system lines, filter weighing error, and flowmeter reading error within the sampling system. Increasing injection pressure decreased P M emissions for the low load cases (modes 1 and 5) of the A V L 8-mode test, but did not significantly affect P M emissions at the high load modes (4 and 8). Increasing injection pressure at these high-load modes seemed to give rise to poor air-fuel mixing. Therefore using increased injection pressure may be a strategy to reduce P M emissions, but for high load test points and the current engine geometry, a suspected air/fuel mixing problem may be producing higher P M emissions than expected. This mixing problem may result in higher P M emissions, effectively overshadowing the reductions in P M emissions normally achieved using higher injection pressure. The effect of retarding injection timing on P M emissions was not consistent across the range of engine operating conditions characterized by modes 1,4,5, and 8. A t idle (mode 1), retarding timing caused P M emissions to increase, a result similar to the response of a combustion system running solely on diesel fuel. A t modes 4 and 8, retarding timing caused P M emissions to decrease, an opposite trend to that occurring at idle. It was suspected that high cylinder pressures resulting from early pilot fuel injection (advanced timing) or unwanted fuel interaction with cylinder wall and piston surfaces was impeding good mixing between injected natural gas and air for the current engine geometry. A low P M emission rate at this high load, retarded timing test 46 point is actually a positive and interesting result if it can be reproduced in future engine testing. A t this high load-mode, it may be possible to mitigate both P M and N O x emissions using a retarded timing strategy. N o consistent trend was established between retarding timing and P M emissions for mode 5. Based on these results, advancing injection timing was not an effective strategy for reducing P M emissions for all engine load/speed conditions. However, results may be different if the suspected air/fuel mixing problem is resolved, and P M emissions at certain modes can be reduced using advanced injection timing. Finally, P M emissions at several modes appeared to be more sensitive to timing variations than other gaseous emissions such as N O x and C O , an important result to keep in mind during engine certification testing. The composition of the H P D I exhaust P M was analysed for volatile organic fraction ( V O F ) fuel-like and oil-like components using a direct capillary injection of filter-borne P M into a gas chromatograph/flame ionization detector (GC/FID) . The basic assumption of this analysis was that all hydrocarbons eluted in the G C before a selected cut-off time were fuel-related and those eluted after the cut-off time were oil-related. This analysis was carried out on particulate filters for each of modes 1,4,5,8 of two A V L 8-mode tests (#1 and #2) carried out in the morning and afternoon of the same day. The difference between the actual P M emissions and the particulate V O F was classified as " soot, sulfate and nitrate ions, and other components". It was found that the oil-like V O F of the P M was consistently highest at high speed condition (modes 5 and 8) for both of tests #1 and 2. The fuel-like component contribution to P M was the same for each respective mode in comparing the results of tests #1 and 2. Results from the present study for selected A V L test modes determined the V O F portion of total P M mass emissions to be between 34-58% for H P D I combustion. This portion of the P M emissions could possible be reduced by 50%> using an oxidation catalyst. Finally, the injected pilot diesel fuel quantity was varied for 3 distinct settings at mode 2 of the A V L 8-mode test to determine whether or not this would affect P M emissions. The difference in the P M emissions for these three injected quantities was not statistically significant, so no conclusions were drawn regarding the effect of increasing this quantity on P M emissions. 47 6.0 Recommendations Based on the results obtained in the various studies undertaken, a number of recommendations can be made concerning future test procedures and engine operating parameters and component design related to minimizing P M emissions. In order to obtain repeatable modal P M emission rates, longer engine conditioning times should be adopted for respective engine load/speed points, and the P M sampling system purged with heated air. This is to avoid any bias error arising from P M emissions characteristics such as: particle deposition, composition (i.e. % volatile organics, soot and other components) and emission rate characteristics that differ across the various engine operating modes of a test protocol. Varying injection timing, or commanded pilot SOI cannot be used in isolation as a strategy for reducing P M emissions without considering engine geometry. Since the results obtained in this research were not conclusive for all engine load/speed test points, it is possible that P M reductions can be achieved by changing injection timing and engine geometry for chosen engine test points. Tests involving injection pressure swings should be re-executed at a time where the maximum gas rail pressure setting can be significantly higher than the minimum gas rail pressure for a given test point. For the study results shown in this report, the maximum gas rail pressure was restricted to 25 M P a due to component design and material selection constraints. A t this time, it is not recommended that increased injection pressure be used as a strategy to reduce P M emissions for the ISX 400 engine running on H P D I combustion until suspected gas mixing and engine geometry issues are resolved. Further studies on the effect of injected pilot fuel quantity on P M emissions should be run when a reliable mass flow measurement technique for the diesel fuel flow has been integrated into the engine data acquisition system. The reported results gave an empirical "least, medium and high" amounts of injected pilot diesel fuel based on the length of the pilot fuel intensifier stroke, which was measured in milliseconds. Once this measurement is reliable, the P M collected on filters from this future study should be analyzed for unburned fuel contribution to the particulate V O F , in order to determine more conclusively whether or not the higher diesel pilot quantity is related to higher P M emissions. These filters could also be analyzed for elemental carbon, or soot, to see i f the pilot fuel quantity affected this component of the particulate matter. 48 Ideally these experiments would be carried out in an optical engine so that the processes of diesel pilot combustion, subsequent natural gas injection, mixing with intake air and combustion could be monitored. Such visual evidence of gas dynamics might give an indication as to what conditions are giving rise to P M formation. 49 7.0 N o m e n c l a t u r e Bhp brake horsepower B M E P brake mean effective pressure B O C beginning of combustion B S F C brake specific fuel consumption B S N Bosch Smoke Number C F R Code of Federal Regulations C N G compressed natural gas [ C 0 2 ] a carbon dioxide concentration in dilution air [ C 0 2 ] e carbon dioxide concentration in exhaust [ C 0 2 ] d carbon dioxide concentration in diluted exhaust [C0 2 ] a > wet wet- basis carbon dioxide concentration in dilution air [C0 2 ] e , wet wet- basis carbon dioxide concentration in exhaust [C0 2 ]d , wet wet- basis carbon dioxide concentration in diluted exhaust [ C 0 2 ] e d r y dry- basis carbon dioxide concentration in exhaust C O V coefficient of variation C V S constant volume sampling D F I - G C ™ Direct Filter Injection Gas Chromatography D l direct injection D M E dimethyl ether dP differential pressure D R dilution ratio E C elemental carbon (soot) E E C European Economic Community E G A exhaust gas analysis E P A Environmental Protection Agency E S C European Stationary Cycle E U I electronic unit injector FID flame ionization detector F T P Federal Test Procedure G C gas chromatography GDIL dilution air flowrate into dilution tunnel GEDF diluted exhaust flowrate GEXH exhaust sample flowrate G P exhaust sample flowrate GTOT diluted exhaust flowrate into dilution tunnel G R P gas rail pressure [ H 2 0 ] e exhaust water concentration H C hydrocarbon H D heavy-duty HPDI high-pressure direct injection (of natural gas) H P L C high-performance liquid chromatograph IDI indirect injection ISO International Standards Organization L D light-duty L N G liquefied natural gas M F M mass flow measurement m V millivolt N G natural gas n m H C non-methane hydrocarbon N T E not to exceed O C organic carbon P A H polycyclic aromatic hydrocarbon P A C polycyclic aromatic compound P D S partial-flow dilution system P M particulate matter P M 2 5 particulate matter of diameter less than 2.5 microns ppm parts per million (molar concentration) psi pounds per square inch Qair dilution air flow rate Q _ d i f f differential heat release Qexh exhaust flow rate Q _ i n , integrated heat release rpm revolutions per minute R H relative humidity S O F soluble organic fraction SOI start of injection Tamb ambient temperature in engine test cell t H C total hydrocarbon T 1 omega temperature reading downstream of Omega heating elements U S United States V O F volatile organic fraction V O C volatile organic compounds *la molar air flow molar exhaust flow diluted exhaust flow 51 8.0 References Abbass, M . K . , Andrews, G . E . , Kennion, S.J., Williams, P.T. , Bartle, K . D . "The Survivability of Diesel Fuel Components in the Organic Fraction of Particulate Emissions from an IDI Diesel", S A E Technical Paper 910487, (1991). Abdul-Khalek, I.A., Kittleson, D . B . , Brear, F. "Nanoparticle Growth During Dilution and Cooling of Diesel Exhaust: Experimental Investigation and Theoretical Assessment", S A E Technical Paper 2000-01-0515, (2000). Abdul-Khalek, I. A . , Kittleson, D . , Brear, F. "The Influence of Dilution Conditions on Diesel Exhaust Particle Size Distribution Measurements" , S A E Technical Paper 1999-01-1142, (1999). Abdul-Khalek, I.A., Kittleson, D . B . , Graskow, B.R., Wei , Q . , Brear, F. "Diesel Exhaust Particle Size: Measurement Issues and Trends", S A E Technical Paper 980525, (1998). Andrews, G . E . , Abbass, M . K . , Abdul Haleem, S., Shen, Y . , Williams, P.T., Bartle, K . D . "The Determination of Lubricating O i l Fraction of Diesel Particulates", Journal of Aerosol Science, V o l . 22, Suppl. 1, pp. S459-S462, (1991). Cartellieri, W. , Tritthart, P. "Particulate Analysis of Light Duty Diesel Engines (IDI&DI) With Particular Reference to the Lube O i l Particulate Fraction", S A E Technical Paper 840418,(1984). Cuthbertson, R . D . , Shore, P.R. "Direct Capillary Injection Gas Chromatography of Filter-Borne Particulate Emissions from Diesel Engines", Journal of Chromatographic Science, v. 26, pp. 106-112,(1988). Clark, N . N . , Gautam, M . , Lyons. D . W . , Bata, R . M , Wang, W. , Norton, P., Chandler, K . "Natural Gas and Diesel Transit Bus Emissions: Review and Recent Data", S A E Technical Paper 973202, (1997). Driscoll, K . E . , Deyo, K . C . , Carter, J .L. et al. "Effects of Particle Exposure and Particle-Elicited Inflammatory Cells on Mutation in Rat Alveolar Epithelial Cells" , Carcinogenesis, 18:423-430, 1997. Engeljehringer, K . , Schindler, W. "Experiences with a Mini-Dilution System for Engine Homologation and Development", S A E Technical Paper 890789, (1989). Fleisch, T . , McCarthy, C , Basu, A . , Udovich, C , Charbonneau, P., Slodowske, W. , Mikkelsen, S.E. , McCandless, J. " A New Clean Diesel Technology: Demonstration of U L E V Emissions on a Navistar Diesel Engine Fueled with Dimethyl Ether", S A E Technical Paper 950061,(1995). Gomes, P .C.F . , Yates, D . A . "The Influence of Some Engine Operating Parameters on Particulate Emissions", S A E Technical Paper 92222, (1992). 52 Graze, R.R. Jr. "Development of a Miniaturized, Dilution-Based Diesel Engine Particulate Sampling System for Gravimetric Measurement of Particulates", S A E Technical Paper 931190,(1993). Hodgins, K . B . , H i l l , P . G , Ouellette, P., Hung, P. "Directly Injected Natural Gas Fueling of Diesel Engines", S A E Technical Paper 961671, (1996). Heywood, J. B . "Internal Combustion Engine Fundamentals", 1st ed. M c G r a w - H i l l , Toronto, (1988). Huisingh, J., Bradow, R, Jungers, R. et al. "Application of Bioassay to the Characterization of Diesel Particle Emissions", (proceedings from Feb. 1978 Symposium on Application of Short-Term Bioassays in the Fractionation and Analysis of Complex Environmental Mixtures, Williamsburg, V A ) , New York, Plenum Press, Environmental Science Research, v. 15, pp. 383-418.(1978). Ikegami, M . , Fukuda, M . , Yoshihara, Y . , Kaneko, J. "Combustion Chamber Shape and Pressurized Injection in High-Speed Direct Injection Diesel Engines", S A E Technical Paper 900440, (1990). Kato, T , Tsujimura, K , Shintani, M , Minami, T . , Yamaguchi, I. "Spray Characteristics and Combustion Improvement of D l Diesel Engine with High-Pressure Fuel Injection", S A E Technical Paper 890265, (1989). Kayes, D . , Hochgreb,S. "Investigation of the Dilution Process for Measurement of Particulate Matter from Spark-Ignition Engines", S A E Technical Paper 982601, (1998). Kittelson, D . B . Personal Communication, March 22, 2001. Kittelson, D . B. , Johnson, J .H. "Variability in Particulate Emission Measurements in the Heavy Duty Transient Test", S A E Technical Paper 910738, (1991). Lapuerta, M . , Armas, O . , Ballesteros, R., and Duran, A . "Influence of Mini-Tunnel Operating Parameters and Ambient Conditions on Diesel Particulate Measurement and Analysis ," S A E Technical Paper 1999-01-3531,(1999). Lipfert, F . W . "Filter Artifacts Associated with Particulate Measurements: Recent Evidence and Effects on Statistical Relationships", Atmospheric Environment, v.28,No.20, pp.3233-3249, (1994). MacDonald, J.S., Plee, S. L . , D ' A r c y , J. B . , Schreck, R . M . "Experimental Measurements of the Independent Effects of Dilution Ratio and Filter Temperature on Diesel Exhaust Particulate Samples", S A E Technical Paper 800185, (1980). MacDonald, J.S., Barsic, N.J . , Gross, G.P. , Shahed, S.P, Johnson, J . H . , "Status of Diesel Particulate Measurements", S A E Technical Paper 840345, (1984). Mayer, W. J., Lechman, D . C , Hilden, D . L . "The Contribution of Engine O i l to Diesel Exhaust Particulate Emissions", S A E Technical Paper 800256, (1980). 53 Needham, J.R., May, M . P . , Doyle, D . M . , Faulkner, S .A. , Ishiwata, H . "Injection Timing and Rate Control - A Solution for Lower Emissions", S A E Technical Paper 900854, (1990). Norton, P., Vertin, K . , Bailey, B . , Clark, N . N . , Lyons, D . W . , Goguen, S., Eberhardt, J. "Emissions from Trucks Using Fischer-Tropsch Diesel Fuel" , S A E Technical Paper 982526, (1998). Pierpont, D . , Reitz, R. "Effects of Injection Pressure and Nozzle Geometry on D.I. Diesel Emissions and Performance", S A E Technical Paper 950604, (1995). Rao, K . K . , Winterbone, D . E . , Clough, E . "Laser Illuminated Photographic Studies of the Spray and Combustion Phenomena in a Small High Speed D l Diesel Engine", S A E Fuels and Lubricants Meeting & Exposition, (1992). Schauer, J.J., Kleeman, M . J . , Cass, G.R. , Simonett, B . R . T . "Measurement of Emissions From A i r Pollution Sources: C\ Through C30 Organic Compounds from Medium Duty Diesel Trucks", Environ. Sci. Technol, 33, p. 1578-1587, (1999). Snelling, D.R. , Smallwood, G.J. , Sawchuk, R . A . , Neil l , W.S. , Gareau, D . , Chippior, W . L . , L i u , F., Giilder, O . , Bachalo, W . D . "Particulate Matter Measurements in a Diesel Engine by Laser-Induced Incandescence and the Standard Gravimetric Procedure", S A E Technical Paper 1999-01-3653,(1999). Smith, A . Tidmarsh, T . D . , Wilcock, M . "The Effect of Engine Tune on the Emissions Performance of a Modern High Speed Direct Injection Diesel Engine", S A E Technical Paper 980175,(1998). Sorenson, S . C , Mikkelsen, S.E. "Performance and Emissions of A 0.273 Litre Direct Injection Diesel Engine Fuelled with Neat Dimethyl Ether" S A E Technical Paper 950064, (1995). Uludogan, R., Foster, D . E . , Reitz, R . D . "Modeling the Effect of Engine Speed on the Combustion Process and Emissions in a D l Diesel Engine", S A E Technical Paper 962056,(1996). Wedekind, B . G . A . , Andersson, J .D. , Hall , D . , Stradling, R., Barnes. C , Wilson, G . " D E T R / S M M T / C O N C A W E Particle Research Programme: Heavy-Duty Results", S A E Technical Paper 2000-01-2851, (2000). Williams, D.J . , Milne, J.W., Quigley, S . M . , Roberts, D . B . , Kimberlee, M . C . "Particulate Emissions From 'In-Use' Motor Vehicles - Diesel Vehicles", Atmospheric Environment, 23, N o . 12, p.2647,(1989). Williams, P.T., Abbass, M . K . , Andrews, G . E . , Bartle, K . D . "Diesel Particulate Emissions: The Role of Unburned Fuel" , Combustion and Flame, v. 75 pp. 1-24, (1989). 54 Appendix A. A Critical Review of the EPA 2007 Mass-Based Particulate Matter Emissions Legislation Target of 0.01 g/bhp.hr For Heavy-Duty Engines The latest Unites States Environmental Protection Agency (US E P A ) year 2007 emission standard for vehicle exhaust particulate matter (PM) is 0.01 g/bhp.hr for heavy-duty engines. This legislation is driven by examination of health study data correlating human mortality and morbidity (increased incidence of new, or aggravation of existing, diseases) with ambient mass-based P M exposure (Dockery and Pope, 1994; J. Schwartz, 1999; Bascom et al., 1996). However, ambient P M and its vehicle-sourced component are heterogeneous mixtures of various sizes and chemical species of particles. Therefore, the properties of P M responsible for human mortality and morbidity still must be identified, as well as the likely mechanisms by which the offensive particles impede normal physiological functions. If studies show that one size classification or chemical composition of P M contributes more strongly to adverse human health effects, then future E P A legislation should not simply target mass-based emissions as in the case of the 2007 standard. The emission rate should be further described in terms of particle sizes and associated number counts, and/or chemical composition. This review uses conclusions from recent investigations to compare the relative impact of the following P M properties on human mortality and morbidity: size class, number count and chemical composition. There is sufficient research so far to support the recommendation that future P M standards should include particle size and number count emissions targets. However, a lack of conclusive data correlating P M chemical composition with adverse health effects makes it difficult to recommend revising current standards to include composition restrictions. Before this discussion can take place, however, it is useful to describe the respective P M properties mentioned above in more detail: 55 Size class. The size distributions of vehicular P M are classified as follows. Coarse particles are those with aerodynamic greater than 2.5 microns; fine particles are those with diameter between 2.5 and 0.1 micron, and ultrafine particles are those with diameter below 0.1 micron. Traditionally, particle sizes are also classified as P M i 0 , P M 2 5 _ 1 0 , P M 2 5 , PMo.i -2.5 and PM0.1, where the subscript indicates the upper bound or range of particle diameter. N u m b e r count. In comparing the effects of different particle characteristics on human health, it is important to consider that particle number count is also an important characteristic when determining P M toxicity. This is because particle size and number count together define the effective surface area of a given P M mass; the importance of surface area on P M toxicology will be discussed in more detail in the following section of this report. Chemical Composition. Particulate matter is composed of soot, volatile organic hydrocarbons originating from fuel and lubricating oil, (either unburned or partially burned in both cases), metallic elements from fuel and lubricating oil additives, trace metals from engine wear, and sulfate/nitrate ions. Each of these species may be toxic to human tissue and ultimately affect human mortality and morbidity. This will be further discussed in a later section of this paper. Finally, the term human morbidity, an umbrella term used to encompass all disease-related consequences exposure to toxic substances, will be used to describe adverse health effects involving the incidence of new disease and/or aggravation of existing disease. Effect of Particle Size and Number Count on Human Mortality and Morbidity Epidemiological study data by Dockery et al. (1992), Schwartz et al. (1996), and Pope (2000) suggested that daily levels of P M 2 5 pollution were more strongly associated with mortality than were P M i 0 levels. Further to this, Seaton et al. (1995) hypothesized that ultrafine P M below 0.1 microns in diameter was responsible for the associations between P M and detrimental health outcomes even at low mass concentrations of ambient P M 0 . 1 . The latter statement suggests that P M mass concentration may not be the sole index by which one should assess P M toxicity. For example, Oberdorster et al. (1995) hypothesized that the surface area of P M of a given mass concentration was a more important property. T o illustrate this, a fixed mass concentration of 10 pg/cm 3 can result in an aerosol with: a) a number count of 1 particle/cm3 of 2.5 [im diameter particles, or b) 2 million particles/cm3 of 0.02 pm diameter particles. The surface area of the aerosol described in b) is much greater than that described in a). 56 A Closer Look at the Effect of Particle Size and Number Count on Human Physiological (Pulmonary and Cardiovascular) Reactions The repercussions of human exposure to high-surface area P M mass concentrations were described by Wichmann et al. (2000) along with the subsequent mechanisms by which ultrafine P M might contribute to toxicity in human tissues: 1. Mass Transport ofAdsorbed Species as a Function of PM Surface Area. The higher surface area associated with ultrafine P M of a given mass concentration results in higher mass transport of species adsorbed to these primary particle surfaces deep into lung tissue. These species include: reactive gases, radicals, transition metals, and organic hydrocarbons. Greater surface area results in higher oxidative stress on lung tissues related to the influx of these adsorbed species (Li et al., 1999). 2. Deposition of Inhaled Particles Due to Physical Particle Size. The probability of a particle depositing itself in the lung is at a minimum for 0.5 (im diameter particles and increases for smaller diameter particles1. 50% of inhaled 0.02 um diameter particles are estimated to deposit themselves in the alveolar regions of the lung and lower regions of the tracheobronchial tree. Once inhaled, these particles penetrate more rapidly into interstitial lung tissue than do larger particles; some studies have even shown translocation of ultrafine P M in remote organs such as the liver and heart. 3. Limitation of Human Immune Response to Ultrafine Particles. The lung's ability to protect its tissues from foreign substances is compromised in the case of ultrafine particles because the lung cells do not "recognize" these particles as well (based on their size). This results in ultrafine particles migrating to the epithelium 2 and interstitial cell sites, where body defences have failed to impede this passage. The migration can be slow, a consideration that inhibits normal pathophysiological reactions from occurring. The process by which tissue is protected is called phagocytosis and is carried out by bodies known as alveolar macrophages. 4. Insolubility of Particles in Lung Fluid. The increased surface area of ultrafine particles increases surface area dependent reactions that reject these particles, such as 1 International Commission on Radiological Protection, 1994. 2 The covering of internal and external surfaces of the body, including lining of vessels and other small cavities. This covering consists of cells joined by small amounts of cementing substances, and can be several layers deep. 57 inflammation and phagocytosis. Thus particles that are not readily soluble in the fluid of the epithelial lining and with high surface area essentially become an interface between other retained particles and lung cells, fluids and tissues that impedes normal processes. So how do these acute physiological changes brought on by small particle penetration of lung tissue cause remarkable health changes? The short-term, or immediate effects are largely pulmonary. Lippmann et al. (2000), Schwartz et al. (1996) and Thurston et al. (1994) found that diseases such as: chronic obstructive pulmonary disease ( C O P D ) , asthma, pneumonia, and other respiratory diseases were more closely associated with particles below the P M 2 5 size classification than with larger particles. However, initial reactions occurring in the lungs cause secondary effects on processes in other body systems, such as the cardiovascular system. Seaton et al. (1995) suggested that low-grade inflammations increased blood coagulation and altered blood rheology, ultimately causing higher cardiovascular mortality and morbidity. L i et al. (1999) provided numerical data to support this hypothesis by monitoring a measurable index of blood coagulation and finding that it increased when ultrafine elemental carbon was introduced into rat trachea. These studies emphasized the causality between local pulmonary effects directly resulting from P M inhalation and the influence that these local reactions can have on other major body systems, namely the cardiovascular system. Effect of Particle Composition on Human Mortality and Morbidity Presently, there is less scientific data correlating the epidemiological effects of P M composition on human mortality and morbidity as compared to data relating the effect of particle size on human health (Brauer, M . , 2001). Those studies that exist tend to have inconclusive data. For example, a study by Imrich et al. (1996) found that the insoluble fraction of a sample of concentrated ambient air pollution particles (CAPs) of particle size range 0.1 - 2.5 um provoked the highest immune response involving alveolar macrophages, which are involved in fighting off foreign particles in the lungs. However, when this insoluble fraction was further analysed for elemental content, no single element emerged as being most significantly toxic to the lungs making it difficult to attribute toxicity to any single component. One problem with such investigations is that only a portion of the particles introduced is actually taken up by the lungs, and the mechanism by which the lungs selectively or randomly take up different types of particles is still unknown. 58 Another study by Lippmann et al. (2000) investigated the effects of particle sulphate content, acidity and size class. From a composition standpoint, it was found that sulphate ions contributed more to respiratory-related hospital admissions and mortality than did particle acidity. However, the author attributed this result to low acidity readings, which may have understated the role of particle acidity adversely affecting health as compared to sulphate ion concentration. Difficulty in Separating the Effects of Particle Size vs. Composition on Human Health Lippmann et al. (2000) also showed that size-related particulate matter properties affected mortality and hospital admissions from pulmonary causes more than particle composition (sulphate or acid components). In addition, Imrich et al. (1996) attributed the uncertainty in detrimental pulmonary effects due to particle composition to a lack of data for particle number count or effective dose actually interacting with pulmonary cells. Without knowing the specific particle count per cell, the investigators suggested that interpretation of biological response due to elemental particle loading would be difficult. In short, it seems as though health researchers are hesitant to consider only the effects of particle composition without knowing number counts and actual particle population of lung cells, even for fixed particle size range. This suggests that particle size and number count data are of primary importance, and knowledge of particle composition is less critical in describing the toxicity of a given P M mass (McTaggart-Cowan, 2001). Further to the question of particle size vs. composition effects, is there possibly an interaction between the two characteristics, making it hard to discern which property poses a more serious threat to human health? The toxicity of particles of known composition has been studied, and the results suggest that toxicity is a function of size more than composition. To illustrate this, MacNee et al. (2000) found that inhalation of a mass concentration of 1 ug/cm 3 of ultrafine, 14 nm-diameter elemental carbon particles produced severe inflammation effects in rat lungs, whereas the same mass concentration of fine, 260 nm-diameter particles of the same composition showed no significant effects. L i et al. (1999) also reinforced this point by quantifying the inflammatory response for 3 different particle diameters, holding the mass exposure constant at 125 fig. He measured the percentage of neutrophils (body chemical produced as first line of defence to foreign substances) produced in a given volume o f 59 bronchial fluid for particle diameters of: 14 nm, 50 nm, and 260 nm. The results were 40, 14 and 4%, respectively, indicating that the smallest particles provoked the highest immune response. It seems from this information that particles of a given composition are toxic by virtue of their size. C o n c l u s i o n Based on conclusive data relating adverse human health effects to particle size and composition, it has been determined that detrimental health effects of ultrafine particles are greater than the health effects of the same mass of fine particles. This result has reinforced the concept that the surface area of a given mass of P M is the most critical parameter in assessing the toxicity of particulate matter to human lungs. B y contrast, a lack of convincing evidence makes it difficult to attribute human mortality and morbidity to specific chemical species. Studies comparing particle composition vs. size effects on human health showed that size had more significant effect on mortality and morbidity. From the evidence provided, then, it is recommended that future E P A vehicle P M emissions legislation should be size and number-based and not simply mass based. N o such recommendation can be made where P M composition is concerned until the effect of different chemical species on human health can be established. 60 R E F E R E N C E S Bascom,R., Bromberg, P . A . , Costa, D . A . , Devlin, R,. Dockery, D . W . , Frampton, M . W . , Lambert, W. , Samet, J . M . , Speizer, F .E . , Utell, M . "Health Effects of Outdoor A i r Pollution", American Journal of Respiratory and Critical Care Medicine, vol. 153, p.3-50 (1996). Brauer, M . University of British Columbia, Vancouver, British Columbia, Canada, Dept. of Occupational Hygiene (Personal Communication) (2001). Dockery, D . W . , Schwartz, J., Spengler, J .D. " A i r Pollution and Daily Mortality: Association with Particulates and A c i d Aerosols", Environmental Research, vol. 59: 362-373 (1992). Dockery, D . W . , and Pope, C A . "Acute Respiratory Effects of Particulate A i r Pollution", Annual Review of Public Health, vol. 15: p. 107-132 (1994). Imrich, A . , Ning, Y . Y . , Kobzik, L . "Insoluble Components of Concentrated A i r Particles Mediate Alveolar Macrophage Responses In Vitro" , Toxicology and Applied Pharmacology, vol. 167, p. 140-150(2000). L i , X . Y . , Brown, D . , Smith, S, MacNee, W. Donaldson, K . "Short-Term Inflammatory Responses Following Intratracheal Instillation of Fine and Ultrafine Carbon Black In Rats", Inhalation Toxicology, vol. 11, p. 709-731 (1999). Lippmann, M . , Ito, K . , Nadas, A . , Burnett, R. "Association of Particulate Matter Components with Daily Mortality and Morbidity in Urban Populations", Health Effects Institute, Report 95, (2000). MacNee, W. , L i , X . Y . , Seaton, A . , Donaldson, K . "Effects of Short-Term L o w Exposure to Carbon Black", American Journal of Respiratory and Critical Care Medicine, vol. 162: 440-446 (2000). McTaggart-Cowan, J. Royal Roads University, Victoria. British Columbia, Canada, Dept. of Atmospheric Science (Personal Communication) (2001). Oberdorster, G . , Gelein, R . M . , Ferin, J., Weiss, B . "Association of Particulate A i r Pollution and Acute Mortality: Involvement of Ultrafine Particles?", Inhalation Toxicology, vol. 7, p. 111-124. (1995) Pope, C A . "Epidemiology of Fine Particulate A i r Pollution and Human Health: Biological References and Who's A t Risk", Environmental Health Perspectives, vol.108, Supplement 4 (August 2000) Schwartz, J. " W h y Are People Dying of High A i r Pollution Days?", Environmental Research, vol. 64, p. 26-35 (1994b). Schwartz, J., Dockery, D . W . , Neas, L . M . "Is Daily Mortality Associated Specifically With Fine Particles?", Journal of Air and Waste Management, vol. 46, p. 927-939, (1996). 61 Seaton, A . , MacNee, W, Donaldson, K , Godden, D . "Particulate A i r Pollution and Acute Health Effects", Lancet, vol. 345, p. 176-178 (1995). Thurston, G . D . , Ito, K . , Hayes, C . G . , Bates, D . V . , Lippmann, M . "Respiratory Hospital Admissions and Summertime Haze A i r Pollution in Toronto, Ontario: Consideration of the Role of A c i d Aerosols", Environmental Research, vol. 65, p. 271-290 (1994). Wichmann, H . E . , Spix, C , Tuch, T . , Wolke, G . , Peters, A . , Heinrich, J., Kreyling, W . G . , Heyder, J. "Daily Mortality and Fine and Ultrafine Particles in Erfurt, Germany. Part I: Role of Particle Number and Particle Mass", Health Effects Institute, Report 98 (2000). 62 G l o s s a r y o f T e r m s Endotoxin. a complex bacterial toxin composed of protein, lipid, and polysaccharide, which is released only upon lysis of the cell. Also , B A C T E R I A L P Y R O G E N . Rheology. - The dynamics of macromolecular liquids - The study of the deformation and flow of matter, usually liquids or fluids, and of the plastic flow of solids. The concept covers consistency, dilatancy, liquefaction, resistance to flow, shearing, thixotrophy, and V I S C O S I T Y . Synonyms: Flowmetry, Velocimetry. Phagocytosis: Process involving white blood cell leaving the blood, with the intent of consuming some microbe by engulfing it. Phago = eating, cytose: cell Macrophage: Epithelium: The cell-eating entity involved in phagocytosis. The covering of internal and external surfaces of the body, including lining of vessels and other small cavities. This covering consists of cells joined by small amounts of cementing substances, and can be several layers deep. Ischemic Heart Disease: Temporary shortage of oxygen into a part of the body, that ultimately can result in cardiac arrest (example:angina pectoris) Lysis: the destruction or splitting of cells by antibodies, chemical agents, or the like. 63 Appendix B. Filter Handling Procedures For PM collection with subsequent VOF analysis, collect PM on EMFAB TX40HI20-WW 90-mm diameter disc filters. B.l Filter Pre-Conditioning o Place a filter taken from the stack kept in the conditioning chamber and place it in a labelled petri dish (the filters should be labelled as follows, according to the medium -E M F A B : EM-#### o Record the humidity and temperature in the weighing room. o Always use tweezers gently to manipulate filters. o Autocalibrate the balance for masses 0, 2.5 and 5 mg. o Start by weighing the two blank filters permanently stored in the conditioning chamber. Their respective masses should be within +/- 0.04 mg of the value recorded for their previous weighing, o Weigh filters: o weigh each filter separately 3 times and take average - to learn about repeatability. o Place one filter per individual labelled petri dish. B.2 PM Sampling Procedures B.2.1 Filter Handling: o Remove filter from petri dish gently using tweezers o Place the filter in the front section of filter holder then immediately replace the front section of the holder so that the filter is sealed off from contamination. o Record the filter codes corresponding to the various modes of the 8-mode test in the P M log book o Remove another filter from its petri dish o Place this filter in the back section of the filter holder and connect this back section back to the front section o Connect the filter holder to its intended location in the P M testing apparatus B.2.2 PM Sampling: o For accurate results, use a dilution ratio of 12, and a sampling time of 10 min., and a sample flowrate if 180 L/min. o E N S U R E that the filter temperature is maintained below 52 ° C . o Record the final values of filter differential pressure, filter temperature, sample mass, total mass, and dilution ratio in the log book for every test point B.2.3 Post-PM Sampling Filter Handling o Repeat steps in 3.4.1 above in reverse order so that filters are removed from the holders and returned carefully (using tweezers) to their individual petri dishes. 64 B.2.4 Post-PM Weighing of Filters o Place exposed filter in control chamber for 24 hours. o Always use tweezers gently to manipulate filters. o Record the humidity and temperature in the weighing room. o Autocalibrate the balance for masses 0, 2.5 and 5 mg. o Start by weighing the two blank filters permanently stored in the conditioning chamber. Their respective masses should be within +/- 0.04 mg of the value recorded for their previous weighing, o Weigh filters: o weigh each filter separately 3 times and take average - to learn about repeatability. o Place one filter per individual labelled petri dish. Keep post-weighed filters in a freezer in order to keep volatile components from being driven off during storage time. B.3 Effect of Filter Conditioning Parameters on Particulate Matter Mass and Composition B.3.1 Relative Humidity in Conditioning Chamber It is important to control the relative humidity of the area where both filter conditioning and weighing take place. Lapuerta et al. (1999) demonstrated this in an experiment that held conditioning chamber temperature constant while increasing the R H from 28% to 79%. A n increase in R H from 45 to 79% resulted in a 10% increase in P M mass measured on the filters. It was suggested that this was a result of water adsorption onto the sooty fraction of the exhaust P M on the exposed filter. This was confirmed by performing a solvent extraction/gas chromatograph (Soxhlet type) on the S O F of the P M , and an H P L C analysis on the sooty fraction of the P M . It was found that the mass of the sooty fraction of the P M increased with increasing R H . B.3.2 Temperature in Conditioning Chamber Lapuerta et al. (1999) also examined the effects of changing conditioning chamber temperature on measured P M mass. N o significant change in P M mass was observed for temperatures varying from 17 to 34 0 C . The R H in the conditioning area was held constant at 45% during this investigation. 65 Appendix C. Sierra BG-1 Micro-Dilution Tunnel Calibration and Operation Procedures READ THIS FIRST. VITAL OPERATIONAL INFO BELOW: • MONITOR TEST C E L L AMBIENT TEMPERATURE (value found on Digalog screen). If it deviates +/- 5°C, re-calibrate the Sierra tunnel • NEVER RUN THE SYSTEM ON SAMPLE MODE WITHOUT FILTERS IN THE FILTER HOLDER ASSEMBLY. This can foul the HEP A fdter upstream of the total flow meter (LFE) and potentially foul the bank of flow elements within the flowmeter itself. • When not taking a PM sample or logging emissions, SYSTEM SHOULD ALWAYS IN PURGE MODE • KEEP BACK DOOR OF CABINET CLOSED AT A L L TIMES unless checking temperature readings on Omega panel. C l Pre-Test Procedures 1. Actuate system valves manually. Click on D I A G , then M A N U A L V A L V E SWITCHPWG. Click on each of valves 0-7 two or three times (doing this will change them from red to green on the screen). Verify that valves # 1 and 2 are opening and closing cleanly, and not hissing air when open or closed. RECORD FAULTY V A L V E OPERATION ON PM TESTING DATA ENTRY SHEET. 2. Perform leak test: click on D I A G , then P E R F O R M L E A K T E S T , with alarm set at +/- 5 psi. For this procedure, cap exhaust inlet to dilution chamber, and connect all other lines including the fdter holder without fdters. Once test is over, RECORD PRESSURE LOSS DURING 180s TEST TIME ON PM TESTING DATA ENTRY SHEET. 3. If system fails leak test, cheek for leaks using Snoop under dilution chamber hood and in main cabinet. 4. Check calibration file coefficients. Open file c:\ BG2\Bgcal l .cfg for correct coefficients as per instructions at top of printout in section B.4. 5. Run system calibration. Click C A L I B R A T E , select file B G C A L 2 , cap lines as in 2. 6. Monitor line temperatures. Check O M E G A thermocouple readings ~ 3 0 ° C during calibration. They will be erratic over the 0-40 L P M flow range, but should settle after these points. 7. Check calibration results. Calibration curve should be linear, and have a zero intercept. Values in % diff. column for flows 100 - 200 L P M should be close to 0 (+/- 0.25%). Click "write new coefficients" box, and then "yes" on the subsequent dialog box to write these calibration coefficients to file. IF THE CALIBRATION VIOLATES ANY OF THE ABOVE CRITERIA, RE-RUN IT. RECORD RANGE OF VALUES IN " % DIFF." COLUMN ON PM TESTING DATA ENTRY SHEET. 66 C.2 Procedures During Testing • D u m m y filter run. Place a set of unweighed, clean filters in the filter holder and run a sample point for 180 s, standby mode 30s, DR=12, Q=180 L P M . Discard these after the run. • Place pre-conditioned and pre-weighed filters in the filter holder assembly. Note the codes for these filters in test log along with the engine operation characteristics • Click on "purge" and then "run until I hit stop" until the engine has stabilized and you have been given the ok to draw a P M sample. O N L Y E X C E P T I O N T O THIS P U R G I N G R U L E IS: L O G G I N G E M I S S I O N S (i.e. don't purge during this time) • Set standby mode to 60s, sample time to 600s, dilution to 12, flow to 180 (for E M F A B filters) • While sample id being collected, check O M E G A thermocouple readings ~ 3 0 ° C • Take dPfiiter readings at t=0 s and t = 600s (ensure dP < 136.95" H20 - this means that dP maxed out if you are at this value!) • TakeTfiKer readings every 3 minutes, it should never exceed 52°C • Place loaded filters in the glove box as soon as possible after the point has been taken C.3 Post-Testing Procedures • Remove the exhaust line probe from the main exhaust line. Cap the port to the exhaust. 67 C . 4 B G - 1 C a l i b r a t i o n F i l e : V e r i f i c a t i o n o f Flow C o e f f i c i e n t s INSTRUCTIONS: The l i n e numbers of t h i s f i l e are shown near the middle o f each l i n e . V e r i f y t h a t l i n e s 22 through 30 ( i n b o l d below) on your system match the numbers shown on t h i s f i l e . These are c a l i b r a t i o n c o e f f i c i e n t s f o r the d i l u t i o n meter. The o n l y time t h a t they sh o u l d change i s i f you c a l i b r a t e d your system a g a i n s t another standard. 007- B) 008- S) 009- S) 010- S) 011- S) < BGCAL1.CFG FILE FOR BG1 REV 4.0> ; < LAST MFC CALIBRATION INFORMATIONS >-TRUE 03-25-1997 d a t e 10:27:05 t i m e 03-25-1997 d a t e 10:27:05 t i m e 03-25-1997 d a t e 10:27:05 t i m e 12-21-1999 d a t e 07 : 08:25 t i m e ; < SOFTWARE CONFIGURATIONS > 200.00 ' SLPM.(200.00) 1 1 ' 2 ; < DILUTION FLOW METER COEFFS > 0.000000E+0 System C a l i b r a t i o n Auto=True DIL - MFC's l a s t 1 y r c a l i b MFC's l a s t 1 y r c a l i b l a s t 24 hour c a l i b l a s t 24 hour c a l i b 012-S) TOT - MFC's l a s t 1 y r c a l i b 000000E+0 000000E+0 000000E+0 000000E+0 000000E+0 000000E+0 000000E+0 000000E+0 ;013 -S) MFC's l a s t 1 y r c a l i b ;014 -S) l a s t 24 hour c a l i b ; 015 -S) l a s t 24 hour c a l i b ;017 -R) C a l i b . s e t t l e r a t e i n ;018 -I) C a l i b . . s e t t l e t i m e i n mins. ;019 -I) A u t o B a l w a i t p e r i o d i n mins. ;020 -I) # o f f a i l u r e s a l l o w e d ;022 -R) A A X A 5 OF F I F T H ORDER ;023 -R) B BX A 4 OF F I F T H ORDER ;024 -R) C C X A 3 OF F I F T H ORDER ;025 -R) D D X A 2 OF F I F T H ORDER ;026 -R) E E X A OF F I F T H ORDER ;027 -R) F F OF F I F T H ORDER ;028 -R) G SPAN VALUE ;029 -R) H ZZERO VALUE ;030 -R) TOTALIZER SCALE (1.00) ; < TOTAL FLOW METER COEFFS >-2.757027E-10 -1.205868E-7 1.667683E-5 -7.395267E-4 1.004081E+0 5. 425207E-2 9. 895219E-1 -1.055490E+0 1.000000E+0 031- R) A AX A5 OF FIFTH ORDER 032- R) B BX A4 OF FIFTH ORDER 033- R) C CX A3 OF FIFTH ORDER 034- R) D DX A2 OF FIFTH ORDER 035- R) E EX A OF FIFTH ORDER 036- R) F F OF FIFTH ORDER 037- R) G SPAN VALUE ;038-R) H ZZERO VALUE 039-R) TOTALIZER SCALE 68 C.5 PM Testing Data Entry Sheet author ambaribeau 14-FeW)1 PRE-TEST PROCEDURES Date Operator(s) |Test Requisition Delta Map Version associated file Fuel Type VALVE CHECK LEAK TEST CALIBRATION Iproblem valves: | Ipressure loss: | psi i 180 s | 0,0 intercept Y/N |(circle one) range of % diff. values % to + % filter code mode "H20 dP(l) "H20 dP(f) C T(l) C |T(3) ic T(6) C T(f) (g) Sample Mass (9) Dilution Mass (9) |Total Mass C |Tamb (test cell) Tomega |#1 #2 Appendix D. Overview of Particulate Matter Emissions Testing and Analyses D.O Overview of Tests and Analyses Performed Table D.O Table of PM Testing and Analyses Date Test Req. Test Description Analyses Procedure Jan. 10,11 2001 329/331 "4 comers" repeatability (Study A) Exhaust gases, gravimetric filter analysis Modes 1,4,5,8 repeated 5x each Gas Rail Pressure (GRP) = 19 MPa Timings: Mode 1: -2°BTDC Mode 4: -6°BTDC Mode 5: -23°BTDC Mode 8: - 17°BTDC Jan. 17, 2001 336,337 Repeatability Testing HPDI AVL 8-Mode Composite AVL 8-mode tests #1,2 (Study B) Exhaust gases, gravimetric filter analysis, VOF composition analysis (modes 1,4,5,8) GRP= 19 MPa Jan. 18,23 2001 339,340 Timing Study Exhaust gases, gravimetric filter analysis GRP = 19 MPa Modes 1,4,5,8 run at 4 different timings: (advanced to retarded): Mode 1: -16°, -14° BTDC, +3° ATDC M o d e 4 : - H ° , - 8 0 , - l ° B T D C Mode 5: -36°, -30°, -16.5° BTDC Mode 8: -29°, -18°, -15° BTDC Jan. 17, 2001 338 Pressure Study exhaust gases, gravimetric filter analysis Mode 1: Plow=10 MPa, Pilot SOI* = TBTDC Mode 4: P,ow=17 MPa, Pilot SOI = 6°BTDC Mode 5: P|0W=13 MPa, Pilot SOI = 23°BTDC Mode 8: P!ow =23 MPa, Pilot SOI= 16°BTDC Phish= 25 MPa for all modes Mar. 26, 2001 388 Pilot Fuel Quantity Study exhaust gases, gravimetric filter analysis Mode 2 Pilot Pulse Width (PW) = 0.48, 0.65, 1.1 ms (fixed injection timing, injection pressure) 70 D.A. Emissions and Fuel Flow Tables and Plots from Repeatability Study A Table DA(i). Mode 1 Emissions (g/bhp.hr, kg/bhp.hr) Number of Sample Points n=5 1 2 3 4 5 avg 3-sigma error Emission CO (g/bhp.hr) 4.18 4.00 4.24 4.09 4.23 4.15 0.30 C02 (kg/bhp.hr) 0.91 0.90 0.91 0.89 0.90 0.90 0.02 NOx (g/bhp.hr) 12.81 12.80 12.74 12.62 12.68 12.73 0.25 CH4 (g/bhp.hr) 4.75 4.31 4.76 4.48 4.73 4.61 0.61 nmHC (g/bhp.hr) 1.84 1.77 1.89 1.84 1.89 1.85 0.15 PM (g/bhp.hr) 0.217 0.231 0.240 0.227 0.248 0.233 0.036 Table DA(iv). Mode 4 Emissions (g/bhp.hr, kg/bhp.hr) Number of Sample Points n=5 n=5 n=4 n=4 1 2 3 4 5 avg 3-sigma error avg 3-sigma error Emission CO (g/bhp.hr) 1.03 1.12 1.16 1.17 1.23 1.14 0.227 1.17 0.142 C02 (kg/bhp.hr) 0.36 0.36 0.36 0.36 0.36 0.36 0.004 0.36 0.002 NOx (g/bhp.hr) 2.77 2.76 2.76 2.76 2.75 2.76 0.023 2.76 0.022 CH4 (g/bhp.hr) 0.22 0.21 0.21 0.21 0.21 0.21 0.005 0.21 0.002 nmHC (g/bhp.hr) 0.14 0.09 0.08 0.08 0.08 0.09 0.078 0.08 0.019 PM (g/bhp.hr) 0.022 0.017 0.015 0.014 0.014 0.017 0.009 0.015 0.004 Table DA(v). Mode 5 Emissions (g/bhp.hr, kg/bhp.hr) Number of Sample Points n=5 1 2 3 4 5 avg 3-sigma error Emission CO (g/bhp.hr) 2.27 2.24 2.26 2.27 2.21 2.25 0.08 C02 (kg/bhp.hr) 0.55 0.55 0.55 0.55 0.55 0.55 0.00 NOx (g/bhp.hr) 3.83 3.87 3.84 3.83 3.90 3.85 0.09 CH4 (g/bhp.hr) 2.75 2.67 2.68 2.75 2.66 2.70 0.12 nmHC (g/bhp.hr) 0.73 0.71 0.71 0.73 0.73 0.72 0.03 PM (g/bhp.hr) 0.114 0.120 0.127 0.122 0.115 0.119 0.016 Table DA(viii). Mode 8 Emissions (g/bhp.hr, kg/bhp.hr) Number of Sample Points n=5 n=5 n=4 n=4 1 2 3 4 5 avg 3-sigma error avg 3-sigma error Emission CO (g/bhp.hr) 0.61 0.60 0.61 0.60 0.60 0.60 0.018 0.60 0.013 C02 (kg/bhp.hr) 0.41 0.41 0.41 0.41 0.41 0.41 0.006 0.41 0.003 NOx (g/bhp.hr) 1.38 1.36 1.36 1.35 1.35 1.36 0.032 1.35 0.012 CH4 (g/bhp.hr) 0.32 0.32 0.31 0.32 0.31 0.32 0.003 0.31 0.002 nmHC (g/bhp.hr) 0.10 0.09 0.09 0.09 0.09 0.09 0.020 0.09 0.004 PM (g/bhp.hr) 0.034 0.016 0.014 0.014 0.014 0.018 0.027 0.014 0.002 71 Table DAI. Mode 1 Emissions and Fuel Flows (g/hr, kg/hr) Number of samples n=5 n=5 M O D E 1 1 2 3 4 5 avg 3-sigma error CO (g/hr) 29.34 28.43 29.72 28.99 29.71 29.24 1.64 C02 (kg/hr) 6.38 6.38 6.38 6.33 6.35 6.37 0.07 NOx (g/hr) 89.97 90.91 89.37 89.42 89.18 89.77 2.11 02 (kg/hr) 60.79 60.79 60.86 60.87 61.17 60.90 0.47 CH4 (g/hr) 33.38 30.62 33.39 31.75 33.29 32.48 3.76 nmHC (g/hr,Cl) 12.92 12.56 13.23 13.05 13.30 13.01 0.89 tHC (g/hr,Cl) 46.30 43.17 46.62 44.80 46.60 45.50 4.50 PM (g/hr) -corrected 1.53 1.64 1.68 1.61 1.75 1.64 0.25 total PM mass on filters (mg) 0.789 0.845 0.869 0.833 0.898 0.847 0.122 PM on backup filter (mg) 0.153 0.158 0.153 0.143 0.173 0.156 0.032 ratio of PM mass primary: backup filter 4.16 4.35 4.67 4.82 4.20 4.44 0.88 Diesel flow (kg/hr) 0.38 0.35 0.23 0.77 0.63 0.47 0.66 CNG flow (kg/hr) 1.62 1.48 1.73 1.43 1.55 1.56 0.36 Table D A 4 . Mode 4 Emissions and Fuel Flows (g/hr, kg/hr) Number of samples n=4 n=4 n=5 n=5 comments No I s' trial outlier No 1st trial outlier M O D E 4 avg 3-sigma error avg 3-sigma error 1 2 3 4 5 CO (g/hr) 234.68 254.31 265.22 267.61 280.16 266.83 31.83 260.40 51.19 C02 (kg/hr) 82.09 81.56 81.28 81.24 81.22 81.32 0.49 81.48 1.11 NOx (g/hr) 630.35 627.63 627.97 629.42 624.49 627.38 6.22 627.97 6.70 02 (kg/hr) 90.90 89.46 88.86 88.40 88.74 88.86 1.32 89.27 2.96 CH4 (g/hr) 49.45 48.60 48.92 48.62 48.47 48.65 0.58 48.81 1.19 nmHC (g/hr,Cl) 31.48 20.66 17.97 17.59 18.00 18.56 4.25 21.14 17.72 tHC (g/hr,Cl) 80.93 69.27 66.90 66.21 66.46 67.21 4.21 69.95 18.76 PM (g/hr) -corrected 4.96 3.94 3.37 3.30 3.29 3.48 0.94 3.77 2.16 total PM mass on filters (mg) 0.884 0.712 0.615 0.601 0.600 0.632 0.161 0.682 0.366 PM on backup filter (mg) 0.261 0.194 0.228 0.167 0.151 0.185 0.101 0.200 0.134 ratio of PM mass primary: backup filter 2.39 2.67 1.70 2.60 2.98 2.49 1.64 2.47 1.43 Diesel flow (kg/hr) 1.00 1.04 1.07 1.10 1.14 1.09 0.13 1.07 0.17 CNG flow (kg/hr) 28.77 28.73 28.74 28.74 28.73 28.73 0.01 28.74 0.05 72 Table D A 5 . Mode 5 Emissions and Fuel Flows (g/hr, kg/hr) Number of samples n=5 n=5 M O D E 5 3-sigma 1 2 3 4 5 avg error CO (g/hr) 164.97 161.69 163.75 164.97 159.37 162.95 7.23 C02 (kg/hr) 39.68 39.38 39.55 39.68 39.53 39.56 0.37 NOx (g/hr) 277.73 278.85 278.34 277.73 281.08 278.74 4.16 02 (kg/hr) 202.41 200.55 201.76 202.41 201.31 201.69 2.36 CH4 (g/hr) 199.20 192.72 194.41 199.20 191.92 195.49 10.51 nmHC (g/hr,Cl) 52.79 51.14 51.81 52.79 52.36 52.18 2.11 tHC (g/hr,Cl) 251.99 243.87 246.22 251.99 244.29 247.67 12.13 PM (g/hr) -corrected 8.27 8.65 9.20 8.83 8.28 8.65 1.17 total PM on filters (mg) 1.180 1.252 1.320 1.260 1.187 1.240 0.173 PM on backup filter (mg) 0.204 0.202 0.205 0.212 0.156 0.196 0.067 PM mass ratio primary: backup filter 4.80 5.21 5.43 4.95 6.59 5.40 2.14 Diesel flow (kg/hr) 1.85 1.87 1.78 1.85 1.73 1.82 0.18 CNG flow (kg/hr) 12.66 12.62 12.81 12.66 12.52 12.65 0.32 Table D A 8 . Mode 8 Emissions and Fuel Flows (g/hr, kg/hr) Number of samples n=4 n=4 n=5 n=5 comments No 1st trial outlier No 1st trial outlier M O D E 8 avg 3-sigma error avg 3-sigma error 1 2 3 4 5 CO (g/hr) 249.94 244.90 248.53 245.11 244.08 245.65 5.90 246.51 7.69 C02 (kg/hr) 168.48 167.10 166.88 166.51 165.66 166.54 1.90 166.92 3.08 NOx (g/hr) 561.79 554.36 553.67 551.00 548.66 551.92 7.84 553.90 14.88 02 (kg/hr) 209.52 208.63 207.90 208.00 207.09 207.90 1.90 208.23 2.72 CH4 (g/hr) 129.26 128.59 128.32 128.64 127.78 128.33 1.19 128.52 1.61 nmHC (g/hr,Cl) 42.35 36.82 36.63 35.68 35.69 36.20 1.81 37.43 8.40 tHC(g/hr,Cl) 171.61 165.41 164.95 164.32 163.46 164.54 2.53 165.95 9.74 PM (g/hr) -corrected 13.99 6.37 5.86 5.70 5.59 5.88 1.03 7.51 10.92 total PM mass on filters (mg) 1.163 0.530 0.491 0.476 0.470 0.492 0.081 0.626 0.903 PM on backup filter (mg) 0.140 0.131 0.139 0.117 0.125 0.128 0.028 0.130 0.029 PM mass ratio primary: backup filter 7.32 3.06 2.53 3.08 2.76 2.86 0.79 3.75 6.03 Diesel flow (kg/hr) 1.75 1.85 2.02 1.80 1.97 1.91 0.31 1.88 0.34 CNG flow (kg/hr) 60.32 60.37 60.26 60.20 60.16 60.25 0.28 60.26 0.26 73 Figure DA1. Fuel Flow and PM Emissions Data for 5 Repeated Mode 1 Test Points Repeatability Study Total (CNG+Diesel) Fuel Flow, Diesel Flow, PM Emissions: Mode 1 (Idle) of AVL 8-Mode Test 2 3 4 Test Point Trial # I PM emissions • - total fuel flow diesel flow Figure DA4. Fuel Flow and PM Emissions Data for 5 Repeated Mode 4 Test Points Repeatability Study Total (CNG+Diesel) Fuel Flow, Diesel Flow, PM Emissions: Mode 4 of AVL 8-Mode Test 2 3 4 Test Point Trial # 3 PM emissions • -total fuel flow -diesel flow Figure DA5. Fuel Flow and PM Emissions Data for 5 Repeated Mode 5 Test Points Repeatability Study Total (CNG+Diesel) Fuel Flow, Diesel Flow, PM Emissions: Mode 5 of AVL 8-Mode Test o •M E o> uj w E Test Point Trial tt I PM emissions • -total fuel flow -*-diesel flow Figure DA8. Fuel Flow and PM Emissions Data for 5 Repeated Mode 8 Test Points Repeatability Study Total (CNG+Diesel) Fuel Flow, Diesel Flow, PM Emissions: Mode 8 of AVL 8-Mode Test 16 14 12 to c 10 o to 8 n E 6 Ul 4 E 2 CL Test Point Trial ft I PM emissions • -total fuel flow -diesel flow D.B Emissions and Fuel Flow Tables and Plots from Repeatability Study B Table DB(i): A V L 8 l-Mode Test #1: Moc al Emissions Data (g/bhp.hr, kg/b TEST MODE 1 2 3 4 5 6 7 8 CO (g/bhp.hr) 5.38 2.18 0.44 1.46 2.66 1.63 1.59 0.74 C02 (kg/bhp.hr) 0.89 0.43 0.38 0.36 0.56 0.42 0.42 0.42 NOx (g/bhp.hr) 11.39 6.03 4.29 2.75 3.85 2.24 1.58 1.37 CH4 (g/bhp.hr) 7.18 2.10 0.28 0.24 3.28 1.62 1.15 0.38 nmHC (g/bhp.hr) 2.32 0.53 0.21 0.10 0.73 0.33 0.23 0.11 tHC (g/bhp.hr) 9.49 2.63 0.49 0.33 4.01 1.95 1.38 0.48 PM (g/bhp.hr) 0.340 0.097 0.035 0.024 0.146 0.068 0.040 0.030 Table DB(ii): A V L 8-Mode Test #2: Modal Emissions Data (g/bhp.hr, kg/b TEST MODE 1 2 3 4 5 6 7 8 CO (g/bhp.hr) 4.67 2.22 0.43 1.40 2.58 1.62 1.53 0.70 C02 (kg/bhp.hr) 0.88 0.42 0.38 0.36 0.55 0.42 0.42 0.42 NOx (g/bhp.hr) 10.54 6.06 4.25 2.73 3.72 2.23 1.57 1.36 CH4 (g/bhp.hr) 5.53 2.14 0.29 0.23 3.14 1.62 1.10 0.37 nmHC (g/bhp.hr) 1.70 0.53 0.18 0.09 0.71 0.33 0.22 0.10 tHC (g/bhp.hr) 7.23 2.66 0.47 0.32 3.85 1.96 1.32 0.47 PM (g/bhp.hr) 0.294 0.115 0.033 0.021 0.111 0.052 0.037 0.024 Table D B 1 : AVL 8-Mode Test #1: Modal Emissions and Fuel Flow Data (g/hr, kg/hr) T E S T M O D E 1 2 3 4 5 6 7 8 EMISSION CO (g/hr) 38.01 89.21 64.77 333.18 195.32 272.20 460.75 299.98 C02 (kg/hr) 6.30 17.61 54.92 83.21 41.13 70.64 120.57 170.06 NOx (g/hr) 80.41 247.44 626.44 627.15 282.66 375.74 455.82 556.49 CH4 (g/hr) 50.67 86.07 40.64 54.05 240.59 271.22 332.56 153.07 nmHC (g/hr,Cl) 16.36 21.82 31.39 21.82 53.93 55.84 65.12 43.33 tHC (g/hr,Cl) 67.03 107.89 72.02 75.87 294.53 327.06 397.68 196.40 PM (g/hr) -corrected 2.40 3.98 5.04 5.49 10.75 11.45 11.56 12.03 total PM mass on filters (mg) 1.118 1.416 1.120 0.859 1.383 1.176 0.993 0.902 PM on backup filter (mg) 0.294 0.254 0.250 0.215 0.255 0.213 0.214 0.230 Mass ratio: Primary: backup filter 2.80 4.57 3.47 3.00 4.42 4.51 3.63 2.92 Diesel flow (kg/hr) 0.59 0.20 0.81 0.88 1.74 1.84 2.15 1.82 CNG flow (kg/hr) 1.56 5.44 18.27 28.76 12.87 23.77 42.43 60.50 soot and other (g/hr) 1.26 3.02 4.14 6.37 Table D B 2 : AVL 8-Mode Test #2: Modal Emissions and Fuel Flow Data (g/hr, kg/hr) T E S T M O D E 1 2 3 4 5 6 7 8 EMISSION CO (g/hr) 32.63 90.71 62.82 321.95 195.39 272.91 441.76 286.62 C02 (kg/hr) 6.16 17.30 54.57 83.40 41.83 70.87 120.12 169.45 NOx (g/hr) 73.57 248.02 618.87 625.65 282.50 376.23 454.70 552.93 CH4 (g/hr) 38.60 87.48 41.55 52.19 237.87 273.34 318.11 149.16 nmHC (g/hr,Cl) 11.85 21.56 26.89 21.54 54.00 56.34 63.34 41.99 tHC(g/hr,Cl) 50.45 109.04 68.44 73.72 291.87 329.68 381.45 191.15 PM (g/hr) -corrected 2.05 4.70 4.79 4.80 8.39 8.84 10.84 9.74 total PM mass on filters (mg) 0.994 1.715 1.089 0.768 1.089 0.912 0.939 0.730 PM on backup filter (mg) 0.186 0.233 0.178 0.130 0.195 0.175 0.175 0.141 Mass ratio: Primary: backup filter 4.35 6.37 5.13 4.89 4.58 4.21 4.36 4.17 Diesel flow (kg/hr) 0.31 0.65 1.06 1.05 1.80 1.92 1.85 1.64 CNG flow (kg/hr) 1.49 5.44 18.36 28.92 13.02 23.83 42.50 60.62 soot and other (g/hr) 0.73 1.79 3.35 4.41 76 Figure DB1. Fuel Flow and PM Emissions Data for AVL Test 8-Mode Test #1 Total Fuel (CNG+Diesel) Flow, Diesel Flow, Total PM Emiss ions , Soot & Other PM Component Emiss ions: A V L 8-Mode Test #1 5 o w c o 5 F E Lii ^"^ <D 5 Q ro AVL Test Mode total PM emissions soot & other PM components —*—total fuel flow -^—diesel flow Figure DB2. Fuel Flow and PM Emissions Data for AVL 8-Mode Test #2 Total Fuel (CNG+Diesel) Flow, Diesel Flow, Total PM Emiss ions , Soot & Other PM Component Emiss ions: A V L 8-Mode Test #2 ion: ow iss hr) LL. £ ro ese L U ese D a 4 5 AVL Test Mode I total P M emissions H I S soot & other PM components —*—total fuel flow —•—diesel flow 77 D.T Timing Study Emissions Data Tables and Plots Table D T I : Mode 1,4 of A V L 8-Mode Test Emissions Data (g/bhp.hr, kg/bhp.hr) Mode 1 Mode 4 16° BTDC 14° BTDC 3° ATDC 11° BTDC 8° BTDC 1° BTDC CO (g/bhp.hr) 2.78 2.85 16.13 6.93 3.67 0.53 C02 (kg/bhp.hr) 0.88 0.87 0.94 0.35 0.36 0.38 NOx (g/bhp.hr) 27.34 25.33 7.70 3.03 2.70 3.28 CH4 (g/bhp.hr) 2.53 2.62 36.57 0.28 0.24 0.25 nmHC (g/bhp.hr) 1.59 1.49 7.89 0.11 0.09 0.08 tHC (g/bhp.hr) 4.12 4.11 44.46 0.39 0.33 0.34 PM (g/bhp.hr) 0.405 0.439 0.746 0.079 0.035 0.015 PM mass on filters (mg) 1.451 1.584 2.420 3.067 1.363 0.588 mass ratio primary/backup filter 4.41 3.34 5.26 13.99 9.43 3.03 Table DT2. Mode 5,8 of A V L 8-Mode Test Emissions Data (g/bhp.hr, kg/bhp.hr) Mode 5 Mode 8 36° BTDC 30° BTDC 16.5° BTDC 29° BTDC 18° BTDC 15° BTDC CO (g/bhp.hr) 1.67 1.78 6.27 1.41 1.09 1.06 C02 (kg/bhp.hr) 0.52 0.53 0.58 0.37 0.42 0.44 NOx (g/bhp.hr) 12.43 6.28 2.36 1.85 1.32 1.33 CH4 (g/bhp.hr) 1.75 2.07 11.31 0.33 0.44 0.50 nmHC (g/bhp.hr) 0.58 0.71 1.91 0.09 0.17 0.09 tHC (g/bhp.hr) 2.33 2.78 13.22 0.42 0.60 0.59 PM (g/bhp.hr) 0.106 0.195 0.175 0.045 0.051 0.018 PM mass on filters (mg) 1.119 1.981 1.679 1.539 1.625 0.580 mass ratio primary/backup filter n/a 4.45 5.89 7.66 4.02 2.15 78 Table DT3. Mode 1,4 of A V L 8-Mode Test Emissions and Fuel Flow Data (g/hr, kg/hr) Mode 1 Mode 4 16° BTDC 14° BTDC 3° ATDC 11° BTDC 8° BTDC 1° BTDC CO (g/hr) 22.57 23.58 105.42 1581.27 838.20 120.19 C02 (kg/hr) 7.11 7.21 6.12 80.42 81.82 88.01 NOx (g/hr) 221.88 209.37 50.32 692.62 616.63 749.68 CH4 (g/hr) 20.54 21.66 238.94 63.64 55.58 57.95 nmHC (g/hr,Cl) 12.89 12.33 51.55 26.20 20.37 19.15 tHC (g/hr,Cl) 33.43 34.00 290.49 89.84 75.96 77.11 PM (g/hr) -corrected 3.29 3.63 4.88 18.10 7.91 3.45 total PM mass on filters (mg) 1.451 1.584 2.420 3.067 1.363 0.588 backup filter mass (mg) 0.268 0.365 0.386 0.205 0.131 0.146 mass ratio primary/backup filter 4.41 3.34 5.26 13.99 9.43 3.03 Diesel flow (kg/hr) 0.83 0.82 0.63 1.05 1.22 1.17 CNG flow (kg/hr) 1.66 1.68 1.72 28.40 28.79 30.34 Table DT4. Mode 5,8 o f A V L 8-Mode Test Emissions and Fuel Flow Data (g/hr, kg/hr) Mode 5 Mode 8 36° BTDC 30° BTDC 16.5° BTDC 29° BTDC 18° BTDC 15° BTDC CO (g/hr) 127.18 133.48 471.80 572.15 443.77 432.78 C02 (kg/hr) 39.85 39.72 43.49 151.85 170.95 178.51 NOx (g/hr) 945.49 472.39 177.36 750.05 538.25 541.66 CH4 (g/hr) . 132.96 155.62 851.03 134.24 177.37 203.99 nmHC (g/hr,Cl) 44.19 53.31 143.42 37.58 68.27 37.86 tHC (g/hr,Cl) 177.15 208.93 994.44 171.81 245.63 241.85 PM (g/hr) -corrected 8.03 14.66 13.16 18.25 20.72 7.52 total PM mass on filters (mg) 1.227 1.981 1.679 1.539 1.625 0.580 backup filter mass (mg) n/a 0.364 0.244 0.178 0.324 0.184 mass ratio primary/backup filter n/a 4.45 5.89 7.66 4.02 2.15 Diesel flow (kg/hr) 2.08 1.93 2.08 1.84 1.94 1.75 CNG flow (kg/hr) 12.29 12.28 14.57 54.29 60.17 63.85 79 F i g u r e D T 1 i . C y l i n d e r P r e s s u r e T r a c e f o r 3 In ject ion T i m i n g S e t t i n g s : M o d e l (Idle) o f A V L 8 - M o d e T e s t C y l i n d e r P r e s s u r e T r a c e s : I n j e c t i o n T i m i n g S t u d y -M o d e 1 ( L o w S p e e d , L o w L o a d ) 7 0 *r 3Q a r - 9 * 5 * * * ^ 1 0 , 1 0 --60 -40 -20 40 60 c r a n k a n g l e ( d e g r e e s ) — ' S O M 4 deg B T D C — S O I = 1 6 deg B T D C — S O I = 3 deg A T D C F i g u r e DT1 ii . D i f f e r e n t i a l H e a t R e l e a s e f o r 3 In ject ion T i m i n g S e t t i n g s : M o d e 1 (Idle) o f A V L 8 - M o d e T e s t D i f f e r e n t i a l H e a t R e l e a s e P l o t s : I n j e c t i o n T i m i n g S t u d y -M o d e 1 ( L o w S p e e d , L o w L o a d ) -100 c r a n k a n g l e ( d e g r e e s ) | — S O I = 1 4 deg B T D C SOI=16 deg B T D C — SOI=3 deg A T D C j F i g u r e D T I i i i . I n tegra ted H e a t R e l e a s e f o r 3 In ject ion T i m i n g S e t t i n g s : M o d e 1 (Idle) o f A V L 8 - M o d e T e s t I n t e g r a t e d H e a t R e l e a s e P l o t s : I n j e c t i o n T i m i n g S t u d y -M o d e 1 ( L o w S p e e d , L o w L o a d ) c r a n k a n g l e ( d e g r e e s ) j — S O I - 1 4 deg B T D C — S O I = 1 6 deg B T D C — S Q I - 3 d e g A T D C F i g u r e D T I i v . C O E m i s s i o n s v s . T i m e f o r 3 In ject ion T i m i n g S e t t i n g s : M o d e 1 (Idle) o f A V L 8 - M o d e T e s t T i m i n g S t u d y : C O v s . T i m e P l o t s M o d e 1: L o w S p e e d , L o w L o a d 400 350 _ 300 | 250 S 200 8 150 100 50 - ' ~ ~ — — " -0 J 1 , , 1 , , , , , 0 20 40 60 80 100 120 140 160 180 200 T i m e (s ) [ — M 1 SQI=14 deg B T D C M l S Q I - 1 6 deg B T D C M1 S Q N 3 deg A T D C [ Figure DT4i. Cylinder Pressure Trace for 3 Injection Timing Settings: Mode 4 of AVL 8-Mode Test Cylinder Pressure Traces: Injection Timing Study -Mode 4 (Low Speed, High Load) S f 60 \ A . 20 . . 0-- 6 0 - 4 0 -20 0 20 40 60 crank angle (dogrees) [•^SOI-11 deg BTDC SOI=8 deg BTDC SOI=1 dag BTDC] Figure DT4II. Differential Heat Release for 3 Injection Timing Setting*: Mode 4 of AVL 8-Mode Tsst Differential Heat Release Plots: Injection Timing Study -Mode 4 (Low Speed, High Load) 3M crank angle (degrees) |—SOI=11 deg BTDC SOI»8deg BTDC S O M deg BTDCJ Figure DT4MI. Integrated Heat Release for 3 Injection Timing Settings: Mode 4 of AVL 8-Mode Test Integrated Heat Release Plots: Injection Timing Study -Mode 4 (Low Speed, High Load) MOO-, crank angle (degrees) j—— SON11 deg BTDC SOI=8 deg BTDC SOI-1 deg BTDCJ Figure DT4iv. CO Emissions vs. Tims for 3 Injection Timing Settings: Mode 4 of AVL 8-Mode Test Timing Study: CO vs. Time Plots Mode 4: Low Speed, High Load 2050 . 1860 £ 1250 \ Time (s) -M4 S0l=11 deg BTDC ~^M4SOI=8 deg BTDC M4 SOI=1 deg BTDC | Figure DT5i. Cylinder Pressure Trace for 3 Injection Timing Settings: Mode 5 of AVL 8-Mode Test Cylinder Pressure Traces: Injection Timing Study -Mode 5 (High Speed, Low Load) 100 •^^60 10 _ _ _ _ _ ^ 20 -. o--60 -20 0 20 crank angle (degrees) 60 —SON30 deg BTDC - SOI=36 deg BTDC —SOI=16.5 deg ATDC Figure DT5ii. Differential Heat Release for 3 Injection Timing Settings: Mode 5 of AVL 8-Mode Test Differential Heat Release Plots: Injection Timing Study -Mode 5 (High Speed, Low Load) crank angle (degrees) SOI=30 deg BTDC — SOI=36 deg BTDC SOI=16.5 deg ATDC Figure DT5iii. Integrated Heat Release for 3 Injection Timing Settings: Mode 5 of AVL 8-Mode Test Integrated Heat Release Plots: Injection Timing Study -Mode 5 (High Speed, Low Load) crank angle (degrees) j SO>30 deg BTDC SOI-36 deg BTDC SON16.5 deg ATDC Figure DT5iv. CO Emissions vs. Time for 3 Injection Timing Settings: Mode 5 of AVL 8-Mode Test Timing Study: CO vs. Time Plots Mode S: High Speed, Low Load 410 360 310 260 210 160 110 0 20 40 60 80 100 120 140 160 180 200 Time (s) | — M S SON30 deg BTDC — M1 S0I=36 deg BTDC M1 SON16.5 deg ATDC | Figure DT8i . Cy l inder P r e s s u r e Trace fo r 3 Injection T i m i n g Set t ings: M o d e 8 of A V L 8-Mode Tes t C y l i n d e r P r e s s u r e T r a c e s : I n j e c t i o n T i m i n g S t u d y -M o d e 8 ( H i g h S p e e d , H i g h L o a d ) >> / Ns y^ * so 60 -a#*^ - 40 -, , o~ crank angle (degrees) SOI=18deg BTDC —SOI=29deg BTDC —~SOI=15 deg ATDC| Figure PT8i i . Difterentiai Heat Re lease for 3 Injection T i m i n g Sett ings: M o d e 8 of A V L 8-Mode Test D i f f e r e n t i a l H e a t R e l e a s e P l o t s : I n j e c t i o n T i m i n g S t u d y -M o d e 8 ( H i g h S p e e d , H i g h L o a d ) crank angle (degrees) p~»SOMBdeg BTDC—--SOI=20 deg BTDC SON 15 deg ATDC| Figure DTSiii . Integrated Heat Release for 3 Injection T i m i n g Sett ings: M o d e 8 of A V L 8-Mode Tes t I n t e g r a t e d H e a t R e l e a s e P l o t s : I n j e c t i o n T i m i n g S t u d y -M o d e 8 ( H i g h S p e e d , H i g h L o a d ) crank angle (degrees) | — — S O I - 1 8 dog B T D C — S O I=29 deg B T D C SOt=15 deg ATDCJ Figure DT8iv. C O E m i s s i o n s v s . T ime for 3 Injection T i m i n g Sett ings: M o d e 8 of A V L 8-Mode Tes t T i m i n g S t u d y : C O v s . T i m e P l o t s M o d e 8: H i g h S p e e d , H i g h L o a d 20 40 60 80 100 120 140 160 180 200 Time (s) | S O M 8 deg BTDC -——MB SQ1=28 deg BTDC M8 SON15 deg BTDC| Figura DT1v. PM, NOx and CO Plots for 3 Injection Timing Settings: Mode 1 of AVL 8-Mode Test Timing Study: PM, NOx, and C O Emissions: Mode 1 (Idle) of AVL 8-Mode Test 0.8 0.7 0.6 • 0 5 1 | 0.3 | 3 -10 r5 Pilot SOI (degrees) IBNOX * CO ePMJ Figure DT4V. PM, NOx and CO Plots for 3 Injection Timing Settings: Mode 4 of AVL 8-Mode Test Timing Study: PM, NOx, and C O Emissions: Mode 4 (Low Speed, High Load) of AVL 8-Mode Test 5 « 0.09 0.08 0.07 . 0.06 i _ 0 0 4 If, 0.03 " ~ 0.02 «• 0.01 Pilot SOI (degrees) |slNOx ACQ » P M | Figure DT5v. PM, NOx and CO Plots for 3 Injection Timing Settings: Mode 5 of AVL 8-Mode Test Timing Study: PM, NOx, and C O Emissions: Mode 5 (High Speed, Low Load) of AVL 8-Mode Test 0.25 02 0.15 J | S £ 0.1 B & uj — s 0.05 Pilot SOI (degrees) • NOx A CO « P M | Figure DT8v. PM, NOx and CO Plots for 3 Injection Timing Settings: Mode 8 of AVL 8-Mode Test Timing Study: PM, NOx, and C O Emissions: Mode 8 (High Speed, High Load) of AVL 8-Mode Test 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0.06 0.05 i 0.04 K _ H 0.03 • £• if 0.02 Ul — 2 0.01 a 0 -20 -15 Pilot SOI (degrees) |«NOx ACQ » P M | Figure DT1vi. Fuel Flow and PM Emissions Data for Mode 1 of AVL 8-Mode Test Timing Study Total (CNG+Diesel) Fuel Flow, Diesel Flow, PM Emissions: Mode 1 (Idle) of AVL 8-Mode Test 3.0 ^ | 2.5 2 2.0 <± 1.0 o £ 0.5 0.0 6 5 co 4 I 1 ° -16 deg BTDC 14 deg BTDC Pilot SOI (degrees) 3 deg ATDC • PM emissions -total fuel flow —*— diesel flow Figure DT4vi. Fuel Flow and PM Emissions Data for Mode 4 of AVL 8-Mode Test Timing Study Total (CNG+Diesel) Fuel Flow, Diesel Flow, PM Emissions: Mode 4 (Low Speed, High Load) of AVL 8-Mode Test 32 32 31 31 30 30 29 29 28 20 15 g o cn 10 | UJ 5 I 0 11 deg BTDC 8 deg BTDC Pilot SOI (degrees) 1 deg BTDC 1 PM emissions -total fuel flow diesel flow Figure DT5vi. Fuel Flow and PM Emissions Data for Mode 5 of AVL 8-Mode Test 17 17 16 16 15 15 14 14 13 13 Timing Study Total (CNG+Diesel) Fuel Flow, Diesel Flow, PM Emissions: Mode 5 (High Speed, Low Load) of AVL 8-Mode Test 16 i S i Z 36 deg BTDC 30 deg BTDC Pilot SOI (degrees) 16.5 deg BTDC 3 PM emissions -total fuel flow ~ * ~ diesel flow Figure DT8vi. Fuel Flow and PM Emissions Data for Mode 8 of AVL 8-Mode Test 68 66 € 64 62 T 60 o 58 u_ 56 § 54 52 50 Timing Study Total (CNG+Diesel) Fuel Flow, Diesel Flow, PM Emissions: Mode 8 (High Speed, High Load) of AVL 8-Mode Test 25 20 „ c o 15 n m 10 ifi £ 5 °-0 29 deg BTDC 18 deg BTDC Pilot SOI (degrees) 9 PM emissions -total fuel flow diesel flow D.P Pressure Study Emissions and Fuel Flow Data Tables and Plots Table DPI. Emissions and Fuel Flow Data for Modes 1,4,5,8 of A V L 8-Mode Test (g/hr, Test Data Mode 1 Mode 4 Mode 5 Mode 8 10 MPa 25 MPa 17 MPa 25 MPa 13 MPa 25 MPa 23 MPa 25 MPa CO (g/hr) 22.76 34.53 275.65 628.26 246.81 186.44 311.90 448.65 C02 (kg/hr) 5.67 6.26 83.82 86.96 42.31 42.19 161.00 157.44 NOx (g/hr) 77.68 78.94 582.62 749.00 184.95 334.52 678.30 713.66 CH4 (g/hr) 28.41 41.19 53.53 59.81 329.62 219.28 186.32 186.05 nmHC (g/hr,Cl) 11.23 12.74 24.41 22.49 69.91 52.17 52.53 45.83 tHC (g/hr.Cl) 39.64 53.93 77.94 82.31 399.53 271.45 238.85 231.89 PM (g/hr) -corrected 1.69 1.46 3.64 4.46 10.56 8.51 8.59 9.44 PM mass on filter(mg) 0.833 0.724 0.569 0.704 1.298 1.105 0.660 0.745 backup filter mass (mg) 0.208 0.179 0.113 0.119 0.223 0.207 0.119 0.133 mass ratio primary/backup filter 3.01 3.05 4.05 4.92 4.81 4.35 4.56 4.61 Diesel flow (kg/hr) 0.06 0.49 1.10 1.47 1.53 2.39 2.08 2.04 CNG flow (kg/hr) 1.60 1.47 29.02 30.16 13.83 12.40 57.28 55.97 Table DP2. Emissions Data for Modes 1,4,5,8 of AVL 8-Mode Test (g/bhp.hr, kg/bhp.hr) Test Data M l M l M4 M4 M5 M5 M8 M8 low P high P lowP high P lowP highP lowP high P CO (g/bhp.hr) 3.34 3.25 4.84 2.52 1.21 0.76 2.59 1.11 C02 (kg/bhp.hr) 0.83 0.56 0.88 0.57 0.37 0.39 0.36 0.39 NOx (g/bhp.hr) 11.41 2.44 11.07 4.53 2.56 1.66 3.09 1.76 CH4 (g/bhp.hr) 4.17 4.34 5.78 2.97 0.23 0.46 0.25 0.46 nmHC (g/bhp.hr) 1.65 0.92 1.79 0.71 0.11 0.13 0.09 0.11 tHC (g/bhp.hr) 5.82 5.26 7.56 3.67 0.34 0.59 0.34 0.57 PM (g/bhp.hr) 0.248 0.205 0.016 0.018 0.139 0.115 0.021 0.023 PM mass on filter (mg) 0.833 0.724 0.569 0.704 1.298 1.105 0.660 0.745 mass ratio primary/backup filter 3.01 3.05 4.05 4.92 4.81 4.35 4.56 4.61 86 Figure DP. Fuel Flow and PM Emissions Data for Modes 1,4,5,8 of AVL 8-Mode Test Pressure Study Total (CNG+Diesel) Fuel Flow, Diesel Flow, PM Emissions: Modes 1,4,5,8 of A V L 8-Mode Test mode 1 mode 1 mode 4 mode 4 mode 5 mode 5 mode 8 mode 8 10 MPa 25 MPa 17 MPa 25 MPa 13 MPa 25 MPa 23 MPa 25 MPa Test Point Trial # I PM emissions -total fuel flow •diesel flow 87 Figure DP1a. Cylinder Prwiure Trace for G«a Injection Pressure Settings of 10 and 25 MPa: Mode 1 (Idle) of AVL 8-Mode Test C y l i n d e r P r e s s u r e T r a c e s : G a s R a i l P r e s s u r e S t u d y - M o d e 1 ( L o w S p e e d , L o w L o a d ) 10 o_ crank angle (degrees) | - - p s 1Q MPa — P=25MPa] Figure DP1b. Differential Heat Release for Gas Injection Pressure Settings of 10 and 25 MPa: Mode 1 (Idle) of AVL 8-Mode Test Di f fe ren t ia l H e a t R e l e a s e P r o f i l e s : G a s R a i l P r e s s u r e S t u d y - M o d e 1 ( L o w S p e e d , L o w L o a d ) crank angle (degrees) ~P=10 MPa - '=25 MPa Figure DP1c. Integrated Heat Release for Gas Injection Pressure Settings of 10 and 25 MPa: Mode 1 (Idle) of AVL 8-Modo Test In tegra ted H e a t R e l e a s e P r o f i l e s : G a s R a i l P r e s s u r e S t u d y - M o d e 1 ( L o w S p e e d , L o w L o a d ) crank angle (degrees) | P=10 MPa ^ P - 2 5 MPa] Figure DP1d. CO Emissions vs. Time for Gas Injection Pressure Settings of 10 and 25 h Mode 1 (Idle) of AVL 8-Mode Test Pressure Study: CO vs. Time Plots Mode 1: Low Speed, Low Load 130 120 110 E 100 D. a 90 O 80 o 70 60 50 100 Time (s) Figure D P 4 a . Cy l inder P r e s s u r e T r a c e for Injection P r e s s u r e Set t ings of 17 and 25 M P a : M o d e 4 o f A V L 8-Mode Test C y l i n d e r P r e s s u r e T r a c e s : G a s R a i l P r e s s u r e S t u d y - M o d e 4 ( L o w S p e e d , H i g h L o a d ) , * 3 Q -go ^ - ^ V 00 10 ^ ^ ^ ^ 20 , , , 0--60 -40 -20 0 20 40 60 crank angle (degrees) p ^ p = 1 7 M P 7 — P = 2 S MPa{ Figure DP4b. Differential Heat Release for Injection Pressure Sett ings of 17 and 25 MPa: M o d e 4 of A V L 8-Mode Test D i f f e r e n t i a l H e a t R e l e a s e P r o f i l e s : G a s R a i l P r e s s u r e S t u d y - M o d e 4 ( L o w S p e e d , H i g h L o a d ) •wen f 1 2 i I • tt I X -( • 50 // N X 0 -40 -20 i 20 40 < crank angle (degrees) j - P=17MPa P=25 MPa | Figure D P 4 c . Integrated Heat Re lease for Injection Pressure Set t ings of 17 a n d 25 M P a : M o d e 4 of A V L 8-Mode Test I n t e g r a t e d H e a t R e l e a s e P r o f i l e s : G a s R a i l P r e s s u r e S t u d y • M o d e 4 ( L o w S p e e d , H i g h L o a d ) E 3 I i 1 I 4 3000-2500-2000 y ^ ^ 1 1500 / 1000 / 500 / 0 ... •* -» — 20 40 60 crank angle (degrees) |~~~ P=17 MPa ^ P = 2 5 MPa | Figure D P 4 d . C O E m i s s i o n s v s . T i m e for Injection Pressure Set t ings of 17 a n d 25 M P a : M o d e 4 of A V L 8-Mode Test F i g 5 . 4 d : P r e s s u r e S t u d y : C O v s . T i m e P l o t s M o d e 4 : L o w S p e e d , H i g h L o a d Figure DPSa. Cylinder Pressure Trace for Injection Pressure Settings of 13 and 25 MPa: Mode 5 of A V L 8-Mode Test C y l i n d e r P r e s s u r e T r a c e s : G a s R a i l P r e s s u r e S t u d y - M o d e 5 ( H i g h S p e e d , L o w L o a d ) so AT _ r TO 20 o_ -—P=13MPa" Figure DP5b. Differential Heat Release for Injection Pressure Settings of 13 and 25 MPa: Mode 5 of A V L 8-Mode Test Differential Heat Release Prof i les : G a s Rai l Pressure Study - Mode 5 High Speed, L o w Load) i 4* crank ang le (degrees) Figure DP5c. Integrated Heat Release for Injection Pressure Settings of 13 and 25 MPa: Mode 5 of A V L 8-Mode Test Integrated H e a t R e l e a s e P r o f i l e s : G a s R a i l P r e s s u r e S t u d y - M o d e 5 ( H i g h S p e e d , L o w L o a d ) crank angle (degrees) E Figure DPSd. C O Emissions vs . Time for Injection Pressure Settings of 13 and 25 MPa: Mode 5 of A V L 8-Mode Test P r e s s u r e S t u d y : C O v s . T i m e P l o t s M o d e 5 : H i g h S p e e d , L o w L o a d ? 170 90 , 70 100 Time (s) - M5 P=13 MPa — M5 P=2S MPa Figure DP8a. Cylinder Pressure Trace for Injection Pressure Settings of 23 and 25 MPa: Mode 8 of AVL 8-Mode Test Cylinder Pressure Traces: Gas Rail Pressure Study - Mode 8 (High Speed, High Load) 110 _r^120 -T 100 jr (jo 60 10 20 . . cu crank angle (degrees) E Figure DP8b. Differential Heat Release for Injection Pressure Settings of 23 and 25 MPa: Mode 8 of AVL 8-Mode Test Differential Heat Release Profiles: Gas Rail Pressure Study - Mode 8 (High Speed, High Load) crank angle (degrees) -P'23MPa r=25MPa~| Figure DP8c. integrated Heat Release for Injection Pressure Settings of23and25MPa: Mode 8 of AVL 8-Mode Test Integrated Heat Release Profiles: Gas Rail Pressure Study - Mode 8 (High Speed, High Load) crank angle (degrees) Figure DP8d. CO Emissions vs. Time for Injection Pressure Settings of 23 and 25 MPa: Mode 8 of AVL 8-Mode Test Fig. 5.4o: Pressure Study: CO vs. Time Plots Mode 8: High Speed, High Load E 200 \ ° 150 100 120 Time (s) M8P=23 MPa — M 8 P=25 MPa [ Figure DP1e. PM, NOx and CO Plots for Mode 1 (Idle) Injection Pressure Settings of 10 and 25 MPa Pressure Study: PM, NOx and CO Emissions, Mode 1 (Idle) of AVL 8-Mode Test o ~ OS If 10 15 20 Injection pressure (MPa) |HNOx A CO »PM| Figure DP4e. PM, NOx and CO Plots for Mode 4 Injection Pressure Settings of 17 and 25 MPa Pressure Study: PM, NOx and CO Emissions, Mode 4 (Low Speed, High Load) of AVL 8-Mode Test 12 e 10 o •88 8 E JZ 111 6 O CO o fl> 4 o <£ z 2 0 T 0.02 0.018 0.016 £ - 0.014 _ 0.012 § £ 0.01 » d 0.008 _ | 0.006 <" 0.004 1 0.002 0 10 15 20 25 30 Injection pressure (MPa) I NOx - C O »PM Figure DP5e. PM, NOx and CO Plots for Mode 5 Injection Pressure Settings of 13 and 25 MPa Pressure Study: PM, NOx and CO Emissions, Mode 5 (High Speed, Low Load) of AVL 8-Mode Test O a. E — 0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 10 15 20 Injection pressure (MPa) • NOx A CO »PM| 25 30 Figure DP8e. PM, NOx and CO Plots for Mode 8 Injection Pressure Settings of 23 and 25 MPa Pressure Study: PM, NOx and CO Emissions, Mode 8 (High Speed, High Load) of AVL 8-Mode Test 23.5 24 24.5 Injection pressure (MPa) loNOx ACO • P M l 0.025 0.02 s CL 0.015 o » £ QJ — 0.005 -D. PW Pilot Fuel Quantity Study: Emissions Tables and Plots Pilot Pulse Width Test Data 0.48 ms 0.65 ms 1.1 ms C O (g/bhp.hr) 1.49 0.94 0.75 C02 (kg/bhp.hr) 0.43 0.43 0.44 NOx (g/bhp.hr) 9.04 8.88 9.03 CH4 (g/bhp.hr) 1.41 0.78 0.66 nmHC (g/bhp.hr) 0.38 0.32 0.30 tHC (g/bhp.hr) 1.78 1.10 0.96 P M (g/bhp.hr) 0.0560 0.0565 0.0543 P M mass on filters (mg) 0.920 0.939 0.908 mass ratio primary/backup filter 3.69 3.26 3.45 Table DPW2. Emissions and Fuel Flows For Various Pilot Fuel Pulse Widths (g/hr, kg/hr) Test Data Pilot Pulse Width 0.48 ms 0.65 ms 1.1 ms C O (g/hr) 61.42 39.46 31.75 C02 (kg/hr) 17.64 18.04 18.32 NOx (g/hr) 373.76 371.53 380.05 CH4 (g/hr) 65.86 65.49 65.69 nmHC (g/hr,Cl) 58.12 32.66 27.77 tHC (g/hr,Cl) 15.59 13.49 12.81 P M (g/hr) -corrected 2.32 2.36 2.28 Diesel flow (kg/hr) 0.21 0.78 1.26 C N G flow (kg/hr) 2.32 2.36 2.28 93 Figure DPWa. Cylinder Pressure Trace for 3 Pilot Pulse Width Settings: Mode 2 of AVL 8-Mode Test Cylinder Pressure Traces: Pilot Pulse Width Study Mode 2 (Low Speed, 25% Load) of AVL 8-Mode Test 160 -60 -30 0 30 60 c r a n k a n g l e ( d e g ) PW=0.48 ms • — P W = 0 . 6 5 m s PW=1.1 m s Figure OPWb. Differential Heat Release for 3 Pilot Pulse Width Settings: Mode 2 of AVL 8-Mode Test Differential Heat Release: Pilot Pulse Width Study Mode 2 (Low Speed, 25% Load) of AVL 8-Mode Test cn <D < E i t o' 150 100 50 -30 -50 -10 •PW=0.48 ms 10 c r a n k a n g l e ( d e g ) 30 50 •PW=0.65 ms •PW=1.1 ms Figure DPWc. PM, NOx and CO Plots for 3 Pilot Pulse Width Settings i f O 3 O a, >< « o z Effect of Pilot Fuel Injection Duration on PM, NOx and CO Emissions: Mode 2 (Low Speed, 25% Load) of AVL 8-Mode Test 10 9 8 7 6 5 4 3 2 1 0 0.2 0.4 0.6 0.8 Pilot Pulse Width (ms) I NOx A C O • P M 0.057 0.0565 © (0 0.056 * •_-I 1 0.0555 <J> J" 0.055 ill ^ S 0.0545 °-0.054 1.2 Figure DPWd. Fuel Flow and PM Emissions Data For 3 Pilot Pulse Width Settings 5 o — CO IX. Pilot Pulse Width Study Total (CNG+Diesel) Fuel Flow, Diesel Flow, PM Emissions: Mode 2 of AVL 8-Mode Test 7.0 6.0 5.0 4.0 3.0 -I 2.0 1.0 0.0 0.48 0.65 Pilot Pulse Width (ms) 1.1 2.5 2.0 W c o 1.5 » XT I 1.0 L_ ^ + 0.5 °-0.0 PM emissions -total fuel flow - _ r diesel flow 9 5 Appendix E . Direct Capillary Gas-Injection Method of Analyzing Particulate Matter Components to Determine Lubricating Oil-Like and Unburned Diesel Fuel-Like Fractions in the Volatile Organic Hydrocarbon Fraction of Exhaust P M E.l Calculation of Fuel and Oil-Derived Hydrocarbons In Engine Particulate Matter Ricardo Consulting Engineers developed a thermal desorption procedure that allows the volatile organic P M fraction to be directly introduced to a capillary gas chromatograph/flame ionization detector (GC/FID) . In addition to mitigating the problem of impurities associated with more conventional solvent extraction procedures, the benefits of this method include speed and improved sensitivity, as a large portion of the filter can be loaded onto the G C . E . l . l Physical Process of Particulate Matter Organic Fraction Volatization A section of the particulate-loaded filter is inserted into a pre-chamber upstream of the G C / F I D and heated to approximately 3 2 5 ° C . The resulting volatized organic hydrocarbons are injected into the capillary G C and a spectrum obtained for this particulate V O F sample. The spectrum is then compared to two spectra from an unburned diesel fuel and a lubricating oil sample, respectively. E.1.2 Interpretation of Chromatographic Spectra Cuthbertson and Shore (1988) showed that the spectrum obtained from analysis of particulate-bound hydrocarbons consists of two distinct humps that correspond to hydrocarbon spectra characteristic of unburned diesel fuel and unburned lubricating oil . A semi-quantitative assessment of the fuel- and oil-derived particulate-bound hydrocarbons is possible by assuming that all hydrocarbons eluting before a given retention time X (the " cut -of f time) minutes are diesel fuel-derived and all those eluting after X minutes are oil-derived. This type of analysis is possible by virtue of the distinct spectra obtained for the individual unburned diesel fuel and oil components in the particulate-bound hydrocarbon. Typically the "diesel fuel-derived" hydrocarbons have a low carbon number, and as such the "hump" in the chromatographic spectrum for unburned diesel fuel occurs at a lower retention time (around 7.5 minutes) than the "hump" in the spectrum for unburned oil, which occurs around 18 minutes. 96 Once the cut-off time described above has been chosen for a particular spectrum, the response of the FID to the unburned diesel fuel is calculated (to determine the area under the spectral curve corresponding to a known mass input of unburned fuel, which yields an area/mass ratio response factor) and used to quantify the mass of unburned diesel fuel. This calculation involves integrating the area under the spectral curve from 0 minutes to the cut-off time for unburned diesel fuel, and then converting this to an unburned diesel fuel mass using the response factor. The same procedure is used to determine the mass of unburned oil. The response of the FID to the unburned oil is calculated (to determine the area under the spectral curve corresponding to a known mass input of unburned oil, which yields an area/mass ratio response factor) and used to quantify the mass of unburned oil. This calculation involves integrating the area under the spectral curve for particulate-bound hydrocarbon from the cut-off time to the last discernable retention time value of this spectrum, and then converting this to an unburned oil mass using the response factor. 97 Table E. l Filter Codes and Corresponding Test Point Descriptions Filter code Primary or backup filter Description of Study Test Mode Comments EM-263 primary Repeatability study: A V L test #1 1 EM-264 backup ii 1 EM-269 primary ii 4 EM-270 backup a 4 EM-271 primary ii 5 EM-272 backup a 5 EM-277 primary a 8 EM-278 backup a 8 EM-281 primary Repeatability study: A V L test #2 1 EM-282 backup ii 1 EM-289 primary ii 4 EM-290 backup a 4 EM-291 primary a 5 EM-292 backup a 5 EM-297 primary a 8 EM-298 backup a 8 EM-261 n/a Travel blank Contains trace contamination from filter holder Table E2. Raw Data from Particulate VOF Analysis: Primary Filters from Repeatability Study B Test Points: Modes 1,4,5,8 of A V L 8-Mode Test AVL test #1 mode 1 primary AVL test #1 mode 4 primary AVL test #1 mode 5 primary AVL test #1 mode 8 primary AVL test #2 mode 1 primary AVL test #2 mode 4 primary AVL test #2 mode 5 primary AVL test #2 mode 8 primary Filter ID EM-263 EM-269 EM-271 EM-277 EM-281 EM-289 EM-291 EM-297 fuel-like VOF mass (mg) 0.264 0.087 0.299 0.064 0.234 0.059 0.218 0.055 oil-like VOF mass (mg) 0.050 0.141 0.348 0.204 0.240 0.245 0.254 0.232 total VOF (mg) 0.313 0.229 0.648 0.268 0.474 0.304 0.473 0.287 Table E3. Raw Data from Particulate VOF Analysis: Backup Filters from Repeatability Study B Test Points: Modes 1,4,5,8 of A V L 8-Mode Test AVL test AVL test AVL test AVL test AVL test AVL test AVL test AVL test #1 #1 #1 #1 #2 #2 #2 #2 travel mode 1 mode 4 mode 5 mode 8 mode 1 mode 4 mode 5 mode 8 blank backup backup backup backup backup backup backup backup EM-Filter ID EM-264 EM-270 EM-272 EM-278 EM-282 EM-290 EM-292 EM-298 261 fuel-like VOF mass (mg) 0.086 0.070 0.086 0.073 0.077 0.095 0.108 0.064 0.037 oil-like VOF mass (mg) < 0.011 < 0.011 0.015 0.011 0.024 0.036 < 0.011 < 0.011 0.039 total VOF (mg) < 0.097 < 0.081 0.101 0.084 0.101 0.130 < 0.119 < 0.075 0.076 99 Figure E l . Chromatographic Spectrum of Particulate VOF on Filter E M - 263 Software Version : 6.1.1.0.0:K20 * Date Sample Name : WESTPORT EM263 PRIMARY Data Acquisition Time Instrument Name : GC3 Channel RackA/ial : 0/0 Operator Sample Amount : 1.000000 Dilution Factor Cycle : 3 08/03/01 09:40:23 02/03/01 11:32:42 , 7 ^ . T^Ti COL 1.000000 Raw Data File : \\klebb\TCData\DATA3\02MAR01003.raw Sequence File: \\klebb\TCData\DATA3\Group.seq 16 18 20 Time [min] Time Component [min] Name Area [MVs] Particulate Analysis 4.755 GROUP 1 13389754.54 18.335 TOTAL 18376941.54 21.585 GROUP 2 4943103.00 36709799.08 100 ' Figure E2. Chromatographic Spectrum of Particulate V O F on Filter E M - 264 i 1 of 1 Software Version : 6.1.1.0.0:K20 Sample Name : WESTPORT EM-26f BACKUP Instrument Name : GC3 Rack/Vial : 0/0 Sample Amount : 1.000000 Cycle : 5 Raw Data File : \\klebb\TCData\DATA3\06MAR01005.raw Sequence File: \\klebb\TCData\DATA3\Group.seq Date Data Acquisition Time Channel Operator Dilution Factor 07/03/01 09:56:52 I 06/03/01 13:07:32 11 A COL 1.000000 RICARDO : •'. 11 -;. I •, I. | i:, : I!: 11:' 11,1, i 11'' 11M j J1: 11; I, |; 11,11111 i: 111; h, i ; I,:! |: u,! J: i! i!:! ;: ^  11 • :!;: i i i, I! ^  n • I i I; 11.1. 10 12 14 16 18 20 22 24 26 28 30 32 34 Time [min] Particulate Analysis Time Component [min] Name Area [uV-s] 4.755 GROUP 1 18.335 TOTAL 21.585 GROUP 2 5565655.70 7395032.39 1813613.94 14774302.03 101 Figure E3. Chromatographic Spectrum of Particulate V O F on Filter E M - 269 of 1 Software Version : 6.1.1.0.0:K20 Sample Name : WESTPORT EM-269 PRIMARY Instrument Name : GC3 Rack/Vial : 0/0 Sample Amount : 1.000000 Cycle : 3 Raw Data File : \\klebb\TCData\DATA3\05MAR01003.raw Sequence File: Wklebb\TCData\bATA3\Group.seq Date Data Acquisition Time Channel Operator Dilution Factor 07/03/01 09:52:38 05/03/01 12:29:21 A COL 1.000000 RICARDO 1 2 0 - 3 . ¥ ioo-§ 6 0 - 3 2 0 - 3 12 16 18 Time (min] 20 28 32 Time Component [min] Name Area [uV-s] Particulate Analysis 4.755 GROUP 1 18.335 TOTAL 21.585 GROUP 2 7587405.45 14331513.30 6744107.85 28663026.60 102 Figure E4. Chromatographic Spectrum of Particulate V O F on Filter E M - 270 age 1 of 1 Software Version : 6.1.1.0.0:K20 Sample Name : WESTPORT EM-270 BACKUP Instrument Name : GC3 RackA/ial : 0/0 Sample Amount : 1.000000 Cycle : 6 Raw Data File : \\klebb\TCData\DATA3\06NMR01006.raw Sequence File: \\klebb\TCData\DATA3\Group.seq Date Data Acquisition Time Channel Operator Dilution Factor 07/03/01 09:57:16 06/03/01 14:28:33 A COL 1.000000 R I C A R D O 2 0 0 - 3 5" E iTJjJJi 10 14 16 18 20 Time [mini 24 30 32 Particulate Analysis Time Component Area [min] Name [uV-s] 4.755 GROUP 1 4736646.12 18.335 TOTAL 6056368.88 21.585 GROUP 2 1310680.31 12103695.31 103 Figure E5. Chromatographic Spectrum of Particulate V O F on Filter E M - 271 jge 1 of 1 Software Version : 6.1.1.0.0:K20 Sample Name : WESTPORT EM-271 PRIMARY Instrument Name : GC3 RackA/ial : 0/0 Sample Amount : 1.000000 Cycle : 4 Raw Data File : \\klebb\TCData\DATA3\05MAR01004.raw Sequence File: \\klebb\TCData\DATA3\Group.seq Date Data Acquisition Time Channel Operator Dilution Factor 07/03/01 09:52:59 05/03/01 14:49:21 A COL 1.000000 RICARDO Time Component [min] Name Area [MVs] Particulate Analysis 4.755 GROUP 1 18.335 TOTAL 21.585 GROUP 2 19144452.50 34930872.71 15786420.21 69861745.43 104 Figure E6. Chromatographic Spectrum of Particulate V O F on Filter E M - 272 Page 1 of 1 Software Version Sample Name Instalment Name Rack/Vial Sample Amount Cycle 6.1.1.0.0:K20 WESTPORT EM-272 BACKUP GC3 0/0 1.000000 2 Raw Data File : \\klebb\TCData\DATA3\07MAR01002.raw Sequence File: \\klebb\TCData\DATA3\Group.seq Date Data Acquisition Time Channel Operator Dilution Factor 07/03/01 09:59:15 07/03/01 09:23:03 A C O L 1.000000 RICARDO 5" E j H i •! I: • j: -: JI! j 11.; 11!:;: i 11; i:: J |.1 h 111! I;! i; i |!'!; I : i: I, i:' i •.,!! 11;!, 11; I:; 111 h 11:; i I j I,, • 11,111! I:: I, 10 12 14 16 18 20 22 24 26 28 30 32 34 .Time IminJ Particulate Analysis Time Component Area [min] Name [uVs] 4.755 GROUP 1 5722001.16 18.335 TOTAL 7984757.33 21.585 GROUP 2 2262756.17 15969514.65 105 Figure E7. Chromatographic Spectrum of Particulate V O F on Filter EM-277 Page 1 of 1 Software Version : 6.1.1.0.0:K20 Sample Name : WESTPORT EM-277 PRIMARY Instrument Name : GC3 RackA/ial : 070 Sample Amount : 1.000000 Date : 07/03/01 09:52:14 Data Acquisition Time : 05/03/01 11:41:02 Channel : A Operator : COL Dilution Factor : 1.000000 m m mum WTO*. IRICARDOI^L^LB Cycle : 2 Raw Data File : \\klebb\TCData\DATA3\05MAR01002.raw Sequence File: \\klebb\TCData\DATA3\Group:seq o 80-CO a: 60-HE3 2 0 - 3 o 1 ft fl 10 12 14 16 18 20 22 24 26 28 30 32 !   Time (min] Particulate Analysis Time Component [min] Name Area [MVs] 4.755 GROUP 1 18.335 TOTAL 21.585 GROUP 2 7580935.80 16401229.11 8820293.31 32802458.22 106 Figure E8. Chromatographic Spectrum of Particulate V O F on Filter E M - 278 Page 1 of 1 Software Version Sample Name Instrument Name Rack/Vial Sample Amount Cycle 6.1.1.0.0-.K20 WESTPORT EM-278 BACKUP GC3 0/0 1.000000 3 Raw Data File: \\klebb\TCData\DATA3\07MAR01003.raw Sequence File: \\klebb\TCData\DATA3\Group.seq Date Data Acquisition Time Channel Operator Dilution Factor 07/03/01 12:12:40 07/03/01 10:03:31 A COL 1.000000 RICARDO 2 0 0 - 3 1 6 0 - 3 1 4 0 - 3 f ^ 2 0 - 3 Time Component [min] Name 10 12 14 18 20 22 24 26 28 30 32 34 Time [min) Area [uVs] Particulate Analysis 4.755 GROUP 1 5083295.60 18.335 TOTAL 7078587.41 21.585 GROUP 2 1995291.81 14157174.81 107 Figure E9. Chromatographic Spectrum of Particulate V O F on Filter E M - 281 Page 1 of 1 Software Version : 6.1.1.0.0:K20 Sample Name : westport em-0281 primary Instrument Name : GC3 Rack/Vial : 0/0 Sample Amount : 1.000000 Cycle : 5 Raw Data File : \\klebb\TCData\DATA3\05MAR01005.raw Sequence File: \\klebb\TCData\DATA3\Group.seq Date Data Acquisition Time Channel Operator Dilution Factor 07/03/01 09:54:01 05/03/01 15:29:38 A COL 1.000000 RICARDO 30 32 34 Time (min] Particulate Analysis Time Component [min] Name Area [MVs] 4.755 GROUP 1 14982401.99 18.335 TOTAL 26528226.65 21.585 GROUP 2 11452674.12 52963302.77 108 Figure E10. Chromatographic Spectrum of Particulate V O F on Filter E M - 282 Page 1 of 1 Software Version 6.1.1.0.0:K20 Date 07/03/01 12:10:40 K B Sample Name Instrument Name WESTPORT EM-282 BACKUP GC3 Data Acquisition Time Channel 07/03/01 11:10:42 " j " * ^ A IHWAIH'I*! Rack/Vial 0/0 Operator COL Sample Amount 1.000000 Dilution Factor 1.000000 Cycle 4 Raw Data File: \\klebb\TCData\DATA3\07MAR01004.raw Sequence File: \\klebb\TCData\DATA3\Group.seq f ^ V 100-3 20 -3 |||i|l||||llll|llll|llll|llll|llll|llll|lllllllll|llll|llll|IIH 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 Time [min] Particulate Analysis Time Component Area [min] Name [MVS] 4.755 GROUP 1 5418292.30 18.335 TOTAL 7915628.59 21.585 GROUP 2 2497336.29 15831257.18 109 Figure E l l . Chromatographic Spectrum of Particulate V O F on Filter E M - 289 Page 1 of 1 Software Version Sample Name Instrument Name Rack/Vial Sample Amount Cycle 6.1.1.0.0:K20 WESTPORT EM-289 PRIMARY GC3 0/0 1.000000 2 Raw Data File : \\klebb\TCData\DATA3\06MAR01002.raw Sequence File: \\klebb\TCData\DATA3\Group.seq Date Data Acquisition Time Channel Operator Dilution Factor 07/03/01 09:55:00 06/03/01 09:07:36 A COL 1.000000 RICARDO E .',: I j! 1  ;1; I: i!;:: j : ! . i: I!, i i! •:, I::! I,. I:!!.: ':! - : I; 1 1 :!'1 • l :' : 1111:1: r ! 1".: 1 : ' ''1 12 14 16 18 20 22 24 26 28 30 32 Time IminJ 34 Time Component [min] Name Area [uV-s] Particulate Analysis 4.755 GROUP 1 18.335 TOTAL 21.585 GROUP 2 7958421.59 18170931.34 10212509.75 36341862.68 110 Figure E12. Chromatographic Spectrum of Particulate V O F on Filter E M - 290 Page 1 of 1 Software Version Sample Name Instrument Name Rack/Vial Sample Amount Cycle 6.1.1.0.0:K20 WESTPORT EM-290 BACKUP GC3 0/0 1.000000 5 Raw Data File : \\klebb\TCData\DATA3\07MAR01005.raw Sequence File: \\klebb\TCData\DATA3\Group.seq Date Data Acquisition Time Channel Operator Dilution Factor 07/03/01 12:51:12 |] 07/03/01 12:13:21 A COL 1.000000 RIGARDO 220-EE 200-_ 1 6 0 - = 140-120-100 -Ej 8 0 _ § 1 6 0 - = llll|IHI|ll!!|lll!jlll!|ll[l|!!l!|llll|ll!l|!llljl 2 4 6 8 10 2 14 16 18 20 22 Time [mini Particulate Analysis Time Component [min] Name Area [uV-s] 4.755 GROUP 1 18.335 TOTAL 21.585 GROUP 2 6006976.70 8567164.36 2560187.66 17134328.73 111 Figure E13. Chromatographic Spectrum of Particulate V O F on Filter E M - 290 Oeneat\ Page 1 of 1 Software Version Sample Name Instrument Name RackA/ial Sample Amount Cycle 6.1.1.0.0:K20 WESTPORT EM-290 BACKUP RPT GC3 0/0 1.000000 8 Raw Data File: \\klebb\TCData\DATA3\07MAR01008.raw Sequence File: \\klebb\TCData\DATA3\Group.seq Date Data Acquisition Time Channel Operator Dilution Factor 07/03/01 16:22:14 07/03/01 15:22:38 A COL 1.000000 RIGARDO 1 4 0 - 3 f ^ V 100-^3 6 0 - 3 2 0 - 3 , III : ,:,! : I ^ ' ••!;.; I, | : - - : ! | -Nlllh'h'i' ' j l| i.,: - : | • -j :-l ^ ,:l ' 10 12 14 16 18 20 22 24 26 28 30 32 34 Time [min] 2 4 6 8 Time Component Area [min] Name [pV-s] Particulate Analysis 4.755 GROUP 1 6269306.83 18.335 TOTAL 9296461.22 21.585 GROUP 2 3027154.39 18592922.43 112 Figure E14. Chromatographic Spectrum of Particulate V O F on Filter E M - 291 3age 1 of 1 Software Version Sample Name Instrument Name RackA/ial Sample Amount Cycle 6.1.1.0.0:K20 WESTPORT EM-291 PRIMARY GC3 0/0 1.000000 3 Raw Data File : \\klebb\TCData\DATA3\06MAR01003.raw Sequence File: \\klebb\TCData\DATA3\Group.seq Date Data Acquisition Time Channel Operator Dilution Factor 07/03/01 09:55:19 II 06/03/01 11:14:30 K A COL 1.000000 n i i l i i i i | i i i i | i i i i | i i i i | i i i i | i i i i l i i i i | i i i i | i i i i | i i i i | i i i i i i i i i | i i i i | i » V0 ' 1 2 ' ^4. 16 18 20 22 24 26 28 - 30 32 34 Time [min] Particulate Analysis Time Component [min] Name Area [uVs] 4 755 GROUP 1 14596171.69 18.335 TOTAL 26472152.27 21.585 GROUP 2 11875980.58 52944304.55 113 Figure E15. Chromatographic Spectrum of Particulate V O F on Filter E M - 292 Page 1 of 1 Software Version Sample Name Instrument Name Rack/Vial Sample Amount Cycle 6.1.1.0.0:K20 WESTPORT EM-292 BACKUP GC3 0/0 1.000000 6 Raw Data File : \\klebb\TCData\DATA3\07MAR01006.raw Sequence File: \\klebb\TCData\DATA3\Group.seq Date Data Acquisition Time Channel Operator Dilution Factor 07/03/01 13:47:16 07/03/01 12:58:54 A COL 1.000000 RICARDO 1 6 0 - 3 , E l l l l l l l l i l l l l l l l l l Time Component [min] Name 4.755 GROUP 1 18.335 TOTAL 21.585 GROUP 2 Area [MVs] 6255228.46 7923394.74 1657650.78 15836273.97 ! l l | l l l l | l l t l | l l l l | l ! j l | l l l l | l l l l j l l ! l | l l l l | l l ! l | l l ! l | l l l l | ^ 10 12 14 16 18 20 22 24 Time (min] Particulate Analysis 28 30 32 34 114 Figure E16. Chromatographic Spectrum of Particulate V O F on Filter E M - 297 'age 1 of 1 Software Version Sample Name Instrument Name Rack/Vial Sample Amount Cycle 6.1.1.0.0:K20 WESTPORT EM-297 PRIMARY GC3 0/0 1.000000 4 Date Data Acquisition Time Channel Operator Dilution Factor 07/03/01 09:56:21 06/03/01 12:10:43 A COL 1.000000 RICARDO Raw Data File : \\klebb\TCData\DATA3\06MAR01004.raw Sequence File: \\klebb\TCData\DATA3\Group.seq 16 18 20 Time [min] Particulate Analysis Time Component [min] Name Area [MVS] 4.755 GROUP 1 18.335 TOTAL 21.585 GROUP 2 7697135.67 17562464.50 9865328.83 35124929.00 115 Figure E17. Chromatographic Spectrum of Particulate V O F on Filter E M - 298 Kage 1 of 1 Software Version Sample Name Instrument Name RackA/ial Sample Amount Cycle 6.1.1.0.0:K20 WESTPORT EM-298 BACKUP GC3 0/0 1.000000 7 Raw Data File: \\klebb\TCData\DATA3\07MAR01007.raw Sequence File: \\klebb\TCData\DATA3\Group.seq Date Data Acquisition Time Channel Operator Dilution Factor 07/03/01 15:12:26 07/03/01 13:44:50 A COL 1.000000 RICARDO f 150 1 too 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 Time [min] Particulate Analysis Time Component Area [min] Name [uVs] 4.755 GROUP 1 4550348.56 18.335 TOTAL 5988730.25 21.585 GROUP 2 1438381.69 11977460.50 116 Figure E18. Chromatographic Spectrum of Particulate V O F on Filter E M - 261 ^ge 1 of 1 Software Version Sample Name Instrument Name : GC3 6.1.1.0.0:K20 Eh~7.Lt -^i*J!i~"~~i- L'^ n<-RackA/ial Sample Amount Cycle 0/0 1.000000 9 Raw Data File : \\klebb\TCData\DATA3\07MAR01009.raw Sequence File: \\klebb\TCData\DATA3\Group.seq Date Data Acquisition Time Channel Operator Dilution Factor 07/03/01 17:14:51 07/03/01 16:22:34 A IHHaH»M COL 1.000000 ¥ 1 2 0 - ^ 8 0 - ^ JiTfTjTT 2 Time Component [min] Name lljl!ll|ll!lj!ll!|!lll|!!l!jlll![!lll|llil|i!!!jll!!|ll!!|llli]IIH 10 12 14 16 18 20 22 24 26 28 30 32 34 Area [uVs] Time [min] Particulate Analysis 4.755 GROUP 1 18.335 TOTAL 21.585 GROUP 2 4061910.79 6783457.57 2709272.98 13554641.34 117 Figure E19. Chromatographic Spectrum of Westport Diesel Fuel Sample Software Version : 6.1.1.0.0:K20 Sample Name : WESTPORT FUEL SUPPLIED Instrument Name : GC3 Rack/Vial : n/0 Sample Amount : 1.000000 Cyde ; 4 Result File : \\klebb\TCData\DATA3\10JAN01004 rst Sequence File: \\klebb\TCData\DATA3\Group.seq Page 1 of 1 J? a , t e . • 02/05/01 12:36:01 Data Acquisition Time : 10/01/01 11-52:01 Channel : A Operator [ C O L Dilution Factor : 1.000000 ——ma I T Time Component Area [min} Name [pV-s] 6.880 GROUP 1 36070102 28 18.335 TOTAL 37852013 97 23.705 GROUP 2 178191169 TimeNnj 2 ° . 2 2 2 4 2 6 29 30 32 34 : Particulate Analysis 75704027.94 118 Figure E20. Chromatographic Spectrum of Westport Lubricating Oil Sample Software Version : 6.1.1.0 0K20 SKXn. •= ^ 0 R T ° » - S U P P L I E D Rack/Via! : o/ 0 Sample Amount : 1.000000 Cycle . 6 Result R'e : \\Webb\TCData\DATA3\10JAN0lOOB r, Date Data Acquisition Time Channel Operator Dilution Factor 02/05/01 1236:27 10/01/01 13:40:56 A COL 1.000000 Page 1 o EE3333 Time Component Area [""OJ^  Name [uVs] 1 4 1 6 IS 20 22 "mB [min] Particulate Analysis 24 26 2 8 30 32 ' j4 1 pf'vnf J £ I A L 35144020.74 23.705 GROUP 2 13641809.18 70288041.48 119 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0080964/manifest

Comment

Related Items