Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Development and application of in-cylinder fuel concentration and pyrometry optical diagnostic tools… Yeo, Jeff 2017

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

Item Metadata

Download

Media
24-ubc_2017_september_yeo_jeff.pdf [ 7.02MB ]
Metadata
JSON: 24-1.0354490.json
JSON-LD: 24-1.0354490-ld.json
RDF/XML (Pretty): 24-1.0354490-rdf.xml
RDF/JSON: 24-1.0354490-rdf.json
Turtle: 24-1.0354490-turtle.txt
N-Triples: 24-1.0354490-rdf-ntriples.txt
Original Record: 24-1.0354490-source.json
Full Text
24-1.0354490-fulltext.txt
Citation
24-1.0354490.ris

Full Text

DEVELOPMENT AND APPLICATION OF IN-CYLINDER FUEL CONCENTRATION AND PYROMETRY OPTICAL DIAGNOSTIC TOOLS IN DIESEL-IGNITED DUAL-FUEL NATURAL GAS ENGINES by  Jeff Yeo  B.A.Sc., The University of British Columbia, 2013  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Mechanical Engineering)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  August 2017  © Jeff Yeo, 2017 ii  Abstract In recent years, government policies have mandated significant reductions in emissions of greenhouse gases (GHG) and particulate matter (PM) from heavy-duty, on-highway transportation applications.  This has necessitated the development of clean engine technologies such as dual-fuel combustion of natural gas (NG) in compression ignition (CI) engines.  With these new engine developments, the need to understand and optimize these technologies to meet emission regulations becomes crucial.  Traditional engine research relies on thermodynamic methods and exhaust analysis to examine performance and emission trends present across real-world operating conditions – more recently, optical tools have become increasingly accessible and are providing new methods of understanding combustion phenomena.  Interpretation however, is often not directly translatable between optical and thermodynamic engines due to the many mechanical differences between the two.  This work aims to provide the foundations for a new “thermo-optical” approach which combines conventional thermodynamic analysis with optical insight into combustion phenomenon to bridge the gap between thermodynamic and optical engine studies.    The development and application of an optical probe performing line of sight pyrometry for in-cylinder soot concentration and temperature measurements, as well as the implementation of a probe for in-cylinder local fuel concentration measurements is detailed.  The probes are operated in a 2-litre single-cylinder research engine capable of thermodynamic (“all-metal”) and optical configurations and were utilized under two vastly different operating strategies.  These strategies used premixed methane with a diesel pilot (DIDF), and High Pressure Direct Injection (HPDI) for non-premixed NG with a diesel pilot.  The pyrometry probe demonstrated the effect of HPDI injection parameters on in-cylinder PM concentration while fuel concentration measurements were iii  used to identify the combustion mode under DIDF conditions, and to provide insight to HPDI injection and combustion characteristics.  The probes’ performance and capabilities were evaluated under thermodynamic and optical configurations, with high-speed cameras complementing the probes during optical operation.  The framework for interpretation in the “thermo-optical” methodology was developed through analysis of the local fuel concentration, soot concentration and temperature, spatially resolved optical results, and conventional apparent heat release rate (AHRR) analysis.    iv  Lay Abstract The increasingly tough emission regulations for heavy-duty diesel trucks have driven the development of new engine technologies towards reducing greenhouse gas and soot emissions.  Due to the relative infancy of these new technologies, engine research geared towards understanding pollutant formation phenomena is critical for optimization and this work aims to showcase new developments in this area.  Two optical engine probes were developed and commissioned in this work with the purpose of measuring the in-cylinder fuel concentration and soot concentration in a diesel-natural gas dual-fuel engine to expose any links that exist between the probe measurements and engine performance and emission trends.  In addition to this, the probes are simultaneously operated alongside high-speed cameras that record each engine combustion event, and further conclusions regarding the interpretation of the probes are made based on comparisons with high-speed imaging.   v  Preface The original vision of this research work with regards to the thermo-optical methodology was developed by my supervisor Dr. Patrick Kirchen.  However, the direction I took with regards to the design of the pyrometry probe was largely my own undertaking along with the implementation of the probes to the engine facility.  Experimental design during probe commissioning activities was based on past works by colleagues Jeremy Rochussen, Mahdiar Khosravi, and Ehsan Faghani, with input and guidance from Dr. Patrick Kirchen.  Custom fabrication work was performed by Roland Genshorek at UBC Mechanical Engineering Machine Shop.  Simultaneous to the development of these probes, engine commissioning activities also took up a large portion of my research effort.  Jeremy Rochussen designed and commissioned the main mechanical and safety systems for the Ricardo Proteus single-cylinder research engine.  I implemented the engine control and data acquisition hardware including wiring test-cell sensors and actuators, and developed the associated software for engine control and data acquisition.  Mahdiar Khosravi was responsible for commissioning the high-speed optical imaging systems.    Both thermodynamic and optical engine tests were a combined effort between Mahdiar Khosravi, Jeremy Rochussen, and myself.  I was responsible for processing thermodynamic and probe data, while Mahdiar Khosravi completed post-processing for the natural luminosity and OH* chemiluminescence imaging.  Jeremy Rochussen wrote the original scripts to process the thermodynamic data, and I modified these to include routines specific to the probes.    vi  The ICOS results shown in Chapter 4 of this thesis has been presented and published as part of the 2016 Combustion Institute Canadian Section Conference and 2016 SAE Powertrain, Fuels, and Lubricants meeting.  The pyrometry probe results shown in Chapter 5 have also been presented and published as part of the 2017 Combustion Institute Canadian Section Conference. • J. Yeo, J. Rochussen, M. Khosravi, and P. Kirchen, “Application of an In-Cylinder Local Infrared Absorption Fuel Concentration Sensor in a Diesel-Ignited Dual-Fuel Engine,” in Proceedings of the Combustion Institute - Canadian Section, 2016. • J. Yeo, J. Rochussen, and P. Kirchen, “Application of an In-Cylinder Local Infrared Absorption Fuel Concentration Sensor in a Diesel-Ignited Dual-Fuel Engine,” SAE Tech. Pap., no. 2016-01–2310, 2016. • J. Yeo, M. Khosravi, J. Rochussen, and P. Kirchen, “Application of an In-Cylinder Line of Sight Two-Colour Pyrometry Probe in an Optical Pilot-Ignited Direct Injection Natural Gas Engine,” in Proceedings of the Combustion Institute - Canadian Section, 2017. I was the lead author for these works with input and assistance from co-authors Jeremy Rochussen, Mahdiar Khosravi, and Dr. Patrick Kirchen.    In addition to the works listed above, I co-authored the following papers: • J. Rochussen, J. Yeo, and P. Kirchen, “Effect of Fueling Control Parameters on Combustion and Emissions Characteristics of Diesel-Ignited Methane Dual-Fuel Combustion,” SAE Tech. Pap., no. 2016-01–0792, 2016. o I was responsible for developing the engine control and data acquisition for the engine facility and subsequently wrote a section of this paper describing these aspects. vii  • M. Khosravi, J. Rochussen, J. Yeo, P. Kirchen, G. McTaggart-Cowan, and N. Wu, “Effect of Fuelling Control Parameters on Combustion Characteristics of Diesel-Ignited Natural Gas Dual-Fuel Combustion in an Optical Engine,” in ASME 2016 Internal Combustion Engine Division Fall Technical Conference, 2016, pp. 1–10. o I was responsible for continued development of engine control aspects with regards to optical camera triggering and skip firing required for optical tests. o I also assisted with rebuilding the optical engine and running optical tests with Jeremy Rochussen and Mahdiar Khosravi.   • J. Rochussen, J. Son, J. Yeo, M. Khosravi, P. Kirchen, and G. McTaggart-Cowan, “Development of a Research-Oriented Cylinder Head with Modular Injector Mounting and Access for Multiple In-Cylinder Diagnostics,” SAE Tech. Pap., no. 2017-24–0044, 2017. o I provided input with regards to the design and geometry of the optical probe bores.   o I designed and built the pressure testing plate used in structural validation testing and supported those tests. o The sample combustion diagnostic results present some probe data from my engine tests.  These results can be found in Chapter 4 and Chapter 5.   viii  Table of Contents Abstract .......................................................................................................................................... ii Lay Abstract ................................................................................................................................. iv Preface .............................................................................................................................................v Table of Contents ....................................................................................................................... viii List of Tables .............................................................................................................................. xiii List of Figures ............................................................................................................................. xiv List of Symbols .............................................................................................................................xx List of Abbreviations ................................................................................................................ xxii Acknowledgements .................................................................................................................. xxiv Dedication ................................................................................................................................. xxvi Chapter 1: Introduction ............................................................................................................... 1 1.1 Motivation ................................................................................................................... 1 1.2 Alternative Clean Engine Technologies ..................................................................... 2 1.3 Dual-Fuel Concept with Natural Gas and Diesel ........................................................ 3 1.3.1 Diesel-Ignited Dual-Fuel .................................................................................... 4 1.3.2 High Pressure Direct Injection ............................................................................ 4 1.4 Comparison of Thermodynamic and Optical Studies ................................................. 5 1.5 Optical Probes ............................................................................................................. 7 1.6 Objectives ................................................................................................................... 7 1.7 Approach ..................................................................................................................... 8 Chapter 2: Background Theory & Review of the State of the Art ........................................... 9 2.1 Dual-Fuel Combustion Using Natural Gas and Diesel ............................................... 9 ix  2.1.1 DIDF Background ............................................................................................. 10 2.1.1.1 Stage 1 Combustion ...................................................................................... 14 2.1.1.2 Stage 2 Combustion ...................................................................................... 14 2.1.1.3 Stage 3 Combustion ...................................................................................... 15 2.1.1.4 DIDF Operating Space and Conceptual Model Summary ............................ 16 2.1.2 HPDI Background ............................................................................................. 20 2.1.3 Optical DIDF Studies ........................................................................................ 21 2.1.4 Utility and Limitations of Optical Diagnostic Tools in Engine Research ........ 29 2.2 In-Cylinder Optical Diagnostic Tools ....................................................................... 30 2.2.1 In-Cylinder Fuel Concentration Measurements ................................................ 31 2.2.2 In-Cylinder Soot Temperature and Soot Concentration Measurements ........... 35 2.2.2.1 Two-Colour Pyrometry Technique ............................................................... 37 2.2.2.2 Heat Transfer Background Theory ............................................................... 37 2.2.2.3 Hottel Broughton Correlation ....................................................................... 39 2.2.2.4 Parameter Selection in the Two-Colour Method .......................................... 39 2.2.2.5 Validation and Application Studies .............................................................. 40 2.2.2.6 Two-Colour Pyrometry Summary ................................................................ 44 2.3 Links in the Thermo-Optical Methodology .............................................................. 45 2.3.1 Thermo-Optical Application of Fuel Concentration Measurements ................ 45 2.3.2 Thermo-Optical Application of Pyrometry ....................................................... 45 2.4 Summary and Literature Gap .................................................................................... 47 Chapter 3: Experimental Facility .............................................................................................. 49 3.1 Ricardo Proteus Single-Cylinder Engine .................................................................. 49 x  3.2 Instrumentation and Data Acquisition (DAQ) .......................................................... 52 3.2.1 Test-Cell and Engine Instrumentation .............................................................. 55 3.2.2 Data Acquisition (DAQ) ................................................................................... 56 3.2.2.1 Combustion Analysis Software..................................................................... 57 3.2.2.2 Externally Logged Signals ............................................................................ 57 3.2.2.3 Data Acquisition Procedure .......................................................................... 58 3.3 Engine Control Unit (ECU) ...................................................................................... 59 3.3.1 FPGA Virtual Instrument .................................................................................. 60 3.3.2 Real-Time Virtual Instrument ........................................................................... 61 3.3.3 Operating PC Virtual Instrument ...................................................................... 61 3.3.3.1 Engine Controls ............................................................................................ 61 3.3.3.2 Safety Functionality ...................................................................................... 63 3.4 Testing Procedure ..................................................................................................... 63 3.5 Experimental Facility Summary ............................................................................... 65 Chapter 4: LaVision Internal Combustion Optical Sensor (ICOS) ....................................... 68 4.1 Internal Combustion Optical Sensor Hardware Overview ....................................... 68 4.2 Operating Theory of the Internal Combustion Optical Sensor ................................. 72 4.3 ICOS DIDF Experimental Results and Discussion .................................................. 75 4.3.1 ICOS DIDF Experiment Design ....................................................................... 75 4.3.2 Effect of Diesel Fuel on Fuel Concentration Measurements (DIDF) ............... 78 4.3.3 Local Air Fuel Ratio Comparison with Mass Based Air Fuel Ratio ................ 80 4.3.4 Uncertainty in Local AFR Measurements ........................................................ 83 4.3.4.1 Volume Errors ............................................................................................... 83 xi  4.3.4.2 I0 Error .......................................................................................................... 85 4.3.4.3 Air Mass Errors ............................................................................................. 87 4.3.5 Evaluation of Combustion Mode Based on Fuel Concentration History ......... 91 4.3.5.1 Methane Comsumption Delay Relative to 5% HRR .................................... 97 4.3.5.2 Fuel Concentration Ratio Comparison with Pressure Ratio ......................... 99 4.3.6 Effect of Methane Port Injection Timing on Local Fuel Concentration ......... 101 4.3.7 ICOS DIDF Testing Summary........................................................................ 104 4.4 ICOS HPDI Experimental Results and Discussion ................................................ 105 4.4.1 ICOS HPDI Experiment Design ..................................................................... 105 4.4.2 Effect of Diesel Fuel on Fuel Concentration Measurements (HPDI) ............. 107 4.4.3 Evaluation of Fuel Concentration Histories .................................................... 108 4.4.4 ICOS HPDI Testing Summary........................................................................ 117 4.5 ICOS Summary ....................................................................................................... 118 Chapter 5: Pyrometry Probe ................................................................................................... 120 5.1 Pyrometry Probe Design ......................................................................................... 120 5.1.1 Flow Restriction Analysis ............................................................................... 127 5.1.2 Pyrometry Probe Optical Components ........................................................... 132 5.2 Pyrometry Probe Calibration .................................................................................. 136 5.3 Pyrometry Probe HPDI Experimental Results and Discussion .............................. 141 5.4 Pyrometry Probe Summary ..................................................................................... 149 Chapter 6: Conclusions and Recommendations .................................................................... 151 6.1 ICOS ....................................................................................................................... 152 6.2 Pyrometry Probe ..................................................................................................... 154 xii  6.3 Recommendations ................................................................................................... 156 6.3.1 Future ICOS Work .......................................................................................... 156 6.3.2 Future Pyrometry Probe Work ........................................................................ 158 References ...................................................................................................................................160 Appendices ..................................................................................................................................168   xiii  List of Tables Table 1 - EPA and Euro VI emission limits for heavy-duty vehicles [3], [4]. ............................... 2 Table 2 - Comparison between an optical and thermodynamic engine. ......................................... 6 Table 3 - Ricardo Proteus single-cylinder research engine specifications. .................................. 51 Table 4 - Engine Sensors used in combustion analysis. ............................................................... 55 Table 5 - Data acquisition chassis module descriptions. .............................................................. 56 Table 6 - Powertrain control modules descriptions. ..................................................................... 59 Table 7 - Detailed experimental conditions in the ICOS DIDF measurement campaign. ............ 77 Table 8 - Selected HPDI operating points. ................................................................................. 106 Table 9 - Cylinder temperatures at combustion top dead center. ............................................... 113 Table 10 - Selected HPDI operating points. ............................................................................... 141  xiv  List of Figures Figure 1 - Conceptual schematics of DIDF (left) and HPDI (right) injection strategies. ............... 3 Figure 2 - HPDI Injector (www.westport.com) .............................................................................. 5 Figure 3 - Karim’s conceptual stages of dual-fuel combustion under heavy- and light-load conditions. ..................................................................................................................................... 11 Figure 4 - Sahoo’s conceptual stages of dual-fuel combustion. ................................................... 13 Figure 5 - Rochussen’s mathematical descriptions of the different stages of DIDF combustion. 17 Figure 6 - Description of flame combustion modes under the experimental conditions described in [16]. ........................................................................................................................................... 19 Figure 7 - OH* chemiluminescence imaging of low- and high-load DIDF operation (left and right respectively). ........................................................................................................................ 23 Figure 8 - OH* chemiluminescence images of DIDF combustion in a RCEM............................ 26 Figure 9 - Simultaneous NL and OH* chemiluminescence image sequences for two different pilot injection pressures at the same global and methane equivalence ratios. .............................. 27 Figure 10 - Conceptual images of reaction zone growth at high and low diesel pilot injection pressures based on OH* chemiluminescence imaging for low pilot ratio DIDF operation. ........ 29 Figure 11 - 2D air/fuel ratio (fuel concentration) profile in an SI engine using Nd:YAG laser Rayleigh scattering........................................................................................................................ 32 Figure 12 - Optical setup used in dual-wavelength PLIF measurements of temperature and fuel concentration. ................................................................................................................................ 33 Figure 13 - Sample fuel concentration measurements (right) from PLIF tracer experiments. ..... 34 Figure 14 - Two-colour pyrometry probe cross section. .............................................................. 42 Figure 15 - Schematic of the optical connections in a pyrometry application. ............................. 43 xv  Figure 16 - Schematic layout of Ricardo Proteus experimental facility in thermodynamic and optical configurations.................................................................................................................... 50 Figure 17 - HPDI head (left) and DIDF head (right). ................................................................... 51 Figure 18 - Experimental facility process and instrumentation diagram for DIDF thermodynamic engine configuration. .................................................................................................................... 53 Figure 19 - Experimental facility process and instrumentation diagram for HPDI thermodynamic engine configuration. .................................................................................................................... 54 Figure 20 - Simplified DAQ and ECU hardware and software hierarchy. ................................... 60 Figure 21 - Standard optical test firing scheme. ........................................................................... 62 Figure 22 - Proteus research engine in HPDI optical configuration. ............................................ 65 Figure 23 - HPDI head with in-cylinder diagnostic probes installed: in-cylinder pressure transducer, ICOS probe, and pyrometry probe. ............................................................................ 66 Figure 24 - Schematic of probe (ICOS and pyrometry) installations and their position relative to the gas jets for HPDI injection. ..................................................................................................... 67 Figure 25 - Simplified cross section view of the absorption path of the LaVision ICOS probe. . 69 Figure 26 - LaVision ICOS probe without its mounting clamp.................................................... 69 Figure 27 - DIDF configuration engine front and side view. ....................................................... 70 Figure 28 - Plan view from piston showing diesel spray pattern relative to the ICOS probe in DIDF configuration. ...................................................................................................................... 71 Figure 29 - DIDF head with in-cylinder diagnostic probes installed: in-cylinder pressure transducer, ICOS probe, and pyrometry probe. ............................................................................ 71 Figure 30 - Plan view from piston showing diesel and NG spray pattern relative to the ICOS probe in HPDI configuration. ....................................................................................................... 72 xvi  Figure 31 - Sample transmitted light intensity to indicate the influence of cycle stages (intake, compression, combustion, exhaust) on the measured intensity for DIDF configuration. ............. 73 Figure 32 - Considered operating points on global equivalence ratio – pilot ratio space. ............ 78 Figure 33 - Local fuel concentration measurements for diesel-only testing. ................................ 79 Figure 34 - Unadjusted local AFR measurement trends during compression .............................. 81 Figure 35 - Comparison of local and global AFR based on measured fuel and air mass flow rates for all measured points. ................................................................................................................. 82 Figure 36 - Compression ratio adjusted (decrease from 14.25:1 to 13.25:1) local AFR measurement trends during compression prior to combustion. .................................................... 85 Figure 37 - I0 adjusted (1% increase) local AFR measurement trends during compression prior to combustion. ................................................................................................................................... 87 Figure 38 - mair adjusted (10% linear decrease from -20° to 0° aTDC) local AFR measurement trends during compression prior to combustion............................................................................ 89 Figure 39 - Sample in-cylinder methane fuel mass mfuel calculated using eq. (14). ..................... 90 Figure 40 - Local fuel concentration with AHRR.  Points 1-5, 1300 bar. .................................... 92 Figure 41 - Local fuel concentration with AHRR.  Points 1-5, 300 bar. ...................................... 93 Figure 42 - Local fuel concentration with AHRR.  Points 6-10, 1300 bar. .................................. 93 Figure 43 - Local fuel concentration with AHRR.  Points 6-10, 300 bar. .................................... 94 Figure 44 - AHRR and normalized fuel concentration.  Point 3, 300 and 1300 bar..................... 95 Figure 45 - AHRR and normalized fuel concentration.  Point 8, 300 and 1300 bar..................... 96 Figure 46 - Methane consumption delay vs. global AFR for all considered operating points. .... 98 Figure 47 - Comparison of fuel concentration ratio and pressure ratio calculated at peak fuel concentration crank angle and -5° aTDC for 1300 bar test cases. .............................................. 100 xvii  Figure 48 - Local fuel concentration with AHRR histories for varying commanded start of port methane fuel injection. ................................................................................................................ 101 Figure 49 - Local fuel concentration with AHRR histories for varying commanded start of port methane fuel injection. ................................................................................................................ 102 Figure 50 - Port injection timing relative to intake valve lift for CSOI variation. ..................... 102 Figure 51 - Comparison of local and global AFRs for methane CSOI timing sweep. ............... 103 Figure 52 - HPDI injection parameters explained. ..................................................................... 106 Figure 53 - HPDI diesel-only heat-release rate and fuel-concentration for thermodynamic and optical engine configurations. ..................................................................................................... 107 Figure 54 - HPDI main diesel spray pattern for the diesel-only test relative to ICOS and pyrometry probes. ....................................................................................................................... 108 Figure 55 - Heat release rate and fuel concentration (fired and non-fired cases) for HPDI fueling in optical engine configuration. .................................................................................................. 109 Figure 56 - Heat release rate and fuel concentration (fired and non-fired cases) for HPDI fueling in thermodynamic engine configuration. .................................................................................... 110 Figure 57 - Short PSEP HPDI fuel concentration measurements ............................................... 112 Figure 58 - High DRP HPDI fuel concentration measurements ................................................. 112 Figure 59 - Short PSEP illuminated non-fired freeze-frames. .................................................... 115 Figure 60 - Short PSEP NL fired freeze-frames. ........................................................................ 116 Figure 61 - DIDF head pre- and post-modifications for pyrometry probe. ................................ 121 Figure 62 - As designed probe bore dimensions on the original diesel head used for DIDF studies. ........................................................................................................................................ 122 Figure 63 - Pyrometry probe cross sections. ............................................................................... 123 xviii  Figure 64 - Sapphire and invar holder brazed assembly. ............................................................ 124 Figure 65 - As designed probe bore dimensions on the custom head used for HPDI studies. ... 125 Figure 66 - Fully assembled pyrometry probe and photodetector assemblies. ........................... 127 Figure 67 - Sketch of the probe restriction showing the flow of hot gases across the front face of the sapphire rod. .......................................................................................................................... 128 Figure 68 - Modelled gas pressure drop between cylinder and probe cavity volume for varying Cd................................................................................................................................................. 130 Figure 69 - Modelled gas velocity through restriction area between cylinder and probe cavity volume for varying Cd. ................................................................................................................ 131 Figure 70 - Pyrometry probe connection diagram ...................................................................... 132 Figure 71 - Pyrometry probe configuration as used in HPDI experiments................................. 133 Figure 72 - Projected maximum FOV under experimental setup geometry (50° cone angle). .. 135 Figure 73 - Detailed pyrometry probe photodetector setup ........................................................ 136 Figure 74 - Pyrometry probe benchtop calibration setup with Labsphere. ................................. 137 Figure 75 - Sample 700nm calibration after a set of optical tests. .............................................. 138 Figure 76 - Comparison calibration before and after 3 hours of continuous thermodynamic operation. .................................................................................................................................... 139 Figure 77 - Diesel-only HPDI test. ............................................................................................. 140 Figure 78 - Optical heat release rate, soot temperature, and KL for Baseline, High DRP, and Short PSEP test points. ............................................................................................................... 142 Figure 79 - Optical heat release rate, soot temperature, and KL for Long GPW, LPI, and Diesel test points. ................................................................................................................................... 143 xix  Figure 80 - Comparison between high-speed natural luminosity imaging and KL factor measured by the probe for diesel-only operation. ....................................................................................... 145 Figure 81 - Comparison between high-speed natural luminosity imaging and KL factor measured by the probe for HPDI late post injection (LPI) operation. ........................................................ 146 Figure 82 - Thermodynamic heat release rate, soot temperature, and KL for Baseline, High DRP, and Short PSEP test points. ......................................................................................................... 147 Figure 83 - Thermodynamic heat release rate, soot temperature, and KL for Long GPW, LPI, and Diesel test points. ........................................................................................................................ 148  xx  List of Symbols A Area A/FCH4,stoich. Stoichiometric air-fuel ratio for methane A/Fdiesel,stoich. Stoichiometric air/fuel ratio for diesel A/Fglobal Quantitative measurement of global air-fuel ratio based on mass flow rate measurements of fuel and air A/Flocal Qualitative measurement of local air-fuel ratio based on infrared absorption probe measurements C1 Blackbody radiation constant C2 Blackbody radiation constant; Diatomic carbon C2/C1 Fuel concentration ratio between two different points of a cycle CA50 Crank angle corresponding to 50% heat release CA95 Crank angle corresponding to 95% heat release CAdelay Crank angle duration corresponding to the delay in methane fuel conversion between 5% heat release to 50% of the initial methane fuel concentration CAHRR,0.05 Crank angle corresponding to 5% heat release CAρ,0.5 Crank angle corresponding to 50% fuel concentration of the reference initial value Cd Discharge coefficient in valve mass flow rate equation Cfuel Local fuel concentration CH Carbyne CH4 Methane CO Carbon monoxide CO2 Carbon dioxide H2O Water I Measured ICOS intensity I0 Reference ICOS intensity iλ Spectral radiation intensity iλ,b Blackbody spectral radiation intensity K Absorption coefficient per unit length L Optical path length LHVCH4 Lower heating value of methane LHVdiesel Lower heating value of diesel ṁ(t) Mass flow rate through a valve restriction as a function of time mair In-cylinder mass of air per cycle ṁair,meas. Measured air mass flow rate ṁCH4,meas. Measured CH4 mass flow rate ṁdiesel,meas. Measured diesel mass flow rate mfuel In-cylinder mass of fuel expressed as a function of crank angle n Index of refraction xxi  NOx Nitrogen oxides O2 Oxygen OH Hydroxyl radical OH* Electronically-excited hydroxyl radical P2/P1 In-cylinder pressure ratio between two different points of a cycle pcr Critical pressure pcyl Cylinder pressure pin Inlet pressure in valve restriction mass flow calculation pout Outlet pressure in valve restriction mass flow calculation Rpilot Ratio of pilot fuel to total fuel on an energy basis T Temperature Ta Apparent temperature Tc Cylinder temperature Tmean-gas Mean gas temperature V Cylinder volume expressed as a function of crank angle α Pyrometry empirical parameter ελ Spectral emissivity θ Angle of incidence κ Heat capacity ratio λ Wavelength σ Absorption strength ϕ  Equivalence ratio  ϕCH4  Equivalence ratio of methane  ϕdiesel Equivalence ratio of diesel ϕglobal Equivalence ratio of diesel and methane combined xxii  List of Abbreviations AFR Air-fuel ratio AHRR Apparent heat release rate aSOI After start of injection aTDC After top dead center BDC Bottom dead center bTDC Before top dead center CA Crank angle CAD Crank angle degree CARS Coherent Anti-Stokes Raman Spectroscopy CAS Combustion Analysis Software cDAQ Compact Data Acquistion CI Compression ignition CN Cetane number CR Compression ratio cRIO Compact Realtime Input Output CRP Colour-ratio pyrometry CSOI Commanded start of injection CWL Center wavelength DAQ Data acquisition DIDF Diesel-ignited dual-fuel ECU Engine control unit EGR Exhaust gas recirculation EOI End of injection EPA Environmental Protection Agency EPT Engine position tracker EU European Union EVO Exhaust valve open FOV Field of view FPGA Field-programmable gate array GHG Greenhouse gas HC Hydrocarbons HPDI High Pressure Direct Injection ICE Internal combustion engine ICOS Internal Combustion Optical Sensor IFPEN IFP Energies nouvelles IR Infrared  IVC Intake valve close LHV Lower heating value LII Laser-induced incandescence  LPI Late post injection LTC Low temperature combustion NG Natural gas NL Natural luminosity xxiii  NMHC Non-methane hydrocarbons OEM Original equipment manufacturer ON Octane number PFI Port fuel injector/injection PLIF Planar laser-induced fluorescence PM Particulate matter PN Particle number RCEM Rapid compression / expansion machine SI Spark-ignited SPC Slightly premixed combustion Tcf Trillion cubic feet TDC Top dead center THC Total hydrocarbons uHC Unburned hydrocarbons VI Virtual instruments xxiv  Acknowledgements This work would not have been possible without the very fruitful collaboration and financial support from Westport Fuel Systems, in particular with Drs. Gordon McTaggart-Cowan, Sandeep Munshi, and Ning Wu.  In addition, I would like to acknowledge further financial support provided by the NSERC CREATE Program in Clean Combustion Engines, NSERC Discovery Grant Program, and the Canadian Foundation for Innovation (CFI).  At UBC, Dr. Steven Rogak has been a great source of valuable insight, and I thank him for the direction he provided throughout both my undergraduate and graduate degrees.  I must also mention Roland Genschorek for his dependability – a great deal of this work required his machining skills, and he never failed to deliver.  To my supervisor, Dr. Patrick Kirchen, thank you for taking a chance on a student who slept through your undergrad class.  You presented me with the opportunity of a lifetime, and I could not have been more fortunate to work with such a patient and helpful supervisor.  I will forever be grateful for the guidance you’ve given me as I would not have made it this far without it.  Jeremy and Mahdiar – cheers to the many good memories we’ve had in the lab.  I don’t think it was possible for me to stay sane rebuilding the same engine as many times as we did without the camaraderie we have.  I am so glad to have been a part of the experience that is Team Proteus, and I am endlessly appreciative for all you’ve done for me.  Of course, to the other folks in the group and especially Jan, it has been an absolute treat.  I promise I’ll set up another karting session before the end.    xxv  A thesis can sometimes be a source of frustration, and for those rough days, I would like to recognize the steadfast foundation my friends provide me.  Whether it be to lend an ear, or for a quick laugh, I can always rely on them for a reprieve from the hectic life of grad school.  I must also show my gratitude to Grant for the carpools, coffee breaks, and the interesting and sometimes bizarre perspective you’ve given me based on your own experience with grad school.  To Sisi, you have my utmost appreciation – not only for the occasional kick-start when motivation is low, but also for your patience and understanding when I find myself entangled with deadlines.  And finally, to my parents, thank you for the unconditional love and support throughout the years and for allowing me to pursue the things I want without question.  Unquestionably, it would not have been possible without you. xxvi  Dedication          dedicated to Jooi1  Chapter 1: Introduction This thesis demonstrates the development, commissioning, and implementation of two optical probes for use in a compression ignition optical research engine at the University of British Columbia Clean Energy Research Centre.  This chapter provides the reader with the motivation, necessary background knowledge, and objectives of the work presented herein.    1.1 Motivation Over the past several decades, increasing global awareness of climate change has driven policymaking and the economic landscape of the world.  Climate change has been greatly attributed to the sharp increase of greenhouse gas (GHG) emissions since the start of the Industrial Revolution and in the past couple decades alone, GHG emissions have increased by 35% from 1990 to 2010 of which carbon dioxides accounts for three-fourths of total emissions [1].  Of the total GHG emissions, the transportation sector has long been a large contributor – in the United States, transportation contributes to about 26% the country’s GHG emissions since 1990 [1] .  This global phenomenon has become a pressing concern and many government policies and regulations that aim to reduce GHG output in the transportation sector are now in place [2]–[4].  The transportation sector can be separated into distinct categories such as air, rail, marine, and automotive.  This work focuses primarily on heavy-duty highway applications, which is best described by long-haul trucks used in cargo transportation.  The aforementioned policies and regulations, set out by major governing bodies including Transport Canada, Environmental Protection Agency (EPA) in the US, and the European Union (EU), enforce emission limits on heavy-duty compression ignition (CI) engines on species including CO, total hydrocarbons (THC), 2  non-methane hydrocarbons (NMHC), NOx, particulate matter (PM), CH4, and particle number (PN).  In Canada, on-road vehicle and engine emission regulations are harmonized with US EPA policies, and can be collectively considered under EPA standards.  The most recent and relevant limits are shown in Table 1. Table 1 - EPA and Euro VI emission limits for heavy-duty vehicles [3], [4].  Limit values Regulation & Governing Body CO (g/kWh) THC (g/kWh) NMHC (g/kWh) NOx (g/kWh) PM (g/kWh) CH4 (g/kWh) PN (1/kWh) Euro VI Steady State (EU) 1.5 0.13 - 0.4 0.01 - 8x1011 Euro VI Transient (EU) 4 - 0.16 0.46 0.01 0.5 6x1011 EPA Transient (US) 21 - 0.19 0.27 0.013 - -  Traditionally, heavy-duty CI engines are fueled by diesel.  However, increasing pressure from governing bodies to reduce GHG emissions has led original equipment manufacturers (OEMs) to pursue alternative fuels and technologies.  As a result, OEMs are now focusing more of their resources on developing future technologies that aim to meet EPA and Euro VI targets [5].  1.2 Alternative Clean Engine Technologies Some of the GHG cutting alternative technologies explored by various research groups and OEMs around the world include different types of combustion strategies, exhaust after-treatment, alternative fuels, and more recently, a complete shift away from internal combustion engines (ICE) and towards the hybridization and electrification of medium- and heavy-duty vehicles [6].  Government incentives and research programs also play a large role in driving innovation in this area with the SuperTruck program being a notable effort in providing a full Class 8 tractor-trailer 3  vehicle demonstration of a wide range of fuel-saving technologies in an attempt to reach 55% brake thermal efficiency [6].  While some variants of these technologies have been successfully introduced to large markets (e.g., Tesla’s success with the electrification of passenger vehicles), other technologies are still in the infantile stages in their development, and have not been brought to commercialization.  Keeping in mind the emphasis of this work on commissioning two optical probes for use in a heavy-duty compression ignition optical research engine, the discussion on clean engine technologies is limited to dual-fuel combustion using natural gas and diesel.    1.3 Dual-Fuel Concept with Natural Gas and Diesel Dual-fuel combustion is used here as the application through which the two optical probes’ performance is evaluated.  The dual-fuel concept uses both natural gas and diesel fuels simultaneously in the engine and in this study, will be referred to as diesel-ignited dual-fuel (DIDF) or High Pressure Direct Injection (HPDI).  A conceptual schematic of both combustion types is shown in Figure 1.  The reader is reminded that this serves as necessary discussion leading towards the commissioning and application of two in-cylinder optical probes.     Figure 1 - Conceptual schematics of DIDF (left) and HPDI (right) injection strategies.  4  1.3.1 Diesel-Ignited Dual-Fuel DIDF operation utilizes a gaseous port fuel injector (PFI) in the intake manifold and a traditional common rail diesel fuel injection system.  The PFI is supplied with methane CH4 (used as a surrogate fuel for NG in this study) and is actuated near intake valve closing.  The intake charge is subsequently inducted into the cylinder, forming a homogeneous (premixed) charge.  Combustion phasing is controlled through the direct injection of pilot diesel fuel, which ignites through compression.  NG as a fuel has high resistance to auto-ignition and has a high ON, allowing it to be used in traditional CI engines at high compression ratios without knock.  However, with the resistance to auto-ignition comes the difficulty in control of combustion phasing.  Diesel is used here as a high CN fuel that readily ignites under compression ignition to allow for controllable combustion phasing.  Current research in DIDF will be discussed in greater detail in §2.1.1.  1.3.2 High Pressure Direct Injection  High Pressure Direct Injection (HPDI) is a technology developed by Westport Fuel Systems Inc.  Like DIDF, HPDI is a dual-fuel concept that directly injects diesel as a pilot ignition source for NG.  However, the NG is directly injected into the cylinder as opposed to the port fuel injection fuel delivery method in DIDF (non-premixed combustion vs. premixed combustion respectively).  In HPDI, a single direct injector with dual concentric needles is used for both diesel and NG injection (Figure 2).  A short diesel pilot injection precedes the main natural gas injection – the NG contributes a majority of the injected fuel energy.  The direct injection of both diesel and NG creates a stratified (non-premixed) mixture within the cylinder, like traditional CI diesel engines.  Current literature in HPDI combustion will be presented in §2.1.2.  5   Figure 2 - HPDI Injector (www.westport.com)  1.4 Comparison of Thermodynamic and Optical Studies An in-depth understanding of fundamental in-cylinder combustion processes is crucial to advances in engine design and optimization but such intricacies can be missed in traditional thermodynamic apparent heat release rate (AHRR) analysis.  In more recent times, advancements in computational tools combined with optical diagnostics and validation in optical engines has provided significant developments in understanding in-cylinder phenomena including mixing, combustion, and emission formation processes [7].  However, optical engines, while useful, are not necessarily representative of “all-metal” or thermodynamic engines due to their inherent mechanical limitations.  Piston bowl geometry, component materials, and piston ring seals are just some of the differences between an “all-metal” engine and its optical counterpart.  A recent study at IFP Energies nouvelles (IFPEN) investigated the major differences between the two configurations and the potential mitigation strategies that can be employed to better align thermodynamic and 6  optical results [7].  In particular, the study notes some of the most significant effects are due to wall heat transfer characteristics, skip-firing strategies (affecting combustion phasing), and dynamic loading effects in the optical configuration.  The table below lists the major changes between optical and thermodynamic engines and discusses the implications on in-cylinder conditions. Table 2 - Comparison between an optical and thermodynamic engine. Aspect Differences Implications on In-Cylinder Conditions Piston Optical: Extended piston w/ quartz window; cylindrical bowl shape  Thermodynamic:  Standard aluminum piston with a toroidal bowl shape • Different mixing characteristics • Different heat transfer properties (e.g., piston surface temperatures) • Compression ratio may not be matched due to bowl geometry • Dynamic loading effects in optical due to increased moving mass and extended piston (e.g., dynamic compression ratio changes) Piston rings Optical: Oil-less, Greased liner and 2 piston rings Thermodynamic: Piston oil jet and 2 piston rings and an additional oil control ring • Different blow-by characteristics  • Change in crevice volumes Injection control Optical: Skip firing Thermodynamic: Continuous injection • Skip firing protects the optical components against damage • Peak in-cylinder pressure is limited in optical • Lower combustion temperatures (lower coolant temperatures) • Recirculated gas composition is altered between the two configurations  Given the known differences between “all-metal” and optical engine configurations, the main goal of this work is to begin the development of a thermo-optical approach that addresses these issues by utilizing optical diagnostic tools (i.e., optical probes) in an “all-metal” engine environment.  7  1.5 Optical Probes This thesis presents the development and application of two different in-cylinder optical diagnostic tools for use in a thermodynamic and optical engine facility.  These probes can provide information about in-cylinder processes that are not normally available in regular thermodynamic analysis and can be operated in continuous high-load conditions.  As such, these optical probes are suited to the development of the thermo-optical approach which requires the probes to be operated in both thermodynamic and optical engine configurations.  The first of these probes is a point location fuel concentration sensor while the second is a line of sight two-colour pyrometer measuring soot concentration and temperature.  The fuel concentration sensor can provide information on fuel conversion mechanisms while the pyrometer can shed light on soot formation and oxidation trends.  The intent with these probes is twofold: to provide optical insight into DIDF and HPDI combustion, and to provide the context and motivation for future combined thermodynamic and optical investigations using these diagnostic tools.    1.6 Objectives 1. Development, commissioning, and application of the in-cylinder fuel concentration sensor and pyrometry probe; development of supporting analysis tools 2. Provide perspective on various combustion strategies in terms of local fuel concentration, soot temperature, and soot concentration 3. Comparison between thermodynamic and optical measurements of local fuel concentration, soot temperature, and soot concentration along with conventional thermodynamic analysis (i.e., heat release rates) 8  4. Comparison of observed combustion trends with existing conceptual understanding of various combustion modes (DIDF/HPDI specific in this investigation) 5. Develop recommendations for interpretation between high-speed spatially resolved optical measurements (natural luminosity and OH* chemiluminescence), probe measurements, and conventional thermodynamic analysis in both optical and thermodynamic engine configurations  1.7 Approach The method used to achieve the objectives listed in §1.6 is detailed throughout the main body of this thesis.  A literature review of current experimental and optical DIDF/HPDI studies was performed to establish current knowledge of the DIDF/HPDI operating space.  Optical works presented insight to DIDF/HPDI experimental methodology and observed trends, which provided the context necessary in developing the thermo-optical methodology.  Following this, an assessment of existing in-cylinder diagnostic tools was carried out with a focus on in-cylinder fuel concentration, soot temperature, and soot concentration measurement probes.  With regards to the two-colour pyrometry probe, this step also identified the basis for the in-house self-cleaning design.  Once the probes were implemented to the facility, commissioning activities took place, and calibration procedures and analysis tools developed.  This eventually led to experimentation under suitable DIDF/HPDI test matrices as a first step towards the thermo-optical methodology.  9  Chapter 2: Background Theory & Review of the State of the Art The review here is divided into sections documenting experimental and optical dual-fuel studies (DIDF and HPDI), followed by an overview of in-cylinder diagnostic techniques.  The goal here is to recognize current works in the field and acknowledge the state of the art.  This provides not only the basis for experimental design, but also the context necessary for the discussion of new results.  An overview of in-cylinder diagnostic techniques looks at other measurement methodologies for measuring fuel concentration, soot temperature, and soot concentration, and identifies similar research using the same types of probes.    The first two sections cover the state of the art in dual-fuel research and discusses the general understanding of DIDF/HPDI combustion phenomena.  The use of optical diagnostics in dual-fuel is also discussed, though the focus is more on the utility and limits of optical tools in engine research.  The latter section focuses on in-cylinder diagnostic techniques with special attention to in-cylinder fuel concentration, soot temperature, and soot concentration measurements as these will be the core of the diagnostics tools used in the thermo-optical methodology.    2.1 Dual-Fuel Combustion Using Natural Gas and Diesel Dual-fuel engine technology is an active area of research and development as OEMs are attempting to achieve the NOx and PM limits set out by the EPA and EU [3], [4].  Some of the combustion strategies being explored focus exclusively on lower pilot injection masses (e.g., less diesel fuel) to better exploit the benefits of NG.  For this reason, much of the literature revolves around small diesel pilot mass.  Here, experimental studies contributing to the understanding of DIDF and HPDI combustion are explored in more detail.  It should be noted that DIDF and HPDI combustion 10  strategies were selected due to the contrast it provides (i.e., premixed vs. non-premixed combustion respectively).  2.1.1 DIDF Background A conceptual model is generally regarded as a high-level representation of a complex system.  The conceptual models of DIDF presented here are a composition of the different stages of combustion and serves aid discussion and explain some of the in-cylinder phenomenon that may be identified through the course of this work.    One of the most widely accepted conceptual descriptions of DIDF operation is proposed by Karim et al. (1993) and treats DIDF combustion in three separate stages of heat release [8]: 1. Combustion of the premixed diesel pilot fuel 2. Combustion of the gaseous fuel component in the immediate vicinity of the ignition and combustion centers of the pilot 3. Pre-ignition reactivity and subsequent turbulent flame propagation through the remaining mixture A schematic representation of the three combustion stages adapted from Karim’s work is presented in Figure 3.   11   Figure 3 - Karim’s conceptual stages of dual-fuel combustion under heavy- and light-load conditions.  The total heat release rate is the summation of the three stages.  Reprinted from [8] with permission from SAE International.  The observations formed throughout Karim’s work provide a suitable setting to discuss some of the main DIDF combustion characteristics.  One of the main observations from Karim’s work include evidence of a lean methane concentration limit below which stage 3 flame propagation cannot sustain itself resulting in a small contribution to the AHRR and high emissions of partial combustion products (uHC and CO).  A key parameter describing the methane concentration is the methane equivalence ratio (ϕCH4) shown below:  𝜙𝐶𝐻4 =?̇?𝐶𝐻4 ∙𝐴𝐹𝐶𝐻4,𝑠𝑡𝑜𝑖𝑐ℎ?̇?𝑎𝑖𝑟 (1) ṁCH4 and ṁair represent the mass flow rates of methane and air into the engine while A/FCH4,stoich is the stoichiometric air-fuel ratio of methane.  For methane equivalence ratios above the lean concentration limit (ϕCH4 ≥ 0.6), stage 3 flame propagation is sustainable, resulting in more 12  complete fuel conversion.  Consequently, stage 3 becomes a significant contributor to the total AHRR with reduced emissions of uHC and CO.  Karim also notes significant dependence of methane utilization (i.e., amount of methane that has been combusted) with respect to pilot injection characteristics despite the same pilot injection mass.  In terms of other emissions, NOx production is largely associated with the quantity of the diesel pilot as well as the overall equivalence ratio and is noted to be lower than in dual-fuel operation with other gaseous fuels such as propane or ethylene due to the lower combustion temperatures.  Finally, Karim found that pre-ignition reactions of methane can produce radicals, aldehydes, and CO in significant amounts despite the absence of a pilot; this pre-ignition reactivity can produce some heat release.  A similar dual-fuel combustion descriptor has been developed by Sahoo et al. (2003) by splitting in-cylinder pressure history into 5 different stages [9], shown schematically in Figure 4.  It should be noted that this is a departure from Karim’s model which is based on HRR. 1. Ignition delay of the pilot diesel (A-B) 2. Combustion of the premixed pilot diesel (B-C) 3. Ignition delay of the primary gaseous fuel (C-D) 4. Rapid combustion of the premixed primary gaseous fuel (D-E) 5. Diffusion combustion (E-F) 13   Figure 4 - Sahoo’s conceptual stages of dual-fuel combustion.  Reprinted from [9] with permission from Elsevier.  In both models, the conversion mechanism for the diesel is assumed to be premixed combustion.  Following this, there is a transition into primary fuel conversion; in Karim’s model, this is described as combustion of the primary fuel in the immediate vicinity and centers of the pilot ignition sites while Sahoo treats the transition to be a short ignition delay of the primary gaseous fuel prior to its auto-ignition.  Finally, in both models, flame propagation dominates the combustion mechanism for the remainder of the gaseous fuel.  Karim’s 3-stage model [8] is frequently referenced by current dual-fuel works [10]–[13] and is widely accepted as an appropriate conceptual model of DIDF.  Each stage of combustion (Figure 3) is explored in more detail to allow the reader to fully understand each of these processes in DIDF operation.  14  2.1.1.1 Stage 1 Combustion Stage 1 combustion is described as the energy release of the diesel pilot.  This is similar to conventional diesel operation where there is generally a pilot injection (premixed combustion) prior to the main diesel injection event (diffusion combustion).  In comparing DIDF pilot characteristics with that of conventional diesel, an increased ignition delay of the pilot is generally observed.  Reasons for this include lower compression temperatures due to the higher specific heat capacity of the fuel, and reduced oxygen concentration in the vicinity of the fuel jet due to air substitution with gaseous fuel [9], [14].  The gaseous fuel can also exhibit significant exothermic pre-ignition reactivity [8], creating competing radicals and partial combustion products that may influence the length of the pilot ignition delay, pressure rise rates, as well as the incidence of knock under high load operation [14].  These effects are complex and hard to quantify; the pilot ignition delay under DIDF operation is ultimately dependent on variables such as the type of gaseous fuels used, in-cylinder charge properties (temperature, residuals), equivalence ratio, heat transfer, and pilot injection parameters, many of which are considerably different in DIDF operation versus diesel [15].    2.1.1.2 Stage 2 Combustion Stage 2 combustion is described as the energy release of the gaseous fuel component that is in the immediate vicinity of the ignition and combustion centers of the pilot.  During the pilot injection process, a small amount of gaseous fuel is entrained with the diesel jets, and is subsequently ignited alongside the diesel pilot auto-ignition to contribute to stage 2 combustion.  Under high-load conditions, the total energy release during stage 2 can be significantly higher than the pilot, combining for a high heat release rate and pressure rise rate while low-load conditions produce 15  significantly lower heat release rates.  Where the gaseous fuel significantly affects the pilot characteristics in stage 1, so too does the diesel pilot injection significantly affect stage 2 combustion; the authors of [8] note that an increase in the pilot fuel quantity will tend to entrain larger amounts of the gas-air mixture and increase the total heat release (e.g., increased stage 2 contribution) despite the same gas equivalence ratios.  The methane equivalence ratio is defined in eq. (1) but is also relevant to any gaseous fuel if the parameters are replaced with their respective gas-specific values.  In contrast to Karim’s model where the gaseous fuel begins combustion in parallel with the diesel pilot, Sahoo et al. (2009) describes the transition to gas combustion as a serial process [9].  Sahoo et al. considers the gaseous component to have its own ignition delay period (process C-D in Figure 4), followed by the unstable and rapid combustion of the primary fuel that precedes diffusion combustion (D-E followed by E-F in Figure 4).  In both models, the ignition of the gaseous fuel is similarly dependent on the in-cylinder variables as described in stage 1 combustion.  2.1.1.3 Stage 3 Combustion Stage 3 combustion is described as the energy release of the gaseous fuel component via turbulent flame propagation through the remainder of the cylinder charge.  The combination of stage 1 and stage 2 combustion form the in-cylinder conditions that dictate stage 3 attributes. The most crucial factor in determining stage 3 characteristics is the gas equivalence ratio, which dominates the heat release rate of this final stage.  The prior stages are pilot-dominated events and are highly dependent on the characteristics of the injection.  Conversely, the gas equivalence ratio determines the extent of the flame propagation or diffusion combustion that occur during stage 3.  The authors 16  of [8], [10] have acknowledged a lean flammability limit (ϕCH4 ≈ 0.6) in their experimentation with DIDF using methane, which is essentially the equivalence ratio below which flame propagation can no longer be reliably sustained.  For this reason, low-load DIDF exhibits extremely poor emissions in terms of CH4 and CO partial combustion products, as total gaseous fuel utilization is low.  Above this lean flammability limit, flame propagation can contribute greatly to total heat release, and significantly higher rates of fuel utilization is achieved.    2.1.1.4 DIDF Operating Space and Conceptual Model Summary A recent investigative campaign [16] took Karim’s 3-stage model and advanced it further with the purpose of developing a map of the DIDF operating space for a limited set of engine parameters.  The combustion stages described in [16], in contrast to Karim’s conceptual descriptors, are mathematically described through the AHRR and its derivative.  Figure 5 illustrates this point. 17   Figure 5 - Rochussen’s mathematical descriptions of the different stages of DIDF combustion.  Reprinted from [16] with permission from SAE International.  In [16], the engine speed was kept constant at 600 rpm under naturally aspirated and non-EGR conditions while fueling parameters were varied. The diesel pilot was varied via injection mass, pressure, and timing, while methane was varied in mass only.  Critical set-point parameters for DIDF experiments are defined by eq. (1) to (4).  These include the pilot ratio Rpilot (ratio of pilot 18  fuel to total fuel on an energy basis), methane equivalence ratio ϕCH4, global equivalence ratio ϕglobal, and global air-fuel ratio (AFR).  While these variables do not fully define the operating point, they are vital in the definition of different operating modes within the DIDF operating space.  𝑅𝑝𝑖𝑙𝑜𝑡 =?̇?𝑑𝑖𝑒𝑠𝑒𝑙 ∙ 𝐿𝐻𝑉𝑑𝑖𝑒𝑠𝑒𝑙?̇?𝑑𝑖𝑒𝑠𝑒𝑙 ∙ 𝐿𝐻𝑉𝑑𝑖𝑒𝑠𝑒𝑙 + ?̇?𝐶𝐻4 ∙ 𝐿𝐻𝑉𝐶𝐻4 (2)  𝜙𝑔𝑙𝑜𝑏𝑎𝑙 =?̇?𝑑𝑖𝑒𝑠𝑒𝑙 ∙A𝐹𝑑𝑖𝑒𝑠𝑒𝑙,𝑠𝑡𝑜𝑖𝑐ℎ?̇?𝑎𝑖𝑟+ 𝜙𝐶𝐻4 (3)  𝐴𝐹𝑅𝑔𝑙𝑜𝑏𝑎𝑙 =1𝜙𝐶𝐻4∙𝐴𝐹𝐶𝐻4,𝑠𝑡𝑜𝑖𝑐ℎ (4) ṁair, ṁdiesel, and ṁCH4 represent the measured mass flow rates of air, diesel, and methane into the engine.  A/FCH4,stoich was previously defined as the stoichiometric air-fuel ratio of methane, and LHVCH4 and LHVdiesel are the lower heating values of the two fuels.    The authors of [16] observed each of the stages of combustion identified in Figure 5 throughout the range of tested fueling parameters and equivalence ratios but found significant variability in the characteristics of each stage.  The authors postulated that this represents a wide range of performance and perhaps changes in the fundamental combustion mechanism (e.g., flame propagation, pre-mixed combustion).  The characteristics of each stage was later used to describe the boundaries within the operating space of DIDF, where changes in combustion mechanism occurs.  These combustion mechanisms or modes are identified as pilot independent combustion, 19  flame propagation combustion, and non-flame propagation combustion, shown in Figure 6, and were developed based on thermodynamic analysis and exhaust stream measurements.    Figure 6 - Description of flame combustion modes under the experimental conditions described in [16].  Reprinted from [16] with permission from SAE International.  The aforementioned models of DIDF use thermodynamic (heat-release rate, in-cylinder pressure) and exhaust emissions analysis to provide the foundation for understanding the key parameters and performance aspects in DIDF operation.  However, the conclusions relating to the actual combustion mode are based on conjecture and are not fully substantiated in the sense that the classification of combustion regimes or type of combustion mode is not quantified via flame-speeds calculations or measurements.    20  In comparison, conventional diesel is a well-established technology that is well understood while DIDF has not reached the same level of maturity.  As a case in point, the diesel literature has developed a complete and widely accepted conceptual model of reacting diesel fuel jets in [17] using laser-sheet imaging techniques, while DIDF is limited in the details of understanding combustion processes.  The models describing DIDF are general in nature, and do not cover the entire gamut of DIDF operating modes, nor does it discuss important in-cylinder aspects such as PM and NOx formation regions, uHC sources, and the local equivalence ratio.  DIDF has increased complexity by virtue of a second fuel.  To develop a comprehensive model of DIDF requires advanced optical diagnostic tools with experimentation across a wide range of operating modes.   2.1.2 HPDI Background HPDI operates on a similar basis as conventional CI diesel engines by directly injecting both diesel and natural gas into the combustion chamber late in the compression stroke.  The amount of diesel fuel is small relative to the NG amount and is injected prior to the introduction of NG to the cylinder.  The diesel auto-ignites and acts as a reliable combustion initiator for the NG; the main heat release is primarily due to non-premixed combustion of the turbulent gas jet.    HPDI is of interest here because it provides a contrasting combustion mode (non-premixed) for discussion.  Unfortunately, there are no examples of a conceptual model describing HPDI combustion and the relevant emissions in detail.  Patychuk et. al [18] provides some discussion on this point, and states that pollutant formation in HPDI is similar to that of conventional diesel engines and conceptual models by Dec and Tree [17], [19] are generally applicable.  The reader is directed to Dec’s model [17] for information regarding diesel emission formation mechanisms.  A 21  concise description of the latest HPDI technology and its general performance and emissions characteristics can be found in [20].  For some examples of detailed works investigating the various combustion aspects of HPDI, the reader is referred to [18] (effect of intake and exhaust valve timing), [21] (effect of late post injection of NG), [22] (effect of injection pressures up to 600 bar), and [23] (effect of high levels of EGR).  2.1.3 Optical DIDF Studies Engine combustion studies using optical diagnostics can provide substantial insight to in-cylinder processes compared traditional analysis techniques.  An example of this is the extensive use of laser-sheet imaging in [17] that facilitated the development of a widely accepted conceptual model explaining reacting diesel fuel jets.  It is important to note however, that similar investigations with optical access requirements is performed in optical engine facilities that have several key differences in its mechanical configuration and operation compared to a conventional thermodynamic or all-metal engine.  Skip firing (a technique used to protect optical components) can contribute heavily to differences in the combustion environment by skipping combustion cycles and therefore change the recirculated gas composition and in-cylinder temperatures.  Changes to compression ratio, cylinder liner or piston materials (quartz windows), and load limitations are just a few of the other notable differences.  With these distinctions in mind, optical studies, while useful, can be limited in their application and utility for real world operating conditions.  This section provides the reader with a summary of some notable optical studies in DIDF combustion. HPDI optical studies are excluded from the discussion as there are currently no published optical HPDI investigations.  However, future work at the test facility described in §3.1 22  will be capable of optical investigations across a range of combustion strategies including HPDI and DIDF and is described in detail in [24].  One of the more recent works in optical DIDF [10] utilize an optical engine with a Bowditch piston to characterize the combustion regimes occurring in DIDF.  The engine used in the investigation features titanium piston with a flat quartz window in place of the traditional piston bowl, and for some experiments, utilized a similar piston with additional quartz windows housed within the piston bowl wall for secondary access.  Dronniou et al. set out to characterize the DIDF combustion regimes as auto-ignition (or sequential auto-ignition), flame propagation, diffusion, or a combination thereof using high-speed cameras simultaneously measuring spatially resolved natural luminosity (NL) and single-shot OH* chemiluminescence.  In addition, non-reacting planar laser-induced fluorescence (PLIF) experiments provided knowledge on the temporal and spatial evolution of the in-cylinder fuel distribution of the pilot fuel as a compliment to the NL and OH* images.  The methane equivalence ratio was varied while holding the pilot injection parameters constant.    For the limited set of test conditions, the authors of [10] noted several points of interest.  For low methane equivalence ratios (ϕCH4 = 0.56), NL and OH* chemiluminescence imaging both showcase a significant central area of low combustion intensity; it is suggested that this may be the cause of significant uHC (unburned premixed fuel) emissions for operating points at lower methane equivalence ratio.  High temperature reaction zones (represented by the OH* chemiluminescence signal) develop from the periphery of the imaging area but do not progress much further towards the central region.  Figure 7 (left) showcases this behaviour.  The AHRR 23  showcases an asymptotic tail due to the slower diffusion combustion and expanding cylinder volume.  For high methane equivalence ratios (ϕ = 0.94), NL and OH* chemiluminescence imaging showcase uniform coverage of the combustion chamber.  The high temperature reaction zones (represented by the OH* chemiluminescence signal) develop from the periphery of the imaging area towards the central bowl region up until full coverage is achieved.  Figure 7 (right) showcases this behaviour.  DIDF strategies utilizing near-stoichiometric equivalence ratios achieves a higher level of combustion efficiency (less uHC, partial combustion products).  The AHRR showcases a bell-like shape not unlike those found in spark-ignited (SI) gasoline engines.     ϕglobal = 0.56 ϕglobal = 0.94 Figure 7 - OH* chemiluminescence imaging of low- and high-load DIDF operation (left and right respectively).  The OH* radical represents high temperature reaction zones associated with a flame front.  Images are taken in an optical engine.  Reprinted from [10] with permission from SAE International.  Under lean conditions, the authors of [10] noted that combustion was observed to be dominated by auto-ignition of the pilot, entrained premixed fuel, and air.  This phenomenon is not 24  immediately obvious in Figure 7, but Figure 8, which shows DIDF experiments in a rapid compression expansion machine, exemplifies pilot dominated combustion in lean conditions.  The subsequent combustion following pilot ignition tends to progress via diffusion mode moving from richer to leaner regions, and is controlled by the mixing between hot burned gases and fresh premixed charge.  Under rich or near-stoichiometric conditions (i.e., ϕCH4 = 0.94), DIDF combustion was observed to be mainly a flame propagation mechanism.  The pilot auto-ignites around the periphery of the cylinder at multiple locations, and propagates inwards towards the central part of the cylinder, consuming fresh premixed charge.  Dronniou’s optical study here demonstrates the utility of imaging systems in combustion research.  The qualitative differences in combustion mechanisms between the high and low load are obvious in the optical images (Figure 7) and provide much needed insight as to why low-load DIDF combustion exhibit such high uHC and CO emissions; in the past, these characteristics were determined via exhaust measurements and actual in-cylinder phenomenon leading to the high emission rates were not well understood.    Another recent optical DIDF study [12] investigates the effect of charge bulk motion on dual-fuel combustion development and related pollutant emissions through variations in methane supply methods as well as the use of variable intake ports (tumble and swirl).  This study has several areas of focus; however, for the sake of brevity, the discussion will be concentrated on the optical results.  The optical evaluation of charge bulk motion was only performed at low-load operating points and is conducive to the previous discussion of [10] regarding low-load DIDF and high levels of uHC and CO (Figure 7, left).  In [10], the methane premixed charge is considered homogeneous while 25  in [12], the variable injection strategies and intake ports can combine for a stratified mixture.  Charge stratification offers a different perspective on DIDF operating strategies and allowed the authors of [12] to evaluate DIDF combustion development under low-load stratified conditions.    The authors discerned from low-load thermodynamic AHRR results that the swirl-only intake port configuration offered better oxidation characteristics during late-stage combustion (i.e., stage 3 combustion [8], stage 2b combustion [16]) and was evidenced by higher AHRR.  The analysis found shorter late-stage burn durations (crank angle duration between 50% to 95% of total heat release, CA50 – CA95), higher late-stage heat release, and lowest emissions of NOx and HC (for the 1000 bar pilot injection pressure test case).  It was hypothesized that the swirl-only intake configuration offered the best mixing characteristics between the hot burned gases and unburned premixed charge and subsequently enhanced oxidation.  The authors then used an in-cylinder endoscope and high-speed CCD camera to measure combustion luminosity; the measured luminance corroborate the above observations and showcase higher peak luminosity (representing higher intensity in combustion), as well as quicker combustion.  It is interesting to note that while the authors also had success with the swirl-only intake configuration in the high-load tests, none of these were observed optically presumably due to load limitations or cleaning requirements for the endoscope.  While the use of optical diagnostics in this study is fairly limited, it still demonstrates the utility of optical access in a research engine in that it can substantiate hypotheses made from thermodynamic data.  A contrasting work in DIDF utilizes a rapid compression / expansion machine (RCEM) [13] for the study of ignition and combustion characteristics in lean premixed methane/air charge.  While 26  RCEMs differ significantly from regular engines, they offer improved optical access.  An intensified high-speed camera with 307nm bandpass filter records OH* chemiluminescence in addition to a photomultiplier recording total UV light emitted by OH*, CH, and C2 radicals.  The authors note that for the tested conditions (diesel pilot pressure 400 bar, timing 4.1 ms bTDC, and homogeneous mixtures of methane at ϕCH4 ≤ 0.65), combustion initiates along the axes of the diesel pilot jets (Figure 8) in contrast to the observations made by [10], [25] where combustion initiates at the periphery of the combustion chamber as shown in Figure 7 (right).    Figure 8 - OH* chemiluminescence images of DIDF combustion in a RCEM.  The OH* radical represents high temperature reaction zones associated with a flame front.  Ignition originates along the axes of the pilot diesel jets and subsequently propagates through the premixed charge.  Reprinted from [13] with permission from SAE International.  27  While it is not immediately clear as to why the various experiments showcase different combustion mechanisms, the authors of [26] were able to replicate the two different combustion mechanisms (ignition at cylinder periphery vs. ignition along diesel jet axes) by altering the pilot injection pressure for a small range of ϕglobal and ϕCH4.  The authors utilized a single cylinder optical research engine equipped with two high-speed cameras for simultaneous NL and OH* chemiluminescence imaging and considered the effect of fueling parameters, including injection pressure (same ϕglobal and ϕCH4 at different pilot injection pressures of 1300 and 300 bar), as shown in Figure 9.    Figure 9 - Simultaneous NL and OH* chemiluminescence image sequences for two different pilot injection pressures at the same global and methane equivalence ratios.  Reprinted from [26] with permission from ASME.  28  Based on the above OH* chemiluminescence observations, the authors of [26] developed a conceptual image (Figure 10) of DIDF reaction zone growth at low pilot ratios (ratio of injected diesel fuel energy to total fuel energy, Rpilot) for high and low pilot injection pressures.  High pressure pilot diesel injections with low pilot ratios resulted in a flame front originating from the piston bowl periphery which propagated towards the center of the combustion chamber, as shown in Figure 10, left.  As the equivalence ratio of the premixed methane decreased, the flame front no longer propagated all the way to the cylinder center and left an unburned area (see Figure 9, OH* chemiluminescence images for high pilot injection pressure at 14.5 CAD aSOI).  Conversely, low pressure pilot diesel injections resulted in ignition in the central region of the bowl, near to the diesel spray.  The reaction zone subsequently propagated away from the ignition sites, which were no longer at bowl periphery, but rather in the central region of the cylinder where the diesel fuel jets originally were (Figure 10, right).  The extent of the propagation away from the diesel jets depended on both the pilot ratio Rpilot and methane equivalence ratio ϕCH4. While the thermodynamic AHRR recorded for the same test points do show some notable differences, they do not provide any indication of the combustion mechanism.  Only through optical diagnostics are the authors able to clearly discern the differences in combustion modes and the implications it may have on combustion efficiency and emission trends.  29   Figure 10 - Conceptual images of reaction zone growth at high and low diesel pilot injection pressures based on OH* chemiluminescence imaging for low pilot ratio DIDF operation.  Left: combustion begins at the bowl periphery and propagates inwards.  Right: combustion begins at the central region and propagates outwards.  Adapted from [26] with permission from ASME.  2.1.4 Utility and Limitations of Optical Diagnostic Tools in Engine Research A common feature between the optical studies in DIDF is that they all offer additional insight to previously unknown combustion phenomenon.  As previously discussed, there is only so much that can be inferred from a thermodynamic analysis, and the value of optical diagnostics should be clear.  However, these optical diagnostics are not without limitations.  As mentioned in [7] and §1.1, optical engines often have significant differences with their all-metal counterparts.  The differences in materials, operating temperatures, skip-firing requirements, cleaning requirements, load limitations, are just a few of the characteristics that make it difficult to directly compare between the two.  While there are some mitigation strategies such as intake and coolant heating, compression ratio adjustments, and the use of simulated EGR, researchers must bear in mind the 30  difficulties in matching engine performance between optical and all-metal configurations and is the motivation behind the development of the thermo-optical approach proposed.    2.2 In-Cylinder Optical Diagnostic Tools This section will outline the various optical diagnostic tools that have been developed for combustion studies with emphasis on in-cylinder fuel concentration, temperature, and soot concentration measurements.  The reader is presented with a basic overview of the operation and design considerations for each technique and show examples of their use in combustion studies as a precursor to the experiments presented later in this thesis.  A few examples of the different types of optical measurements are listed below: 1. OH* chemiluminescence imaging of DIDF combustion representative of high-temperature reaction zones or flame front propagation [10], [13], [26] 2. Planar laser-induced fluorescence (PLIF) imaging of the temporal and spatial evolution of diesel fuel jets [10] 3. Laser-sheet imaging of reacting diesel fuel jets [17] 4. Simultaneous chemiluminescence, soot laser-induced incandescence (LII), OH* PLIF, used to image early-injection LTC in a heavy-duty diesel engine [27] 5. Two-colour pyrometry probes used for measuring in-cylinder soot temperature and soot concentration [28]–[33] 6. Spatially resolved colour-ratio pyrometry temperature and soot concentration estimation [34]–[41] 7. In-cylinder fuel concentration [42]–[45]  31  The listed studies are by no means a comprehensive review of the literature but is intended to show the different optical tools required for the various measurements.  It should be noted that the methodology significantly varies, and can be passive (e.g., natural luminosity) or active (e.g., laser-induced incandescence), spatially resolved (e.g., 2D camera imaging) or line-of-sight (e.g., 1D probe based measurements), or even a point location measurement.  A common theme, especially for any spatially resolved work (e.g., PLIF, soot LII), is the significant difficulty involved with setting up optical equipment, focusing photodetectors and/or cameras, and mitigating the limitations in engine operation because of optical access.  With these intricacies in mind, this investigation aims to lay the groundwork for a thermo-optical approach that addresses many of these complexities through the “on-engine” calibration of a series of optical probes with the spatially resolved measurements.  The thermo-optical approach revolves around the use of temporal and spatial optical data from the optical configuration to guide interpretation of the probes in the all-metal engine configuration.  Further discussion on the links in the thermo-optical methodology is presented in §2.3.  2.2.1 In-Cylinder Fuel Concentration Measurements Measuring the in-cylinder fuel concentration can provide significant insight to various combustion phenomenon, especially in the context of premixed gaseous fuels.  Assessment of fuel concentration histories show fuel conversion mechanisms as well as differences between operating strategies and can provide some indication of gaseous fuel emissions.  In-cylinder fuel concentration measurements are difficult in the sense that there is no natural luminosity of the fuel to take advantage of.  Consequently, a source of fuel vapor/droplet excitation or illumination is 32  required, as can be found in laser-based studies such as Rayleigh scattering technique and PLIF [43], [44].    In [43], irradiation of laser light on gaseous fuel molecules causes Rayleigh scattering.  The scattering intensity that can be detected is characteristic of the irradiated chemical species and is a function of the intensity of the laser source and the concentration of fuel molecules.  A CCD camera was used to capture the two-dimensional fuel concentration distribution over various fuel injection and intake flow conditions in an SI engine.  A sample of the measurement output is shown in Figure 11.  Figure 11 - 2D air/fuel ratio (fuel concentration) profile in an SI engine using Nd:YAG laser Rayleigh scattering.  The figure has thresholding applied with the black areas representing air/fuel ratios below 13 and white areas representing air/fuel ratios above 13.  Reprinted from [43] with permission from SAE International.  A more modern fuel concentration diagnostic technique in literature is discussed in [44] and uses PLIF as the measurement method.  Here, a tracer is added to the fuel (gasoline) such that laser-33  induced fluorescence of the fuel is possible.  A laser is introduced to the combustion chamber via windows in the side of the cylinder liner and the subsequent fluorescence is imaged through a Bowditch piston equipped with a fused silica window.  Again, a CCD camera is used to capture the two-dimensional fuel concentration distribution under negative valve overlap conditions.  A schematic of the optical setup used as well as a sample of the fuel concentration measurements are shown in Figure 12 and Figure 13.  Figure 12 - Optical setup used in dual-wavelength PLIF measurements of temperature and fuel concentration.  Reprinted from [44] with permission from SAE International.  34   Figure 13 - Sample fuel concentration measurements (right) from PLIF tracer experiments.  Reprinted from [44] with permission from SAE International.  These setups are complicated, and can be extravagant in cost.  A simpler setup was developed in [42], [45] which utilizes a light absorption method at a single point within the combustion chamber.  In [42], the authors developed a modified spark plug with optical access and an attached measurement volume.  A fiber optic cable directs light from a chopped (modulated) tungsten halide lamp to the modified spark plug which illuminates the measurement volume.  The source light is subsequently reflected back, and a second fiber optic cable is used to deliver the reflected light to a bandpass filter and photodetector.  The light within the measurement volume is absorbed by the fuel molecules in the space, and the signal attenuation is representative of the fuel concentration.  As a similar version of this sensor is employed in this investigation [46], details of its operating 35  theory and setup are discussed later on in Chapter 4: LaVision Internal Combustion Optical Sensor (ICOS).   2.2.2 In-Cylinder Soot Temperature and Soot Concentration Measurements Soot temperature and concentration measurements provide key information regarding combustion processes such as NOx formation regions or soot formation and oxidation characteristics.  The pyrometry technique employed in this investigation can simultaneously measure both soot temperature and soot concentration at high-speeds, and the two measurements are grouped together here for brevity.  A number of methods exist for the measurement of temperature; the sodium reversal method, Rayleigh scattering, velocity of sound, and Coherent Anti-Stokes Raman Spectroscopy (CARS) are some examples of temperature measurement techniques.  However, a review of in-cylinder diagnostic techniques by Brehob and Kittelson [47] has identified two techniques that are more applicable to CI engines: two-colour pyrometry and CARS.  CARS is an advanced technique utilizing multiple lasers to serially excite gaseous molecules to higher energy states at species specific wavelengths.  The molecules subsequently emit radiation as the excited molecules relax into lower energy states, and through the use of narrow spectral linewidths, interference by radicals or high soot incandescence can be avoided.  The resulting radiation signal can then be traced back to a temperature and species concentration.  Two-colour pyrometry, on the other hand, is relatively simple, inexpensive, and easier to implement.  It is a passive method relying on soot incandescence to provide a signal for measurement.  The signal is measured at two different wavelengths using 36  narrow band-pass filters, and it is correlated back to soot temperature and soot concentration.  The details of its operation will be documented in the next section.  As with thermometry, there is also a number of PM measurement methods utilized in CI engines.  Point location sampling, laser-induced incandescence, light extinction, and two-colour pyrometry are several of the options noted by [47].  Point sampling relies on physically removing a sample from the combustion chamber, but does not provide the instantaneous in-cylinder soot concentration and has issues with smearing and coagulation of soot particles during the process of sampling.  In contrast, both light extinction and two-colour pyrometry can provide instantaneous soot concentration information, at the cost of optical access requirements.  As with the earlier discussion regarding CARS, light extinction is relatively more advanced than two-colour pyrometry, as it requires a light source passing through a measurement volume.  The light passing through the measurement volume is attenuated, and the light extinction can be used to track the instantaneous soot concentration.  Due to the active nature of the technique, PM can be measured throughout the entire engine cycle, whereas two-colour pyrometry can only be employed when there is significant soot incandescence.  Since the two-colour method has the versatility to simultaneously measure soot concentration and temperature at a relatively low cost of implementation, the next sections document its operating theory and application to combustion engines.   37  2.2.2.1 Two-Colour Pyrometry Technique Two-colour pyrometry is an analysis technique used to calculate the soot temperature and in-cylinder soot concentration, based on the measured thermal radiation from in-cylinder soot particles at two different wavelengths.  The measured radiation intensity is combined with a series of assumptions, empirical equations, and heat transfer equations to estimate the soot temperature and concentration and will be discussed in the following sections.  The technique has been used in the past with some success in the study of combustion processes [28]–[41].  Some of these measurements were spatially resolved works using cameras (2D) [34]–[41] while others are line of sight (1D) probe-based measurements [28]–[33].  These measurements have the potential to provide spatial insight to soot formation zones, and can also provide information regarding PM formation and oxidation rates when high-speed instrumentation is used.  In the current investigation, a pyrometry probe has been designed and built with the aim of employing the two-colour pyrometry technique to the engine facility described in Chapter 3: Experimental Facility and constitutes a significant portion of the work presented herein.  This section provides the reader with the necessary background to understand the two-colour pyrometry technique and presents some notable studies in which the pyrometry technique is utilized.    2.2.2.2 Heat Transfer Background Theory A blackbody is defined as an object that absorbs all incident electromagnetic radiation, and when in thermal equilibrium, emits radiation according to Planck’s law (i.e., the spectral radiation intensity from a blackbody is a function of temperature alone).  Because of the blackbody property that all incident radiation is absorbed, its emissivity is unity, ε = 1, and is therefore a perfect emitter.  A grey-body, on the other hand, has an emissivity ε < 1.  Most objects found in the real world are 38  grey-bodies and their emissivity as a function of wavelength and temperature ελ(λ,T) can be described by eq. (5) which is the ratio of the monochromatic (single wavelength) radiation intensity of a grey-body iλ(λ,T) to its blackbody counterpart ib,λ(λ,T) at the same temperature.  𝜀𝜆(𝜆, 𝑇) =𝑖𝜆(𝜆, 𝑇)𝑖𝑏,𝜆(𝜆, 𝑇) (5) The monochromatic radiation intensity of a grey-body can be rewritten in a way that is defined by its apparent temperature Ta, or the equivalent blackbody temperature that produces the same monochromatic intensity ib,λ(λ,Ta) as the grey-body .  This is defined in eq. (6).  𝑖𝜆(𝜆, 𝑇) = 𝑖𝑏,𝜆(𝜆, 𝑇𝑎) (6) Through these series of definitions and the fact that a blackbody is a perfect emitter, it should become clear that apparent temperature Ta is less than that of the true object temperature T.  Finally, the monochromatic intensity emitting from a blackbody follows Planck’s Law and is a function of λ and T and is expressed in eq. (7) below.  𝑖𝑏,𝜆(𝜆, 𝑇) =2𝐶1𝜆5 [𝑒(𝐶2𝜆𝑇) − 1] (7) C1 and C2 are known blackbody radiation constants where 𝐶1 = 3.742×108𝑊 ∙𝜇𝑚4𝑚2 and 𝐶2 =1.439×104𝜇𝑚 ∙ 𝐾.  However, the emissivity of a diesel flame is not known, and an empirical relationship estimating flame emissivity is presented below.  39  2.2.2.3 Hottel Broughton Correlation Hottel and Broughton [48] proposed an empirical relationship for the emissivity of a luminous gas flame in eq. (8):  𝜀𝜆(𝜆) = 1 − 𝑒−𝐾𝐿𝜆𝛼  (8) K is the absorption coefficient per unit length, and L is optical path length or flame thickness.  The product of KL is directly related to the particulate matter concentration in the flame as shown by [28].  This correlation shows that soot emissivity is dependent on both the wavelength as well as the optical and physical properties of the soot and is described by Hottel and Broughton using the index parameter α.  Matsui et al. [28] performed a series of validation studies that confirmed Hottel and Broughton’s findings and determined that α =1.39 for measurement wavelengths in the visible spectrum.  Finally, by combining eq. (5) - (8) with the measured radiation intensity at two wavelengths, a system of equations with one unknown, flame temperature T, can be numerically solved using a nonlinear system solver (9).    [1 −𝑒𝐶2𝜆1𝑇 − 1𝑒𝐶2𝜆1𝑇𝑎1 − 1]𝜆1𝛼= [1 −𝑒𝐶2𝜆2𝑇 − 1𝑒𝐶2𝜆2𝑇𝑎2 − 1]𝜆2𝛼 (9) With the temperature known, KL is determined using eq. (5) and (8).   2.2.2.4 Parameter Selection in the Two-Colour Method As previously mentioned, the selection of parameter α is dependent on wavelength.  A summary of the technique by Ladommatos and Zhao [49], [50] shows that the method is not very sensitive 40  to α within the visible wavelength measurement range (α is generally selected to be 1.39), while significant sensitivity to α selection is noted in the infrared (α ranges from 0.9 to 1 for a range of wavelengths from 1 μm to 5 μm [28]).  As such, many studies utilize pyrometry within the visible spectrum to take advantage of this trait.  In addition to a constant index α, there are also other benefits in selecting visible wavelengths.  Combustion emission in the infrared is quite strong (i.e., wall radiation and reflections), and visible wavelengths avoid this issue altogether.  Signal sensitivity is also improved due to the larger spectral emissive power differences in the visible band with respect to temperature.  In conjunction with this, signal differences between wavelengths are larger in the visible, improving signal to noise ratio.  Finally, wavelength selection should consider other factors such as avoiding spectral bands where combustion radicals may emit strongly.  The end result from the measurement of apparent temperatures Ta, is the soot temperature, as well as the KL factor, which represents the soot concentration within the measurement volume or area.  2.2.2.5 Validation and Application Studies Matsui et al. [28] performed the first validation studies using the two-colour method in a diesel flame.  They investigated the application of the two-colour method using wavelengths in the visible region and performed validation studies for the calculated emissivity of a diesel flame and for the calibration of absolute temperature. Also addressed is the effect of a non-uniform flame temperature and soot concentration distribution on the measurement.  Calibration was performed with a custom blackbody furnace along with a standard tungsten strip lamp.  Their main goals were to: 41  1. Establish the two-colour method as a technique for measuring diesel flame temperature and soot concentration 2. Verify the emissivity estimation provided by the Hottel and Broughton correlation 3. Compare the two-colour method with an established emission-absorption method for temperature measurement  More recently, a two-part review of the method was provided by Ladommatos and Zhao [49], [50].  They examined the theoretical basis of the method with consideration to the assumptions used in the evaluation of flame temperature and soot concentration while the second part details practical problems involved in building a two-colour measurement system.  The summary provided by Ladommatos and Zhao is concise, and the reader is directed to those papers for a complete review.  The discussion of two-colour pyrometry and its applications will be limited to those of the line of sight probe type.  While there are also variants of pyrometry that use high-speed 2D imaging cameras for spatially resolved results, these setups are beyond the scope of this investigation and the reader is directed to [37] for reference.  As previously mentioned (§2.2) there have been a number of studies using the two-colour method throughout the past couple decades [29]–[31].  These studies all commonly feature some form of probe assembly utilizing a sapphire rod and fiber optic connections to external photodetectors.    In [29], a pyrometry probe assembly was developed and installed in place of an exhaust valve.  As the probe assembly is passed through the coolant jacket, special care was taken in the design to seal the optical components and the combustion chamber from leaks.  In addition to the coolant 42  seals, the probe was designed to seal combustion pressures.  The assembly is consisted of a sapphire rod brazed in invar, a set of apertures, and finally a fiber optic adapter.  These components were connected to a trifurcated cable, passing the combustion signal to 3 bandpass filters at 550nm, 700nm, and 850nm respectively (allowing for up to 3 combinations of wavelengths in the method).  Photodiodes measured the filtered intensity and was recorded through a data acquisition system.  A cross section of the probe and a schematic of the setup is found in Figure 14 and Figure 15.  Figure 14 - Two-colour pyrometry probe cross section.  The probe is consisted of an outer sleeve that houses the main optical components and protects it from coolant and oil.  Reprinted from [29] with permission from SAE International. 43   Figure 15 - Schematic of the optical connections in a pyrometry application.  Here, 3 wavelengths are used rather than 2 allowing for up to 3 combinations of colours.  Reprinted from [29] with permission from SAE International.  One of the key features in the probe developed by Yan and Borman is the inclusion of self-cleaning geometry in front of the sapphire rod.  The probe geometry utilizes a flow restriction area in front of the sapphire rod, forcing hot combustion gases in and out of the cavity surrounding the rod with the cycles of combustion.  While this restriction also reduces the viewing area, the forced flow is believed to clean the window by continuously passing hot combustion gases across the tip of the sapphire rod, prevent thermophoretic deposition of soot, as well as by the mechanical “sweeping” action of high velocity gases.  A flow model of the restriction was developed by the authors which predicts flow velocities across the window as high as 40m/s.  In [30], a commercial pyrometer developed by AVL (VisioFEM) is utilized.  While the details of the probe geometry and setup are sparse, the overall assembly is similar.  A probe with a sapphire 44  lens is inserted into the combustion chamber, and a bifurcated fiber optic cable transmits combustion luminosity to a set of bandpass filters and photodetectors.  One thing to note here is the exclusion of any sort of soot-deposition mitigation strategy.  The authors remark that AVL recommends a particular testing methodology that tracks the attenuation of the probe as the window fouls and can be used to correct for the measured radiation signals.  While this methodology is not ideal for long measurement campaigns, it offers some insight to potential mitigation strategies to window fouling.    In more recent applications of pyrometry [31], [32], a pyrometry prototype with a heated window feature is utilized in the study of exhaust and in-cylinder soot in diesel engines.  Again, the details of the probe design are sparse; however, the authors mention the active heating and subsequent high temperature of the sapphire window (600°C) as a mitigation strategy for reducing window contamination.    2.2.2.6 Two-Colour Pyrometry Summary Due to the simplicity of the technique, two-colour pyrometry was selected for use throughout this investigation in the measurement of in-cylinder temperature and soot concentration.  The motivations behind the measurement selection were documented, and with the on-going development of the spatially resolved optical equipment and analysis tools in the engine facility described in Chapter 3: Experimental Facility, the probe lends itself to future growth in its role at the facility as a part of the thermo-optical diagnostic tool chest.  The details of the custom probe design and implementation in the test facility is discussed in Chapter 5: Pyrometry Probe.  45  2.3 Links in the Thermo-Optical Methodology The above sections showed some in-cylinder optical diagnostic tools that are used in combustion studies with an emphasis on fuel concentration, soot temperature, and soot concentration measurements.  To better link these optical diagnostics to conventional thermodynamic results, the following discussion will provide some examples from literature to show the utility of these measurements.   2.3.1 Thermo-Optical Application of Fuel Concentration Measurements Grosch et al. [42], and Koenig and Hall [45] present similar investigations for the crank angle resolved determination of fuel concentration and air/fuel ratio in SI engines.  In both cases, a modified spark plug provides an optical measurement volume that enables an infrared absorption method to determine the local fuel concentration.  The details of its operation are documented in detail in [42], [45] and §4.2 while the current discussion will focus on the utility of the results.  In [42], fuel concentration variations in motored and fired engine cycles for an SI engine were investigated.  The fuel concentration was compared against pressure data where phenomenon including delayed combustion and misfire were identifiable in the crank-angle resolved fuel concentration history. In [45], the mixture formation aspects are considered and the in-cylinder mixture inhomogeneities at all parts of the cycle for a variety of engine speeds are discussed.    2.3.2 Thermo-Optical Application of Pyrometry Yan & Borman [29] employed pyrometry in a single cylinder research engine.  The head was modified (an exhaust valve was removed) to facilitate the implementation of the probe body.  The 46  authors evaluated the novel self-cleaning design of the probe as well as the effect of the field of view (14° vs. hemispherical) on the temperature and KL results.  They noted that the probe exhibits considerable cycle-to-cycle variations.  Early stage soot production was also noted to be rapidly oxidized due to the high-temperature with peak soot temperatures ranging from 2420 to 2500K.    Kunte et al. [33] used probe-based pyrometry to evaluate the influence of fuel composition, injection parameters, and exhaust gas recirculation on the temporal soot evolution during diesel combustion.  The authors used AHRR comparisons and found that the start of the KL signal coincides with the beginning of the diffusion phase of combustion.  Another key finding was that the in-cylinder KL values corresponded well to exhaust PM concentration.    Kirchen et al. [32] evaluated the effect of two different fuels on the combustion and soot emissions in a common-rail diesel engine.  Pyrometry and cylinder pressure measurements were compared, along with exhaust soot measurements.  It was proposed that three parameters associated with the KL signal be used to characterize results: KLmax, KLend, and the ratio KLmax / KLend.  KLmax is the peak KL for a given test point, and KLend is the corresponding end-of-cycle representative KL value.  The authors noted strong correlation between KLend and exhaust-stream measurements (R2 ~ 0.9).  The other parameters KLmax and the ratio KLmax / KLend were used to compare the two fuels, and it was shown that a reduction in soot emissions for one of the fuel was likely due to an inhibited soot formation process, associated with the lower aromatic content of the fuel.  A more recent investigation by Vögelin et al. [31] implemented probe-based pyrometry in a multi-cylinder diesel engine.  A qualitative correspondence between the EGR rate and soot temperature 47  was found in addition to good agreement between in-cylinder KL values to exhaust PM mass measurements.  The probe was also used as an engine-out soot predictor for transient operation based on a steady-state calibration.    2.4 Summary and Literature Gap This chapter discussed the generally accepted conceptual models of DIDF combustion as well as current works that build on these models in the ongoing pursuit of increased performance in new engine development.  The HPDI combustion concept was also introduced, and a review of thermodynamic investigations was presented.  Limitations in thermodynamic analyses were identified, and a subsequent literature search for optical DIDF and HPDI works was carried out.  A review of the state of the art in literature did not uncover any optical HPDI works, though several optical DIDF investigations identified areas of focus for DIDF:  combustion mechanisms present in the DIDF operating space, as well as the low-load emission characteristics of DIDF (high uHC and CO).    In addition to the DIDF and HPDI literature assessment, combustion diagnostics for in-cylinder fuel concentration, soot temperature, and soot concentration were considered.  Existing studies were used to examine each type of measurement with the intention of understanding key parameters, design ideas, and limitations for each tool, and provided the inspiration for some of the design features applied to the custom pyrometry probe that is developed in this investigation.  Further discussion on the utility of these diagnostics was also presented in the context of how the probes may contribute to the thermo-optical methodology.  48  Upon consideration of the literature, a significant gap is observed.  In many of the studies discussed, crucial differences in the operation of optical vs. all-metal engine configurations (as noted by [7]) are not discussed.  This investigation aims to reduce the diagnostic limitations present in each of the two operating configurations by using a full optical diagnostic setup (i.e., optical engine with quartz window access) as a calibration tool for less intrusive optical measurement probes.  The knowledge gained through the simultaneous use of high-speed imaging and in-cylinder probes in an optical engine environment can be extended via in-cylinder probes to the thermodynamic configuration without the same mechanical limitations as optical engines.  This thermo-optical methodology can be used to discern in-cylinder combustion phenomena as well as identify how closely matched an optical test is to its thermodynamic equivalent.    49  Chapter 3: Experimental Facility This chapter describes the single cylinder engine facility used throughout this investigation.  The single cylinder engine facility was developed during the duration of this work, and much of the documentation below outlines the development of the optical probes and instrumentation surrounding the facility.  Descriptions of the engine facility, instrumentation, data acquisition, engine control, and optical probes are provided for the reader.   3.1 Ricardo Proteus Single-Cylinder Engine The Ricardo Proteus single-cylinder, four-stroke optical research engine was used for all testing performed in this investigation.  The Proteus research platform was developed by Ricardo with a combination of OEM and Ricardo designed components and is based around a family of Volvo inline 6-cylinder heavy-duty turbodiesel engines used in on-highway transport applications in the 1970s.  The Ricardo Proteus features the ability to switch between optical and thermodynamic configurations; the optical configuration includes an engine block extension and Bowditch piston to provide optical access via quartz glass piston bowl whereas the thermodynamic configuration is all-metal as in a conventional engine (Figure 16).  To facilitate the probe based measurements, a custom cylinder head was developed to provide additional diagnostic ports as well as to support installation of Westport’s HPDI injector.  The Proteus was operated in four different configurations which are listed below: 1. OEM Volvo TD122 head in thermodynamic configuration (DIDF only) 2. OEM Volvo TD122 head in optical configuration (DIDF only) 3. Custom UBC designed head in thermodynamic configuration (HPDI only) 4. Custom UBC designed head in optical configuration (HPDI only) 50   Figure 16 - Schematic layout of Ricardo Proteus experimental facility in thermodynamic and optical configurations.  The Proteus engine specifications are the same as in past thermodynamic and optical DIDF / HPDI studies at the same facility [11], [16], [24], [26], [46] and are listed in Table 3.  The main differences between the two heads are the revised intake and exhaust port layouts, smaller valves, and relocation of the injector to a central upright position to facilitate the installation of the ICOS and pyrometry probe.  Details regarding the OEM Volvo TD122 cylinder head and custom UBC cylinder head can be found in [24].  Images of both DIDF and HPDI cylinder heads are shown in Figure 17.  51  Table 3 - Ricardo Proteus single-cylinder research engine specifications. Displaced Volume 1998 cc Stroke 152mm Bore 130mm Connecting Rod 275mm Compression Ratio  14.25:1 (thermo) / 13.25:1 (optical) Speed Limit  2050rpm (thermo) / 1200rpm (optical) Load Limit 170bar (thermo) / 110bar (optical) Number of Valves 2 Thermodynamic Piston Bowl Eccentric toroid (80mm) Optical Piston Bowl Cylindrical (80mm) Intake Surge Tank Volume Approx. 1000L Exhaust Valve Open 145° aTDC @ 1.0mm lift Exhaust Valve Close 390° aTDC @ 1.0mm lift Inlet Valve Open -390° aTDC @ 0.6mm lift Inlet Valve Close -150° aTDC @ 0.6mm lift Diesel Pump (DIDF only) Bosch CP3 Diesel Direct Injector (DIDF only) Bosch CRIN2, 5-hole, 142° inc. angle Methane Port Injector (DIDF only) Bosch NGI2 Diesel / Natural Gas Injector (HPDI only) Westport HPDI Developmental Injector   Figure 17 - HPDI head (left) and DIDF head (right).  The HPDI head is a fully custom in-house design whereas the DIDF head is a Volvo TD122 design that has been modified to allow for in-cylinder diagnostics.  52  3.2 Instrumentation and Data Acquisition (DAQ) Prior to the first investigations in 2015 [11], [16] at the test facility described above, the facility was in the process of being set up.  The test-cell was previously uninstrumented, and without an engine control system in place, part of the research effort in this investigation was regarding installing hardware and developing software related to the data acquisition and engine control aspects.  Through this work, a data acquisition and engine control solution was procured and installed along with relevant engine sensors and actuators.  Subsequently, software for data acquisition and engine control was developed using the LabVIEW programming environment.  The instrumentation used throughout this investigation along with the data acquisition methodology is described in this section while the engine control aspects are discussed in §3.3.  A simplified process and instrumentation diagram for the DIDF and HPDI thermodynamic configurations are shown in Figure 18 and Figure 19. 53   Figure 18 - Experimental facility process and instrumentation diagram for DIDF thermodynamic engine configuration. 54   Figure 19 - Experimental facility process and instrumentation diagram for HPDI thermodynamic engine configuration. 55  3.2.1 Test-Cell and Engine Instrumentation The test-cell and engine is equipped with many diagnostic tools; many of these sensors are for auxiliary systems related to engine health and test-cell safety but are not critical to combustion analysis.  These sensors include a myriad of thermocouples, pressure sensors, dyno speed sensors, and chemical alarms.  Sensors that were crucial in the collection of combustion related engine data are listed in Table 4. Table 4 - Engine Sensors used in combustion analysis. Measurement Sensor Description Methane mass flow rate • Endress & Hauser Promass 80A Coriolis Meter Diesel mass flow rate • Mettler Toledo Viper Ex MB SM12 Gravimetric Scale Intake mass airflow rate • Bosch OEM Hot Film Sensor Intake manifold absolute pressure transducer • Kistler 4005B Piezo-resistive Pressure Sensor  • Kistler 4618A Amplifier In-cylinder pressure transducer • Kistler 6125C Piezo-electric Pressure Sensor  • Kistler 5010B Charge Amplifier Engine position and crank-synchronous timing • BEI H25 Incremental Quadrature Optical Encoder (1440 pulses / revolution) • Hamlin 55505 Hall Effect Sensor mounted on flywheel (100 teeth / revolution) • Hamlin 55505 Hall Effect Sensor mounted on cam-gear (2 teeth / revolution) Emissions measurements • AVL CEB NA2 Emissions Analysis Bench (CH4, NOx, O2, CO, CO2) • Bosch OEM LSU 4.9 Lambda Sensor Local infrared absorption fuel concentration sensor • Details in Chapter 4: LaVision Internal Combustion Optical Sensor (ICOS) Two-colour pyrometry probe • Details in Chapter 5: Pyrometry Probe  56  The Mettler Toledo Gravimetric Scale is connected to the lab computer directly through RS-232 serial port, while the AVL Emissions Analysis Bench data is communicated through Ethernet.  All other sensors are connected to the lab computer via the data acquisition system described in the next section.  3.2.2 Data Acquisition (DAQ) Data acquisition of most engine and test-cell sensors was performed via National Instrument’s CompactDAQ (cDAQ) family of products.  The major components of the data acquisition system are outlined in Table 5.   Table 5 - Data acquisition chassis module descriptions. Module Model Description NI cDAQ-9188 Chassis 8-slotted Ethernet Chassis for hot-swappable input/output modules NI 9411 – Differential Digital Input Engine crank encoder input – used for triggering engine synchronous measurements and TDC location NI 9205 – Analog Input Module Slow speed analog voltage input – used for general test-cell and engine auxiliary sensors NI 9213 – Thermocouple Input Thermocouple input – used for general test-cell and engine temperature measurements NI 9215 – Simultaneous Analog Input High-speed (crank-angle resolved) analog voltage input used for engine intake and in-cylinder pressure measurement  The NI cDAQ chassis is a modular platform that allows for any combination of input or output modules within the cDAQ family of products.  The modules listed in Table 5 were specifically chosen with engine operation in mind and were employed in this situation to allow for acquisition and logging of engine sensors.  The cDAQ chassis is connected via Ethernet to the engine computer and National Instrument’s Combustion Analysis Software (CAS) was used to view real-time 57  sensor data and to perform logging functions, in addition to communicating live sensor data to the user interface PC described in §3.3.3.  3.2.2.1 Combustion Analysis Software CAS is a comprehensive combustion package that provides real-time processing of engine data including in-cylinder pressure pegging, heat release rate calculation, user-defined post-processing, and has many other advanced functions such as knock and misfire detection.  However, throughout this investigation, only the live-view and raw data-logging capabilities of CAS were employed and post-processing scripts written in MATLAB were used to perform the combustion analysis after each test.  These post-processing scripts take parameters such as mass flow rates (air, fuel), in-cylinder pressure, pyrometry probe intensities, etc., to calculate AHRR, mean effective pressure, soot temperature and concentration, and equivalence ratios, for example.  Details of these calculations are excluded except for the ICOS and pyrometry probe calculations, which are discussed in the following chapters.  The reader is directed to National Instruments Powertrain Controls systems literature for more information on CAS.  A list of signals logged by CAS is provided in Appendix A.  3.2.2.2 Externally Logged Signals CAS performs the main data logging functionality; however, the DAQ lacks additional communication ports for the AVL emissions bench and gravimetric scale.  These measurements are directly accessed by the operating PC via Ethernet and RS-232 connections and therefore must be logged separately.  Signals are logged for a 3-minute duration and averaged to reduce the effect of random fluctuations.    58  3.2.2.3 Data Acquisition Procedure The data acquisition is generally performed at 100Hz for most of the engine and test-cell sensors.  Signals that require crank-angle resolved resolution (i.e., in-cylinder pressure, intake manifold absolute pressure, pyrometry probe photodetectors) are triggered by a crankshaft encoder with 1440 ticks per revolution equating to 0.25° crank-angle resolution.  Signals from the AVL emissions bench and gravimetric scale are recorded at 2Hz.  The ICOS probe is operated on a separate system, though it takes encoder signals from the engine for crank-angle resolved measurements.    When performing thermodynamic experiments, the engine is allowed to stabilize at each test point (i.e., stable emissions, exhaust temperature, coolant temperature, etc.) before raw data is recorded through CAS for 100 consecutive fired cycles.  Simultaneously, gravimetric scale and AVL emissions bench data (if applicable) are measured over a duration of 3 minutes to reduce random fluctuations in the signals.  Both sets of data are averaged in post-processing.  In contrast, optical tests are skip fired.  The total number of cycles to measure can vary, though limitations in the imaging equipment prevents more than 15 imaged cycles.  Generally, the engine is brought up to a nominal speed (e.g., 600rpm) before a test sequence starts.  A sequence is usually 17 motored cycles followed by 3 fired cycles with only the 3rd cycle being imaged.  This sequence is repeated 15 times, for a total sequence length of 300 cycles.  To ensure all the data is captured, 330 cycles are recorded by CAS.  Gravimetric scale and AVL emissions bench data are not recorded due to the short duration and transient nature of each optical run.  A schematic of the skip firing scheme is shown in §3.3.3.1.  59  3.3 Engine Control Unit (ECU) National Instrument’s Powertrain Control hardware was used to control the fuel injection system.  The control hardware is based on a field-programmable gate array (FPGA) and uses the NI cRIO-9068 (CompactRIO) chassis with selected Powertrain Control modules as hardware drivers for the various sensors and actuators on the engine.  The individual modules are listed in Table 6.   Table 6 - Powertrain control modules descriptions. Module Model Description NI cRIO-9068 Chassis  8-slotted Ethernet Chassis for hot-swappable input/output modules w/ FPGA and real-time processor NI 9751 – Direct Injector Driver Module Hardware driver for diesel common rail and HPDI injectors NI 9752 – Automotive AD Combo Module Analog and digital input module for engine control feedback NI 9754 – Engine-Synchronous TTL Output Module Digital outputs for engine synchronous camera and optical sensor triggering NI 9757 – O2 Sensor Module O2 wide-band sensor controller NI 9758 – Port Fuel Injector Driver Module Hardware driver for port fuel injectors and common rail solenoids  To control the system, a custom LabVIEW user-interface was developed.  The LabVIEW project consists of three distinct virtual instruments (VIs) that communicate at different hardware levels, as discussed below.  Virtual instruments are developed within the LabVIEW graphical programming language and are designed to interact with data acquisition and hardware drivers.  In this case, the three VIs are programs that allows the user to interact with the engine control hardware.  The cRIO chassis is wired to engine sensors and actuators, and is controlled by the FPGA, real-time, and operating PC VIs.  The DAQ and ECU interaction is described by Figure 20. 60   Figure 20 - Simplified DAQ and ECU hardware and software hierarchy.  CAS – Combustion Analysis Software, DAQ – data acquisition, ECU – engine control unit, FPGA – field programmable gate array, IMV – inlet metering valve, VI – virtual instrument.  3.3.1 FPGA Virtual Instrument This VI is the lowest level in the software hierarchy and runs a hardware-timed loop (40 MHz clock on the chassis) that checks hardware sensors and actuates injectors and solenoids per higher level VI commands.  A flywheel hall-effect sensor provides engine position while a crank mounted encoder and cam phasor provides TDC and cycle specific references.  The FPGA directly controls 61  the diesel direct injectors, port fuel injectors, HPDI injector, diesel common rail solenoids, O2 wideband sensor, and camera transistor-transistor logic triggers.  The subVIs within this program are described in more detail in Appendix B.  3.3.2 Real-Time Virtual Instrument The real-time VI acts as an intermediary between the main user interface and the FPGA.  While the real-time VI could be used as the main user-interface, sensors not connected to the cRIO chassis (i.e., fuel gravimetric scale and emissions bench analyzer) would not be accessible.  As a result, the real-time VI was developed as a barebones communication VI between the main user-interface and the FPGA.  The real-time VI runs off the onboard computer on the cRIO chassis alongside the FPGA.  3.3.3 Operating PC Virtual Instrument The main user-interface is run on the operating PC through this VI.  This VI provides the operator with live sensor information from the ECU through the real-time VI, emissions bench data and fuel scale readings through direct communication to the PC, and test-cell sensors through CAS as described in 3.2.2 Data Acquisition (DAQ) and Appendix A.  In addition to live sensor information, the main function of this VI is to provide an interface for user inputs to the ECU.  3.3.3.1 Engine Controls The VI allows the operator to change the adjustable engine operating parameters listed in Appendix C - Table C.1.  These parameters allow the user to control the overall operation of the engine by setting the injection timing, duration, fuel pressure (for DIDF common rail only), skip 62  firing sequence, and camera triggers.  Details on specific parameters and their usage can be found in Appendix C - Table C.1.  One of the more complex features of the VI is the ability for the user to program a skip-firing scheme.  Parameters for this are described in Appendix C - Table C.1 under Sequence Parameters.  Users can set a repeatable sequence anywhere from 1 – 20 cycles long with individual cycle control over all direct injectors (including split injections).  Camera triggers and PFI commands are controllable on an on-off basis for each cycle.  A standard optical test skip firing scheme is described in Figure 21.    Figure 21 - Standard optical test firing scheme.  The sequence of 17 skipped and 3 fired cycles is repeated 15 times.  In HPDI operation, the skipped cycles are used to clear the NG lines of diesel by injecting gas only very late in the cycle (150° aTDC).  Only the 3rd consecutive fired cycle is imaged and analyzed based on the results of an investigation looking at the stability of consecutive fired cycles in a skip firing scheme.    For each cycle, up to 4 direct injection events (2 diesel, 2 NG) can happen in the current software implementation (though with software modification, up to 5 each is possible).  If installed, an additional 2 port fuel injection events can occur per cycle.  For HPDI conditions, a diesel pilot and single NG injection is the normal mode of operation.  A second NG injection is usually reserved for special operating conditions (i.e., late post injection, optical gas clearing shots).  In DIDF 63  conditions, only the diesel direct injection settings are utilized, and the PFI is operated on an on-off basis per cycle.    3.3.3.2 Safety Functionality In addition to the live sensor readouts, logging (AVL emissions, gravimetric scale), and engine control operation, the VI also includes provisions for safety checks against process limits.  The VI checks specific sensors for their process values against critical limits; process values that are checked are listed in Appendix D - Table D.1.   3.4 Testing Procedure The procedures for setting up thermodynamic and optical tests are vastly different.  In the thermodynamic configuration, the engine can be motored infinitely as long as there is enough oil pressure to protect moving components.  Optical tests however, are much more involved and requires much closer attention to detail due to the limited lifetime of the grease applied to the piston liner.  Assuming the engine has been prepared for operation (i.e., coolant/oil temperature sufficiently high, fueling systems enabled, etc.), the startup procedure for thermodynamic tests begins by motoring the engine to a nominal speed before enabling the injectors.  If the combustion strength is high enough, the engine speeds up to a specified hold speed where a load is applied by the dynamometer as resistance.  Generally, injection parameters are set to a well characterized repeatability point for warmup before moving to operating points of interest.  As previously 64  mentioned, once an operating point is set (i.e., injector duration, timing, pressure, etc.), the engine is allowed to stabilize before any data is taken.    For optical tests, startup procedures are much more complicated.  Preparation for a test includes a partial engine teardown to facilitate cleaning of the optical quartz window.  In addition to this, each set of tests requires a regreasing of the optical piston liner as the piston rings are not lubricated with oil like the thermodynamic configuration.  The process of teardown and reassembly takes approximately 3 hours.  Following this, the usual startup procedures include heating the intake, coolant, and oil.  During this time the optical equipment is setup, and pre-start checklists are reviewed.  Desired injection parameters and skip-firing sequences can also be set in the software prior to the test.  These parameters cannot change during the test, and must be carefully reviewed.  When the test is ready to begin, the engine is motored to a speed of 600rpm without a hold limit.  Simultaneously, diesel/NG fuel systems are enabled and brought to pressure (this prevents diesel buildup in the gas line).  Once fuel pressure and engine speeds are at nominal values, injection is enabled and data acquisition for 330 cycles is started for both CAS and ICOS systems.  At the end of the test when injection is automatically shut off, the operator must turn off the dynamometer and disable the fuel systems.    Due to the high preparation effort required for optical tests, thermodynamic campaigns are generally performed beforehand to identify operating modes of interest.  This allows parameters such as mass flow rates, emissions, and equivalence ratios to be determined prior to an optical test where such measurements are not possible.  With careful planning and exploratory thermodynamic testing, operating points can be selected for an optical measurement campaign.    65  3.5 Experimental Facility Summary This chapter covers portions of the Proteus experimental facility buildup, and focuses on the data acquisition and engine control aspects in detail.  These systems make up a large part of the work involved in setting up the test facility but are by no means the only important aspects.  The data acquisition and engine control unit were largely developed by Jeff Yeo while other features of the experimental facility (i.e., engine auxiliary systems, fueling systems, control panel, HPDI head design, 2D camera systems) were designed and/or built by colleagues Jeremy Rochussen, Mahdiar Khosravi, and, Jeff Son.  Several images of the fully assembled experimental facility, as well as a conceptual model of the probe installations are shown below.  Figure 22 - Proteus research engine in HPDI optical configuration. 66   Figure 23 - HPDI head with in-cylinder diagnostic probes installed: in-cylinder pressure transducer, ICOS probe, and pyrometry probe. 67   Figure 24 - Schematic of probe (ICOS and pyrometry) installations and their position relative to the gas jets for HPDI injection.  View is of the fire deck from underneath the UBC cylinder head. 68  Chapter 4: LaVision Internal Combustion Optical Sensor (ICOS) The ICOS is an in-cylinder, local infrared absorption sensor used to determine the temporally resolved, local air-fuel ratio and methane fuel concentration in an engine operating under different fueling conditions.  In this study, the ICOS was applied to DIDF and HPDI fueling conditions.  This type of measurement can provide insight to in-cylinder mixing and fuel conversion processes and is useful for comparisons with traditional combustion diagnostics such as heat release rate.  The development of the ICOS is described in [42], while the relevant aspects of its design and measurement analysis are briefly summarized here before results are discussed.    4.1 Internal Combustion Optical Sensor Hardware Overview The ICOS uses IR absorption to measure the local, in-cylinder fuel concentration. A broadband lamp with an integrated chopper is used as a light source. The chopped light is guided via fiber optic cable through the probe (installed in the cylinder head) towards an in-cylinder measurement volume.  The light is attenuated by the fuel in the measurement volume and is reflected back using a mirror through a second fiber optic cable for measurement. A schematic of the probe cross section describing its optical components is provided in Figure 25 while the integration of the ICOS system with the engine facility is indicated in Figure 18.  An image of the actual probe is shown in Figure 26. 69   Figure 25 - Simplified cross section view of the absorption path of the LaVision ICOS probe.  Reprinted from [42] with permission from SAE International.   Figure 26 - LaVision ICOS probe without its mounting clamp.  70  As the ICOS provides a local measurement, its placement in the cylinder relative to the fuel injector, cylinder walls, and piston bowl will influence the measured fuel concentration.  This is particularly true for HPDI operation due to the non-premixed nature of air/fuel mixture, and the fuel concentration signal is expected to be most sensitive to the injector spray orientation and the ICOS position in the cylinder.  In contrast, DIDF operation is expected to be less sensitive to sensor location due to the homogeneous nature of the air/fuel mixture. The placement of the ICOS probe in both DIDF and HPDI configurations of this investigation is such that the diesel spray jets do not directly impinge on the sensor.  The main benefits of such a position include a lower sensitivity to diesel, as well as slower mirror contamination rates.  A discussion of the specific placement of the sensor relative to fuel jets and its effect on measurements will follow in the results section.  Figure 27 to Figure 29 show the installation of the ICOS in the DIDF configuration, while Figure 30 shows the positioning of the ICOS under HPDI conditions.  Figure 27 - DIDF configuration engine front and side view.  Injector is oriented 75° from the fire deck and the spray included angle is 142°.    71   Figure 28 - Plan view from piston showing diesel spray pattern relative to the ICOS probe in DIDF configuration.  Freeze-frame from an externally illuminated high-speed NL video in the optical configuration.  Figure 29 - DIDF head with in-cylinder diagnostic probes installed: in-cylinder pressure transducer, ICOS probe, and pyrometry probe.  72   Figure 30 - Plan view from piston showing diesel and NG spray pattern relative to the ICOS probe in HPDI configuration.  4.2 Operating Theory of the Internal Combustion Optical Sensor The recorded voltage from the ICOS is an intensity measurement of the attenuated and reflected light (absorption signal) and is optically filtered at 3.4 μm (290nm full-width at half maximum).  The wavelength is chosen for its sensitivity to hydrocarbon absorption lines [51], [52].  As a result, the sensor is sensitive to both diesel and methane; however, for both HPDI and DIDF configurations there is a low sensitivity to diesel due to sensor placement and is discussed in detail later.  Figure 31 shows a sample measurement of the transmitted light intensity (i.e., spectral transmitted light intensity at 3.4μm after the light has passed through the measurement volume twice, see Figure 25) for premixed methane during intake, compression, combustion, expansion 73  and exhaust (DIDF configuration).  The intensity fluctuates during intake as the methane mixes with the air.  The subsequent drop in intensity is due to the increasing fuel concentration during compression, followed by a sharp rise in intensity (concentration reduction) because of methane consumption by combustion.  During expansion and exhaust, the transmitted intensity is at a maximum (the expected fuel concentration is a minimum relative to the other parts of the combustion cycle).  It should be noted that high transmitted intensity in the expansion and exhaust strokes does not necessarily mean all the fuel has been combusted – it could remain as partial combustion products that do not absorb at 3.4μm (i.e., CO).  Figure 31 - Sample transmitted light intensity to indicate the influence of cycle stages (intake, compression, combustion, exhaust) on the measured intensity for DIDF configuration.  The transmitted intensity is measured after the light has passed through the measurement volume twice.  74  The local fuel concentration (Cfuel) is related to the transmitted light intensity signal using the Lambert-Beer law and can be expressed as a qualitative local fuel concentration:   ln (𝐼(𝜃)𝐼0) = −𝜎 ∙ 𝐶𝑓𝑢𝑒𝑙(𝜃) ∙ 𝐿 ⟹ ln (𝐼(𝜃)𝐼0) ∝ −𝐶𝑓𝑢𝑒𝑙(𝜃) (10) where I is the measured, transmitted light intensity at each crank angle and I0 is the average transmitted light intensity within a specified crank angle range (averaged between 250° to 300° aTDC (DIDF) and between -200° and -180° aTDC (HPDI) where expected fuel concentration is minimum for the cycle).  I0 is determined independently for each cycle and absorption strength σ and optical path length L are assumed constant and arbitrarily set to a value of 1.0.  The result of these assumptions is a qualitative measurement of the local fuel concentration.  A quantitative fuel concentration measurement requires consideration of the absorption spectrum of the fuel (i.e., absorption strength σ vs. wavelength with respect to temperature, pressure) and will be the subject of a future investigation.    Along with the fuel concentration (Cfuel) measurements, air-fuel ratios can be calculated with additional parameters.  Information regarding air mass flow rates is required to specify the local air/fuel mixture.  A mass air-flow sensor is used to estimate the in-cylinder mass of air per cycle mair. The measured mass flow rate is converted to mass per cycle based on engine RPM, though the residual gas fraction was not considered.  Finally, eq. (11) is used in combination with the results from eq. (10) to calculate the local AFR at each crank angle.  Again, this is referred to as a qualitative measurement as the absorption strength is assumed to be constant in the fuel concentration calculation.  75   𝐴𝐹𝑙𝑜𝑐𝑎𝑙(𝜃) ∝𝑚𝑎𝑖𝑟𝐶𝑓𝑢𝑒𝑙(𝜃) ∙ 𝑉(𝜃) (11) In eq. (11), V(θ) is the cylinder volume at each crank angle.  Further details regarding ICOS operating theory can be found in [42] where information about chopper speeds, lamp specifications, hardware requirements, are also provided.  4.3 ICOS DIDF Experimental Results and Discussion This section documents the results and discussion from initial ICOS experiments. These experiments were performed on the DIDF thermodynamic engine configuration utilizing the original head described in §3.1 and served as the basis for the development of the analysis tools used in later optical experiments.  4.3.1 ICOS DIDF Experiment Design Operating modes with significantly different fuel conversion mechanisms are selected here for further examination based on past DIDF studies at the same facility [11], [16], [26].  Specifically, some of these low pilot ratio operating modes were observed to have different trends in flame propagation (Figure 9 and Figure 10) due to variations in diesel pilot injection pressure and a discussion of these trends can be found in §2.1.2 and in [16], [26].  The selected operating points may showcase the combustion behaviour exhibited in Figure 10 and the ICOS is used to evaluate whether the instrument can provide insight to in-cylinder combustion modes without the use of the full optical configuration.  Crucial parameters (eq. (1) to (4)) were defined in §2.1.1.  76  A description of the DIDF operating space was proposed by Rochussen in [16] which plots global equivalence ratio ϕglobal versus pilot ratio Rpilot.  The plot also consists of isolines of constant ϕCH4 and constant pilot injection mass.  The ICOS experimental test matrix is subsequently plotted in Figure 32, and shows the nominal operating points as well as the as tested operating points.  The indicated operating points represent a variation in the methane injection mass for two constant diesel injection rates and each operating point is tested at 300 and 1300 bar pilot injection pressure.  Additionally, a timing sweep of port injected CH4 was performed at a single operating point (point 3) though these points were excluded from Figure 32 for clarity.  Full details of each operating point measured are provided in Table 7.  77  Table 7 - Detailed experimental conditions in the ICOS DIDF measurement campaign. Point Label φCH4 φglobal Rpilot Ppilot [bar] Diesel Injection Timing [CAD aTDC] Diesel Injection Duration [ms] Methane Injection Timing [CAD aTDC] Methane Injection Duration [ms] GIMEP [bar] DIDF test matrix for 2 constant diesel flow rates, 2 diesel injection pressures, and varying φCH4 1-1300 0.72 0.76 0.05 1300 -8 0.34 -330 24.2 5.91 2-1300 0.64 0.72 0.10 1300 -8 0.34 -330 21.9 5.82 3-1300 0.57 0.64 0.11 1300 -8 0.34 -330 19.6 5.57 4-1300 0.50 0.56 0.11 1300 -8 0.34 -330 17.3 5.08 5-1300 0.43 0.51 0.15 1300 -8 0.34 -330 15.0 4.38 6-1300 0.75 0.93 0.20 1300 -8 0.46 -330 24.2 5.85 7-1300 0.67 0.87 0.24 1300 -8 0.46 -330 21.9 6.17 8-1300 0.59 0.75 0.22 1300 -8 0.46 -330 19.6 5.96 9-1300 0.51 0.68 0.25 1300 -8 0.46 -330 17.3 5.54 10-1300 0.44 0.58 0.25 1300 -8 0.46 -330 15.0 5.00 1-300 0.72 0.76 0.06 300 -8 0.80 -330 24.2 6.82 2-300 0.64 0.72 0.12 300 -8 0.80 -330 21.9 6.40 3-300 0.56 0.62 0.10 300 -8 0.80 -330 19.6 5.85 4-300 0.49 0.56 0.12 300 -8 0.80 -330 17.3 5.20 5-300 0.42 0.50 0.17 300 -8 0.80 -330 15.0 3.80 6-300 0.73 0.91 0.21 300 -8 1.20 -330 24.2 7.40 7-300 0.65 0.81 0.20 300 -8 1.20 -330 21.9 7.19 8-300 0.58 0.71 0.20 300 -8 1.20 -330 19.6 6.71 9-300 0.50 0.66 0.24 300 -8 1.20 -330 17.3 6.11 10-300 0.43 0.60 0.29 300 -8 1.20 -330 15.0 5.40 DIDF test matrix for CSOI timing sweep.  3-1300 and 3-300 from the DF test matrix are used to supplement these test points for the -330° aTDC CSOI case. 3-1300-280 0.56 0.61 0.08 1300 -8 0.34 -280 19.6 5.65 3-1300-230 0.56 0.62 0.10 1300 -8 0.34 -230 19.6 5.59 3-1300-170 0.57 0.61 0.06 1300 -8 0.34 -170 19.6 5.54 3-300-280 0.56 0.67 0.16 300 -8 0.80 -280 19.6 5.89 3-300-230 0.56 0.67 0.16 300 -8 0.80 -230 19.6 5.87 3-300-170 0.56 0.59 0.04 300 -8 0.80 -170 19.6 5.88  78   Figure 32 - Considered operating points on global equivalence ratio – pilot ratio space.  Points 1-5 and 6-10 as labelled in the figure correspond to nominal constant diesel mass flow rates of 0.12 kg/hr and 0.38 kg/hr respectively.  The variability in actual Rpilot values is attributed to diesel mass flow rate uncertainty, particularly at low injection masses.  Nominal points refer to the set of ideal parameters (Rpilot and ϕglobal) that were targeted.  4.3.2 Effect of Diesel Fuel on Fuel Concentration Measurements (DIDF) The IR absorption probe depends on the vibrational-rotational absorption band of hydrocarbons around 3.4μm [42], [51], [52] and is therefore sensitive to both methane and diesel.  However, as seen in Figure 28, the ICOS sensor is situated between two diesel jet and as a result, low sensitivity to diesel is expected in this configuration.  Nevertheless, diesel effects on local fuel concentration and local AFR measurements may not be negligible and is subsequently evaluated by performing diesel-only testing with the ICOS.  Diesel-only tests (i.e. without any premixed methane) were 0 0.1 0.2 0.3 0.4 0.50.40.50.60.70.80.91Rpilotglobal   CH4 = 0.62 CH4 = 0.49 CH4 = 0.36 CH4 = 0.2312345678910Nominal pts.1-5 300bar1-5 1300bar6-10 300bar6-10 1300barmd = 0.12kg/hrmd = 0.38kg/hr79  performed for each set of pilot parameters as in the DIDF test matrix.  Figure 33 shows the diesel-only fuel concentration histories for the two considered injection quantities and pressures.  Figure 33 - Local fuel concentration measurements for diesel-only testing.  The legend corresponds to points identified in Figure 32.  CSOI timing is the same for each point and end of injection (EOI) timings are shown by their corresponding colours.  Test points are labeled by the injection pressure followed by the points for which the diesel pilot mass corresponds to.  Point 1300 (6-10) is not shown due to contamination of sensor system.  The fuel concentration signal does not change significantly with the introduction of diesel fuel into the combustion chamber.  If diesel fuel droplets were to interfere with the signal, increases in the signal are expected near CSOI (-8° aTDC and after).  The slight increase in fuel concentration around TDC is attributed to combustion noise and soot.  In comparison, a nominal DIDF test point (i.e., Point 3-1300, Figure 40) has peak methane fuel concentration signals about an order of magnitude higher than concentrations noted for the diesel-only tests (0.4 for DIDF compared to -10 -5 0 5 10 15 20-0.1-0.0500.050.10.150.2CAD aTDCLocal Fuel Concentration [a.u.]  1300 (1-5)300 (1-5)300 (6-10)CSOIEOI80  0.03 for diesel).  For the ICOS installation location, pilot injection mass, and pilot injection pressures considered in this study, diesel is not a significant contributor to the local fuel concentration and local AFR measurements. As such, the resulting measurements are interpreted as indicating only methane concentration.    4.3.3 Local Air Fuel Ratio Comparison with Mass Based Air Fuel Ratio Early port injection of methane in DIDF creates a homogeneous air-fuel mixture in the cylinder.  As a result, the probe’s ability to resolve in-cylinder methane fuel concentration can be evaluated through comparison with global measurements for air and methane mass flow rates.  The methane mass flow rate was measured with an in-line Coriolis flow meter and coupled with a mass air-flow sensor to calculate CH4 equivalence ratio ϕCH4.  For comparison, ϕCH4 is converted to global AFR using eq. (4).  Prior to combustion, the AFR is expected to remain constant after intake valve closing and during combustion. However, the local AFR plotted in Figure 34 are calculated from ICOS absorption data and shows an upward trend throughout compression and especially so from -20° aTDC onwards for all test cases.  This may be attributed to several different factors and is discussed later in §4.3.4.  As a result, a representative end value of local AFR was determined for each operating point as the average of 100 cycles at -5° aTDC (selected prior to TDC to avoid combustion interference).  The in-cylinder local AFR determined in this fashion is compared to the global AFR determined from fuel and air mass flow rates in Figure 35.   81      Figure 34 - Unadjusted local AFR measurement trends during compression prior to combustion.  The legends correspond to points identified in Figure 32.  Shaded regions indicate +/- one standard deviation over 100 cycles.  The dashed line at -20° aTDC represents the approximate inflection point in the local AFR results.  Here, unadjusted describes the data as being raw (i.e., directly calculated from eq. (11)); later discussion presents a sensitivity analysis of the local AFR by modifying the raw data. 82   Figure 35 - Comparison of local and global AFR based on measured fuel and air mass flow rates for all measured points.  This relationship is shown to be linear and can be used as a calibration curve to determine the true local AFR in other experimental configurations where the in-cylinder mixtures may be stratified.  The local AFR is noted to be linearly proportional to the global AFR in Figure 35.  While there may be some non-linearity due to temperature and pressure effects in the IR absorption method, for the points measured, there is good agreement between the two.  Consequently, the local AFR as measured by the ICOS can be used as an indication of the in-cylinder local AFR when used in conjunction with global measurements as a calibration.  This calibration assumes that the in-cylinder mixture is homogeneous, though there may be some non-homogeneity effects due to residuals and non-perfect mixing.    20 25 30 35 40 45303540455055Global AFRLocal AFR [a.u.]  AFRlocal = 0.972 * AFRglobal + 9.69Experimental DataR2 =0.9683  4.3.4 Uncertainty in Local AFR Measurements The trend seen in Figure 34 is that of increasing local AFR during compression for all test cases and in particular, a significant rise after -20° aTDC.  Given that port injection occurs far upstream from the intake valve and is assumed homogeneous by the time the intake charge is inducted into the cylinder, it is expected that the local AFR should be stable (i.e., steady state) throughout compression rather than increasing. Local AFR (eq. (11)) determination requires measurements or estimates of in-cylinder air mass, cylinder volume, and I0 (reference intensity).  In addition, the temperature and pressure effects on fuel absorption strength due to the stretching vibrations of C-H bonds should be considered.  Doppler broadening effects on absorption lines can largely be ignored for gasoline and single hydrocarbon species during compression [42].  For this discussion, the sensitivity of the local AFR with respect to cylinder volume V, reference intensity I0, and air mass mair is assessed.    4.3.4.1 Volume Errors Volume errors are evaluated on the basis that the parameters used in determining the crank angle resolved volume vector are not accurately known.  The parameters required to calculate the volume at each crank angle are: engine bore, stroke, connecting rod length, crank radius, displacement volume, and clearance volume.  These parameters are set based on the engine specifications, and have not been confirmed except for the clearance volume.  In addition, encoder position errors can cause phasing errors between the calculated and true cylinder volumes.  The sensitivity of the local AFR to cylinder volume errors was assessed by introducing artificial crank-angle phase shifts (representing encoder position errors), and by reducing the compression 84  ratio (CR) used to calculate the volume (representing errors in the clearance volume).  From these artificial perturbations, it was found that crank-angle phasing only resulted in minor changes to the trends in local AFR and were not substantial enough to correct the rising trends.  CR reduction, however, resulted in more notable changes in the local AFR trends.  Figure 36 shows the local AFR for an artificial reduction in compression ratio from 14.25 to 13.25:1 (estimated 15 cm3 additional clearance volume from original 150 cm3 due to head modifications to fit various probes).  While a significant change, it shows the sensitivity of AFR trends to volume errors; the rise in local AFR prior to combustion is nearly non-existent up until -20° aTDC when compared to the unadjusted AFR in  Figure 34.  The cause of the increase in local AFR after -20° aTDC is unknown at this time and will require further investigation.  Under consideration of this CR adjustment, the local AFR is approximately constant during the compression process, as would be expected for a homogeneous mixture.  The increase in the local AFR after -20° aTDC is expected to be due to some other physical effect such as non-homogeneities or blow-by.  No attempt is made to correct for these artifacts as they are not quantified in the current study.  Further to this, the AFR has a considerable dependence on I0 selection as discussed later.  85      Figure 36 - Compression ratio adjusted (decrease from 14.25:1 to 13.25:1) local AFR measurement trends during compression prior to combustion.  The legends correspond to points identified in Figure 32.  Shaded regions indicate +/- one standard deviation over 100 cycles.  The dashed line at -20° aTDC represents the approximate inflection point in the local AFR results.  AFR remains constant during the majority of compression compared to unadjusted results (Figure 34).  This serves as an indication of the measurement sensitivity to errors in volume calculations.  4.3.4.2 I0 Error The local AFR measurements are also sensitive to the selection of I0, or the reference intensity representing zero attenuation to the signal.  An artificial perturbation to I0 is introduced by setting 86  the reference intensity, calculated as an average of the measured intensity between 250° aTDC to 300° aTDC for each cycle, slightly higher.  This is equivalent to assuming that the calculated I0 has been somewhat attenuated by residual gases or particulates.  Figure 37 shows the effect of a 1% increase to I0.  The adjustment represents the maximum difference between the calculated I0 and the peak cycle intensity (i.e., least attenuated signal) for a given operating point.  Figure 31 provides an example of the intensity measurement for a single cycle – the peak cycle intensity is found shortly after combustion and there is a period of stability in the intensity during the exhaust stroke where fuel concentration is expected to be close to zero.  The sensitivity of the measurement to I0 shows the importance of properly selecting a range at which fuel concentration is minimal and can be difficult in a real engine environment given the effects of residual mass and uHCs.   87      Figure 37 - I0 adjusted (1% increase) local AFR measurement trends during compression prior to combustion.  The legends correspond to points identified in Figure 32.  Shaded regions indicate +/- one standard deviation over 100 cycles.  The dashed line at -20° aTDC represents the approximate inflection point in the local AFR results.  The end values of AFR is higher in magnitude and exhibits significant rise in AFR prior to combustion compared to the as-measured case (Figure 34) and CR adjusted case (Figure 36).  4.3.4.3 Air Mass Errors The local AFR measurement requires an estimate of the air mass, and can be susceptible to errors in the mass flow measurement or blow-by effects.  A simulation of the blow-by effect is achieved by gradually reducing the air mass during part of the compression stroke as shown in eq. (12) and 88  (13).  This represents a gradual 10% linear reduction in the cylinder air mass between -20° to 0° aTDC.  𝑚𝑎𝑖𝑟,𝑎𝑑𝑗𝑢𝑠𝑡𝑒𝑑(𝜃) == {𝑚𝑎𝑖𝑟 [1 −𝜃 − 𝜃1𝜃2 − 𝜃1∗ 0.1] , −20° ≤ 𝜃 ≤ 0° 𝑎𝑇𝐷𝐶𝑚𝑎𝑖𝑟 , 𝜃 < −20° 𝑎𝑇𝐷𝐶 (12)  𝜃1 = −20° 𝑎𝑇𝐷𝐶, 𝜃2 = 0° 𝑎𝑇𝐷𝐶 (13) Applying these two equations to the air mass measurement decreases AFR during the late stages of compression where fuel mass is possibly being lost, resulting in a more stable response in the local AFR measurements.  Figure 38 shows the result of applying eq. (12) (adjusted air mass) to eq. (11).  89      Figure 38 - mair adjusted (10% linear decrease from -20° to 0° aTDC) local AFR measurement trends during compression prior to combustion.  The legends correspond to points identified in Figure 32.  Shaded regions indicate +/- one standard deviation over 100 cycles.  The dashed line at -20° aTDC represents the approximate inflection point in the local AFR results.  This artificial consideration of blow-by reduces the rising trend from -20° aTDC to combustion.   From Figure 38, the significant rise in AFR following -20° aTDC has been reduced and the trend is approximately “flat” throughout compression for most test cases.  While this adjustment does not include a corresponding adjustment to the fuel concentration, it aims to show the potential effect blow-by may have on the local AFR measurement.  90  Another way to check if blow-by is an issue is to calculate the methane fuel mass mfuel using eq. (14).  This equation scales the local fuel concentration by the cylinder volume and the results are shown in Figure 39.  𝑚𝑓𝑢𝑒𝑙(𝜃) = 𝐶𝑓𝑢𝑒𝑙(𝜃) ∗ 𝑉(𝜃) (14)  Figure 39 - Sample in-cylinder methane fuel mass mfuel calculated using eq. (14).  There is a noticeable decrease in the fuel mass starting approximately at -20° aTDC indicating possible blow-by phenomenon.    In Figure 39, the same inflection point at -20° aTDC is observed.  The fuel mass decreases through most of the compression stroke and the inflection point at -20° aTDC indicates an increase in the rate of mass loss and can imply blow-by.  At around -3° aTDC, combustion reaches the sensor location and the fuel mass decreases to 0.  91  The combination of volume, I0 selection, and cylinder air mass shows the importance of parameter determination for accurate AFR measurements. Perturbations in these parameters were used as a sensitivity analysis for local AFR.  It is expected that AFR remains constant throughout compression and that some combination of uncertainties in these variables lead to the rising trends seen in the unadjusted local AFR results (Figure 34).  Other effects may also play a role, including in-cylinder mixture non-homogeneity, residual gases, and non-linearity of the technique (i.e., fuel concentration vs. I/I0, as well as pressure and temperature effects [42]).  However, these effects are not addressed in this study and would require additional investigation.  Regardless, the local AFR measurements in this study are considered qualitative and the main results discussed herein are for the unadjusted data.  4.3.5 Evaluation of Combustion Mode Based on Fuel Concentration History Previous optical DIDF investigations [11], [26] have shown that for low pilot ratios, low and high pressure diesel injections exhibited different fuel conversion characteristics.  High diesel pilot injection pressures resulted in ignition starting at the outside bowl periphery and subsequent flame propagation towards the center of the cylinder.  As methane concentration decreased, the flame propagation (characterized by OH*) was no longer sustained all the way to the center of cylinder due to the methane lean flammability limit.  On the other hand, low diesel pilot injection pressures resulted in ignition sites around the middle of the bowl area similar in shape to the diesel fuel jets.  The reaction zone in this case grew outwards.  Figure 10 provides a conceptual image of these phenomena.  While this insight is important for understanding pollution formation mechanisms, it requires an optically accessible engine.  Here, it will be explored if the ICOS can be used to infer the combustion mode without optical access.   92  Due to the ICOS probe’s installation location in the central region of the combustion chamber (shown in Figure 27 and Figure 28), it is expected that the local fuel concentration may be used to provide insight into the different fuel conversion trends without the use of an optical engine.  For high pilot injection pressure points, delayed fuel concentration reduction (with respect to pilot ignition characterized by AHRR) is expected in the central region of the combustion chamber, as the reaction initiates away from the probe.  Subsequently the local fuel concentration will decrease as the flame approaches the center of the cylinder, where the sensor is located.  Conversely, for low pilot injection pressures, rapid consumption of the fuel (near-simultaneous to pilot ignition) is expected near the centrally located probe, as this is where the reaction initiates.   Figure 40 - Local fuel concentration with AHRR.  Points 1-5, 1300 bar.  A) Initial increase in fuel concentration due to diesel auto-ignition away from the ICOS sensor.  B) Peaks and discontinuities in the fuel concentration signal following the main combustion event are attributed to saturation in the absorption intensity signal.  C) Slow decay in the fuel concentration may be an indication of incomplete combustion at the location of the ICOS sensor.  0200400HRR [J/CAD]  1-13002-13003-13004-13005-1300-5 0 5 10 15 2000.20.4CAD aTDCLocal FuelConcentration [a.u.]A B C 93   Figure 41 - Local fuel concentration with AHRR.  Points 1-5, 300 bar.  Figure 42 - Local fuel concentration with AHRR.  Points 6-10, 1300 bar.  A) Initial increase in fuel concentration due to diesel auto-ignition away from the ICOS sensor.  B) Peaks and discontinuities in the fuel concentration signal following the main combustion event are attributed to saturation in the absorption intensity signal.  0200400HRR [J/CAD]  1-3002-3003-3004-3005-300-5 0 5 10 15 2000.20.4CAD aTDCLocal FuelConcentration [a.u.]05001000HRR [J/CAD]  6-13007-13008-13009-130010-1300-5 0 5 10 15 2000.20.4CAD aTDCLocal FuelConcentration [a.u.]A B 94   Figure 43 - Local fuel concentration with AHRR.  Points 6-10, 300 bar.  Figure 40 through Figure 43 compare the local fuel concentrations and AHRR for varying CH4 equivalence ratios and pilot injection pressures.  For all test cases, the fuel concentration decreases as expected near TDC due to consumption of methane during combustion (methane consumption causes an increase in measured intensity; see Figure 31 for intensity trends within a cycle).  However, the different pilot pressures result in differences in the local fuel concentration relative to the heat release rate.  To better examine these behaviours, 4 test points are taken for comparison in Figure 44 and Figure 45 at two pressures (300bar and 1300bar) and two pilot ratios (Point 3 and Point 8, Figure 32).  To facilitate this comparison, the fuel concentration signal is normalized to a representative value at -5° aTDC and is averaged across 100 consecutive cycles. 0200400HRR [J/CAD]  6-3007-3008-3009-30010-300-5 0 5 10 15 2000.20.4CAD aTDCLocal FuelConcentration [a.u.]95    Figure 44 - AHRR and normalized fuel concentration.  Point 3, 300 and 1300 bar.  Vertical lines in the top and bottom plots represent the CA5 position and CA for 50% reduction in methane concentration, respectively.  Concentrations are normalized to respective values at -5 CAD aTDC.  Point 3 nominal Rpilot = 0.093, ϕglobal = 0.539.  96   Figure 45 - AHRR and normalized fuel concentration.  Point 8, 300 and 1300 bar.  Vertical lines in the top and bottom plots represent the CA5 position and CA for 50% reduction in methane concentration, respectively.  Concentrations are normalized to respective values at -5 CAD aTDC.  Point 8 nominal Rpilot = 0.234, ϕglobal = 0.702.  The 1300 bar test points (Figure 40, Figure 42, and detailed in Figure 44 and Figure 45) show a small peak in the fuel concentration signal prior to its decay.  This aligns with previous results from optical investigations [11], [26] where high pilot diesel injection pressures resulted in ignition at the periphery of the bowl with combustion moving towards the center of the bowl.  The fuel concentration peaks correspond to the beginning of heat release (auto-ignition of diesel pilot fuel) and may indicate a sharp rise in pressure (and therefore concentration) due to combustion elsewhere in the combustion chamber.  In this case, the pressure is increased, while the fuel near the sensor is not consumed.  This hypothesis is examined in detail in §4.3.5.2.  97  Point 5-1300 (Figure 40, black line) exhibits the extreme case of low methane and diesel mass and results in very slow fuel concentration decay, presumably due to the very lean mixture.  Past optical studies [10], [11], [26] have shown that at low methane mass and pilot ratios similar to Point 5-1300, combustion does not always fully reach the central region, leaving an “unburned” area.  It is possible that the extended duration in methane decay rates for these two points correspond to this “unburned” region.     For lower injection pressures (300 bar, Figure 41, Figure 43, and detailed in Figure 44 and Figure 45), all test points indicated similar trends of methane consumption through a rapid drop in concentration, approximately coincident with the start of heat release.  For these low pilot injection pressures, no fuel concentration peak is observed prior to fuel concentration decay, as in the 1300 bar cases.  This agrees with the expectation that ignition sites around the central region of the bowl will consume the methane near the probe measurement volume early in the combustion event.  The AHRR continues well beyond the end of local fuel consumption implying that combustion is sustained elsewhere in the chamber corresponding to the expected outward flame propagation.  4.3.5.1 Methane Comsumption Delay Relative to 5% HRR To further evaluate the presence of the two combustion modes shown in Figure 10, a new metric is introduced as CAdelay, or consumption delay.  This is defined by eq. (15) and represents the change in crank position from the start of combustion (defined as 5% cumulative heat release) to the crank position when a significant decrease in methane concentration is observed (50% of the initial value).   98   𝐶𝐴𝑑𝑒𝑙𝑎𝑦 = 𝐶𝐴𝜌,0.5 − 𝐶𝐴𝐻𝑅𝑅,0.05 [𝐶𝐴𝐷 𝑎𝑇𝐷𝐶] (15) In (15), CAρ,0.5 is the crank angle at which the methane concentration is observed to be 50% of the methane concentration measured at -5° aTDC.  CAHRR,0.05 is the 5% cumulative heat release crank angle.  Figure 44 and Figure 45 are representative plots of both 5% cumulative heat release and 50% methane consumption (represented by the dashed and dotted lines respectively).  It is expected that for the high injection pressure cases where the flame is anticipated to propagate inwards from the periphery, the consumption delay is more sensitive to AFR due to changing flame speeds [53].  For lower injection pressures, due to the proximity of the ignition sites to the centrally located ICOS sensor, CAdelay is expected to be less sensitive to AFR, and subsequently, flame speed.    Figure 46 - Methane consumption delay vs. global AFR for all considered operating points.  Point 5-1300 is an extreme case of low methane and low diesel injection mass; this outlying point (top right) exhibits extremely high methane consumption delay (possibly due to incomplete combustion at the ICOS sensor).  20 25 30 35 40 45024681012Global AFRCAdelay [CAD aTDC]  1-5 (1300 bar)1-5 (300 bar)6-10 (1300 bar)6-10 (300 bar)99  The methane consumption delay CAdelay is plotted against global AFR in Figure 46.  Operation with a low diesel pilot ratio, and high injection pressure (Points 1-5, 1300 bar, red circles), show an increase in CAdelay with increasing AFR.  This is due to leaner mixtures resulting lower flame speeds and a longer flame transit time from the bowl periphery to the ICOS location near the center of the bowl.  The lower injection pressures for the same points show a smaller sensitivity of CAdelay to AFR.  However, points 6-10 (higher diesel injection mass) do not show the same trend between the two pressures; these points were not considered in the optical investigations and the fuel conversion may not necessarily take place through flame propagation [26].    4.3.5.2 Fuel Concentration Ratio Comparison with Pressure Ratio An ideal gas model is used here to verify that the increase in fuel concentration prior to combustion for the 1300 bar cases (as seen in Figure 40 and Figure 42) is due to the pressure rise from combustion.  Equation (16) gives the ideal gas relation for pressure ratio (P2/P1) and fuel concentration ratio (C2/C1) assuming an isentropic compression process (hypothesized to be the process experienced by the central gases as combustion burns around the periphery).  κ is the heat capacity ratio of the gas, which ranged between 1.30 and 1.32.  Details on its calculation is shown in Appendix E.  𝐶2𝐶1= (𝑃2𝑃1)1/𝜅 (16) State 1 is taken at -5° aTDC (nominal value of fuel concentration) and state 2 is taken at the peak fuel concentration crank angle (CA).  Equation (16) is used to predict the fuel concentration ratio 100  using pressure data and is compared against actual concentration ratio (as measured by the ICOS) in Figure 47.    Figure 47 - Comparison of fuel concentration ratio and pressure ratio calculated at peak fuel concentration crank angle and -5° aTDC for 1300 bar test cases.  Error bars show the standard deviation in fuel concentration ratio except for test points exhibiting saturation in the signals.  Solid line is calculated from eq. (16).  Figure 47 shows the comparison of fuel concentration to pressure ratio and indicates that the ideal gas model agrees with the experimental trends, with some exceptions.  In particular, points 6-10 at 1300 bar (with higher Rpilot, diesel injection mass) do not match the model as well.  This may correspond to a different combustion mode in comparison to the lower pilot ratios (points 1-5 at 1300 bar).  In addition, it was noted that this analysis is very sensitive to the crank angle phasing between the ICOS and in-cylinder pressure data, as well as the noise in these respective signals.  101  4.3.6 Effect of Methane Port Injection Timing on Local Fuel Concentration A port methane injection timing sweep (commanded start of injection, CSOI) was carried out at point 3 (see Figure 32) to evaluate the effect of methane mixing time on combustion.  Point 3 was selected as it is well characterized in previous optical investigations.  In the experimental setup, the port injector is located in the intake runner approximately 60 cm away from the intake valve.  Methane CSOI varied from -330° aTDC to -170° aTDC (the latter timing coincides with intake valve closing, IVC); test details are in Table 7.  The AHRR and fuel density histories are shown for both injection pressures in Figure 48 and Figure 49.  Methane injection timings and durations are shown in Figure 50, relative to the intake valve lift.    Figure 48 - Local fuel concentration with AHRR histories for varying commanded start of port methane fuel injection.  Point 3, 1300 bar.  CSOI is indicated in CAD aTDC. 0200400HRR [J/CAD]-5 0 5 10 1500.20.4CAD aTDCFuel Density [a.u.]  -330° CSOI, 1300 bar-280° CSOI, 1300 bar-230° CSOI, 1300 bar-170° CSOI, 1300 bar102   Figure 49 - Local fuel concentration with AHRR histories for varying commanded start of port methane fuel injection.  Point 3, 300 bar.  CSOI is indicated in CAD aTDC.  Figure 50 - Port injection timing relative to intake valve lift for CSOI variation.  Shaded regions denote the port injection durations for the 4 CSOI timings (-330° [red], -280° [green], -230° [blue], -170° [yellow], aTDC).  There is overlap between the injection timings shown. 0200400HRR [J/CAD]-5 0 5 10 1500.20.4CAD aTDCFuel Density [a.u.]  -330° CSOI, 300 bar-280° CSOI, 300 bar-230° CSOI, 300 bar-170° CSOI, 300 bar103  Changing methane injection timing does not appear to have a significant effect on heat release rates for this operating point.  However, the later injection timings (-230° and -170° aTDC) result in a higher local fuel concentration (lower AFR) prior to the start of combustion compared to the earlier timings (-330° and -280° aTDC).  The later injection timings may be too close to intake valve closing (IVC, -150° aTDC) that the injected methane has a full cycle to mix in the intake manifold prior to induction into the cylinder.   Figure 51 - Comparison of local and global AFRs for methane CSOI timing sweep.  Early CSOI corresponds to -330° and -280° aTDC, late CSOI corresponds to -230° and -170° aTDC.  Figure 51 compares the local AFR determined using the ICOS with the global AFR (from mass flow measurements), for each port injection timing sweep.  The global AFR does not show any significant differences for the different injection timings, whereas the local, in-cylinder AFR 104  changes between the early and late injection timings.  This implies that the local fuel concentration may be the influenced by the injection timing.    4.3.7 ICOS DIDF Testing Summary The goal of the initial ICOS DIDF experiments was to evaluate the in-cylinder, local infrared absorption sensor for resolving the local AFR and fuel concentration.  This was performed using a single cylinder DIDF engine operating in a thermodynamic configuration.  Diesel-only testing was performed to evaluate the effect of diesel on ICOS measurements and was found to be negligible for the tested configuration, as the injected fuel was not seen to reach the measurement volume.  A methane timing sweep was also performed at a single point and indicated that for the engine and fuel injection configuration considered, there is a sensitivity of the local AFR to port injection timing.  For this timing sweep, there were no significant effects on other macroscopic parameters, such as the global AFR or methane emissions.  The local in-cylinder AFR measured using the ICOS agreed with the global AFR (based on mass flow rates of air and methane) at each considered operating point.  The measured fuel concentration and associated local AFR were noted to be sensitive to various experimental uncertainties, particularly the reference signal intensity (I0), the actual cylinder content’s mass, and the cylinder volume.  The measured fuel concentration histories agreed with observations made in previous optical studies [26], [54].  Operating points noted to have significantly different fuel conversion modes were replicated in the present study and the resulting local fuel concentration measurements matched well with expected fuel conversion behaviour.  In particular, operating points where the flame propagates from the bowl perimeter to the piston center (low pilot ratio, high injection 105  pressure) resulted in a delay in the local methane consumption relative to the global heat release, while operating points showcasing flame propagation from the central position (low pilot ratio, low injection pressure) did not exhibit the same delay.  4.4 ICOS HPDI Experimental Results and Discussion The knowledge gained from running the ICOS under DIDF operating conditions provides insight on one type of combustion strategy using port fuel injectors.  DIDF creates a homogeneous mixture of NG inside the combustion chamber, which allows the ICOS to provide an estimate of the global, premixed equivalence ratio.  In contrast, HPDI operation creates significant fuel-air stratification due to the direct injection of NG. To better understand how the ICOS operates under highly stratified conditions, a series of ICOS HPDI experiments were carried out simultaneous to other optical diagnostics including the pyrometry probe and spatially resolved NL imaging to determine its suitability for direct injection NG experiments.    4.4.1 ICOS HPDI Experiment Design An HPDI test point with a diesel pilot and NG main injection formed the baseline from which several injection parameters were varied, based on the work of Ehsan Faghani [55]. There, injection strategies and their effect on PM emissions for an HPDI thermodynamic single-cylinder research engine were evaluated.  From the baseline point, other test points varied by: increasing gas pulse width, decreasing pulse separation, adding a late post injection, increasing diesel rail pressure, and running diesel-only.  These points were selected based on their propensity for soot production but are not necessarily representative of real operation.  A schematic explaining HPDI injection parameters is shown in Figure 52 while a summary of the test points run is in Table 8.  106   Figure 52 - HPDI injection parameters explained.  The NG main injection is substituted with diesel in the diesel-only test.  Table 8 - Selected HPDI operating points.  GPW: Gas Pulse Width; DRP: Diesel Rail Pressure; PSEP: Pilot Separation; LPI: Late Post Injection.  These points were run in both thermodynamic and optical configurations with the ICOS and pyrometry probe installed.  High-speed spatially resolved NL and OH* videos were recorded for test points in optical configuration. Measurement Point Injection Pressure Pilot Injection (diesel) Main Injection (diesel) Main Injection (NG) Post Injection (NG) (bar) Timing (aTDC) Duration (ms) Timing (aTDC) Duration (ms) Timing (aTDC) Duration (ms) Timing (aTDC) Duration (ms) Baseline 200 -7°  0.7 - - 3° 1.2 - - Long GPW 200 -7° 0.7 - - 3° 1.5 - - High DRP 220 -7° 0.7 - - 3° 1.2 - - Short PSEP 200 -7° 0.7 - - -3° 1.2 - - LPI 200 -7° 0.7 - - 3° 1.2 11° 0.9 Diesel-Only 200 -7° 0.7 3° 1.5 - - - -  Baseline Long GPW High DRP Short PSEP LPI Diesel-Only Equivalence Ratio 0.59 0.97 0.79 0.53 1.13 0.12 GIMEP (bar) Optical / Thermo 4.19 / 5.81 9.11 / 8.47 8.50 / 7.54 2.94 / 5.32 8.81 / 9.00 0.52 / 1.39   107  4.4.2 Effect of Diesel Fuel on Fuel Concentration Measurements (HPDI) DIDF and HPDI tests were performed with their respective cylinder heads and injectors.  Therefore, the effect of the diesel pilot on fuel concentration measurements is reevaluated under diesel-only conditions for the HPDI injector.  The fuel-concentration history for diesel-only conditions are shown below.  Figure 53 - HPDI diesel-only heat-release rate and fuel-concentration for thermodynamic and optical engine configurations.  Under both thermodynamic and optical testing conditions, the fuel concentration signal from the ICOS is virtually unchanged despite diesel pilot and main injections at -7° and 3° aTDC respectively.  The diesel-only test shows that the diesel jet does not impinge on the sensor itself 108  (similar to DIDF), as expected based on the spray geometry shown in Figure 30 and is further confirmed by inspecting the spatially resolved NL frame shown in Figure 54.  Because the diesel spray does not enter the ICOS detection volume, it is assumed that the subsequent results presented herein are for NG fuel concentration only.  Figure 54 - HPDI main diesel spray pattern for the diesel-only test relative to ICOS and pyrometry probes.  4.4.3 Evaluation of Fuel Concentration Histories The fuel concentration histories here are assessed here to determine the suitability of the ICOS for HPDI experiments.  Figure 55 and Figure 56 below shows a summary of all HPDI ICOS measured fuel concentrations for fired and non-fired conditions.  109   Figure 55 - Heat release rate and fuel concentration (fired and non-fired cases) for HPDI fueling in optical engine configuration.  Test points are labelled as per Table 8.  110    Figure 56 - Heat release rate and fuel concentration (fired and non-fired cases) for HPDI fueling in thermodynamic engine configuration.  Test points are labelled as per Table 8.  In Figure 40 and Figure 55, a significant difference in fuel concentration is noted between DIDF and HPDI.  DIDF is a premixed fueling strategy using a port fuel injector to supply the methane 111  fuel, while HPDI uses a dual co-concentric needle injector to directly inject NG.  As expected, there is no fuel concentration signal leading up to the injection events, and for the fired cases where diesel is injected to ignite the NG, there is limited NG reaching the sensor, evidenced by the low fuel concentration signal.  This trend is not noted for the short PSEP and high DRP optical tests where there is an initial spike, indicating that some fuel reaches the ICOS detection volume.  In the non-fired cases where the diesel pilot is disabled, higher fuel concentration signals are noted through the expansion stroke.  In general, the non-fired cases showcase a sharp increase in fuel concentration, hypothesized to be the initial divergence of gas jet (i.e., partial gas impingement on the sensor) and recirculation from jet impingement on piston bowl wall. This is followed by a reduction in the fuel concentration and eventual leveling out, presumably due to mixing with the remaining fresh air and resulting in well-mixed conditions.  In the optical configuration, fired tests for short PSEP and high DRP show an appreciable ICOS signal and are examined in closer detail.  In particular, these show the highest peak fired ICOS fuel concentration signals and are isolated in Figure 57 and Figure 58 along with the corresponding fuel concentrations for the thermodynamic engine configuration.  For the short PSEP, the introduction of NG is earlier in the cycle, with a higher piston position.  This creates different in-cylinder conditions at the time of NG ignition relative to the other HPDI points (i.e., cylinder temperature, mixing, etc.) and could be the reason why the main NG injection is detected initially before combustion consumes the fuel.  Likewise, with the increased diesel rail pressure (high DRP point), there is increased diesel and NG fuel mass (same injection duration) and a different distribution of locally rich zones, and this can contribute to ICOS detection of NG prior to its consumption.  112   Figure 57 - Short PSEP HPDI fuel concentration measurements for both fired and non-fired cases.  The initial spike in NG fuel concentration, highlighted in red, is present in both fired (optical only) and non-fired tests.  Figure 58 - High DRP HPDI fuel concentration measurements for both fired and non-fired cases.  The initial spike in NG fuel concentration, highlighted in red, is present in both fired (optical only) and non-fired tests. 113  The thermodynamic engine configuration resulted in advanced combustion timing, as well as higher overall fuel concentration due to the higher CR (relative to the optical configuration).  The optical tests were performed with intake heating to account for the effect of higher CR, but combustion phasing was still advanced under thermodynamic conditions.  A comparison of the calculated cylinder mean gas temperature Tmean-gas at TDC (eq. (17)) is presented in Table 9 and shows that TDC temperatures are higher under thermodynamic conditions.  Increased TDC temperatures may explain the advanced combustion phasing.  𝑇𝑚𝑒𝑎𝑛−𝑔𝑎𝑠(𝜃) =𝑃𝑐𝑦𝑙(𝜃) ∙ 𝑉(𝜃)𝑚𝑎𝑖𝑟𝑅𝑎𝑖𝑟 (17) Table 9 - Cylinder temperatures at combustion top dead center.  Temperatures are calculated based on the in-cylinder pressure as well as air mass flow rates. Test Point Thermodynamic TDC T (K) Optical TDC T (K) Baseline 784 740 Long GPW 838 744 Short PSEP 791 744 High DRP 814 749 LPI 840 747  In eq. (17), Pcyl(θ) instantaneous in-cylinder pressure, V(θ) is the instantaneous volume, mair is the cylinder air mass, and Rair is the air gas constant.  In both short PSEP and high DRP cases (optical), the fired fuel concentration signal spikes up in synchronization with the initial rise for the non-fired tests and are circled in red in Figure 57 and Figure 58.  This seems to indicate that the initial gas jet divergence reaches the sensor before combustion consumes the fuel at the sensor location.  The fired thermodynamic tests do not share the same distinct increase in fuel concentration signals and may indicate differences in mixing (piston bowl shape) or earlier combustion phasing consuming the fuel prior to reaching the sensor (see TDC temperature comparisons in Table 9).  114  Differences in fuel mixing characteristics are particularly evident in Figure 55 and Figure 56 when comparing the non-fired fuel concentration between the two engine configurations.  The thermodynamic configuration exhibits a more gradual rise in fuel concentration versus the optical tests, which is likely a function of the bowl geometry.    Illuminated videos of non-fired tests were recorded using the high-speed NL camera to better understand the gas jets and how it affects the sensor signals.  Due to the passive nature of the imaging diagnostics, it was difficult to visualize and interpret the exact mixing characteristics of the gas jets.  The gas jet is normally invisible and the jet is only partially visible with external lighting and is reliant on some of the diesel that gets entrained in the spray (a characteristic of the development HPDI injector employed).  This provides some indication of the gas distribution in the cylinder but is not a complete tracer since the diesel droplets will evaporate.  For some measure of how the NG spray affects the ICOS signals, a series of non-fired freeze-frames for the short PSEP test is compared with the fuel concentration signals below.  115         Figure 59 - Short PSEP illuminated non-fired freeze-frames.  The development of NG jets is shown from 0° to 5° aTDC.  Main NG CSOI is at -3° aTDC for this test.  ICOS Sensor 116  Based on the illuminated, non-fired gas jet images, it is conceivable that the initial spike in fuel concentration is due to the initial penetration of the gas jet.  The jets are first visible at 0° aTDC, and at 3° aTDC, are at their longest visible length.  The fuel concentration reaches its first peak at approximately 4-5° aTDC and is assumed to be due to the gas divergence from the jet axis.  The second and larger fuel concentration peak at 9° aTDC in the non-fired test is attributed to the recirculation of the gas following its impingement along the piston bowl walls.  In contrast, at 5° aTDC the fired test shows a decline in fuel concentration.  The same set 0° to 5° aTDC freeze-frames for the fired test is shown in Figure 60.       Figure 60 - Short PSEP NL fired freeze-frames.  The development of combustion is shown from 0° to 5° aTDC.  Main NG CSOI is at -3° aTDC for this test.  From the images, combustion (approximated by the overall NL coverage of the combustion chamber) accelerates significantly between 4° to 5° aTDC, especially at the outer periphery of the ICOS Sensor 117  bowl where the ICOS is located.  This is in line with the decrease in fired fuel concentration signal at the same crank angles (4°-5° aTDC), and supports the idea that for the other HPDI conditions tested, combustion consumes the fuel before it reaches the ICOS detection volume.  For the stratified conditions tested, the ICOS provides some insight to combustion phenomena though combustion occurs fast enough that the fuel concentration signal at the ICOS location is restricted.  An evaluation of the injector orientation or ICOS mounting geometry may provide stronger signals in a direct injection environment.    4.4.4 ICOS HPDI Testing Summary The ICOS HPDI tests contrast with the DIDF experiments where the methane fuel was premixed.  Diesel-only tests were repeated with the new head and HPDI injector and it was confirmed that the ICOS sensor was insensitive to the diesel spray due to position of the ICOS, relative to the diesel sprays.  For the HPDI test points evaluated, the fired fuel concentration signal was generally lower than previous (fired) DIDF tests.  This is attributed to the stratified nature of the injection strategy, as well as the rapid onset of combustion consuming the NG before substantial amounts can reach the sensor location.  A series of pilot-disabled, non-fired tests were completed to corroborate this observation.  These tests were combined with 2D-spatially resolved NL imaging to provide an indication of gas jet characteristics.  The NG jets were difficult to visualize, but using high-speed, externally illuminated imaging indicated the early position of the gas jets. When compared to the fuel-concentration histories, the initial increase in fuel-concentration under stratified conditions is due to the gas divergence from the gas jet axis followed by a second, larger peak from the reflection of the main gas jet against the piston bowl and cylinder walls.  Future tests under slightly premixed combustion (SPC) conditions may present more opportunities to 118  apply the ICOS to HPDI.  Under the stratified HPDI conditions tested here, there is more utility in using the ICOS to examine jet structures and in-cylinder mixing under non-firing conditions.  4.5 ICOS Summary This chapter documents the application of an in-cylinder local infrared absorption sensor for resolving the local AFR and fuel concentration under premixed (DIDF) and stratified (HPDI) fueling strategies.  The ICOS proved to be useful in evaluating the DIDF premixed equivalence ratio, as well as differences in combustion modes (flame propagation from the bowl periphery inwards vs. central ignition and flame propagation outwards).  HPDI tests indicated that the current positioning of the ICOS limited its suitability for characterizing the fuel concentration history during combustion; however, it is well suited for characterizing the fuel concentration when the NG can reach the ICOS detection volume prior to combustion.  The latter was evident for non-fired tests, as well as test conditions with longer NG combustion delays due to lower temperatures and fueling strategies.  Fuel concentration signals were also extremely sensitive to premixed vs. non-premixed conditions, and for HPDI, jet impingement on the piston bowl can reflect a significant amount of NG back towards the sensor.    Regardless of the combustion strategy employed, both engine configurations benefited from comparisons against high-speed, spatially resolved imaging and provides additional insight not previously available under thermodynamic testing.  The combination of 2D optical and conventional thermodynamic analysis, along with local fuel concentration provides qualitative insight into DIDF and HPDI combustion modes.  Ultimately, the application of the ICOS may 119  enable researchers to have more insight into in-cylinder combustion phenomena without requiring expensive and tedious optical access.  120  Chapter 5: Pyrometry Probe In conjunction with the ICOS probe used throughout this investigation, a line of sight two-colour pyrometry probe was developed to determine in-cylinder temperature T and soot concentration KL.  These parameters are important for characterizing the temporal soot evolution and can also provide insight to combustion temperatures which may affect NOx emissions.  Comparison between in-cylinder KL and exhaust-out PM emissions is also possible.  This probe is a passive measurement approach that relies on thermal radiation from the soot, in contrast to the ICOS’ use of an external lamp.  A photo of the pyrometry probe and ICOS installed onto the HPDI and DIDF head is shown in Figure 23 and Figure 29 respectively.  5.1 Pyrometry Probe Design The concept of the pyrometry probe employed in this investigation is based on the original designs of Yan and Borman [29] and is shown in Figure 14.  In their work, an optical radiation probe consisting of a specially designed trifurcated fiber optical bundle and sapphire rod window was developed and used to measure the combustion radiation in a Cummins single cylinder diesel engine.  An important feature of Yan and Borman’s probe is that it included a self-cleaning window, a feature this investigation aims to replicate in the design of the new probe.  The self-cleaning window is critical because it allows the window to remain clean under steady-state full load conditions without significant degradation of the optical signal when testing for longer durations such as in thermodynamic tests.  Yan and Borman utilized their probe to perform in-cylinder pyrometry measurements for soot concentration and temperature and the operating principle of this method is described in detail in §2.2.2.  121  The probe was originally designed for the DIDF head (Figure 17, right).  Several iterations of the probe design were created, some of which revolved around implementing the probe in the existing glow plug bore.  These early designs avoided machining the DIDF head, which limited possible solutions.  Focus then shifted to modifying the head to fit the probe.  A spare head was cut into sections for laser scanning, and this facilitated the development of a parametric model which was used to constrain the pyrometry probe location and design.  The details of this process (i.e., cutting the spare head and developing a parametric 3D model) can be found in [24].  The selected bore location was the only spot on the DIDF head that had the space to house such a large probe.  A photo of the head before and after modification with the probe installed is shown in Figure 61 along with the as-designed probe bore dimensions for scale in Figure 62.  Figure 61 - DIDF head pre- and post-modifications for pyrometry probe.  A hole was machined through the coolant passage to allow the installation of a probe sleeve.  122   Figure 62 - As designed probe bore dimensions on the original diesel head used for DIDF studies.  The bore is partially exposed to the coolant channels in the head.  Dimensions are in inches.  The finalized install location of the probe for this head required several key aspects to be addressed.  The probe passes through a coolant channel, and this sets the main requirement of the probe such that there is an external sleeve to house and protect the optical components.  The space constraints on the DIDF head determined the overall form factor of the probe sleeve.  In addition, the probe must seal against combustion pressures, and there is an external and internal clamping mechanism 123  included in the sleeve to provide a seating force against combustion.  And finally, perhaps the most crucial to the functionality of pyrometry, the probe must pass combustion light through a robust optical interface.  A comparison between the final design and Yan & Borman’s implementation [29] is shown in Figure 63.     Figure 63 - Pyrometry probe cross sections.  Left: as designed probe assembly for DIDF.  Right: Yan and Borman’s design.    The optical pathway consists of an opening in the probe sleeve, a sapphire rod brazed into an invar holder, and a series of apertures prior to a trifurcated fiber optic cable.  These components are labelled in Figure 63 above.  The opening at the bottom of the probe sleeve is both a view and flow restriction design feature.  The role of this restriction is to drive the flow of high temperature and high-velocity gases across the front sapphire rod interface to prevent thermophoretic deposition of 124  soot on the window.  This is an innovation developed by [29], and analysis regarding the flow across this interface is presented in §5.1.1.  Details of the collection optics following the sapphire rod and invar holder assembly are discussed in §5.1.2.  With regards to the sapphire rod, the design required a mechanical seal against combustion pressures.  To achieve this seal, the sapphire rod is joined to a mating part and clamped against a copper washer to provide a full seal against combustion pressures.  In [29], the sapphire rod is silver brazed into an invar part.  Invar exhibits similar thermal expansion coefficients as sapphire at the temperatures expected during continuous operation (6E-6/°C invar compared to 7E-6/°C sapphire at 300°C [56]) and is highly stable across a wide temperature range.  This is desirable in that thermal stresses will be minimized during the high temperatures experienced during operation.  Steel, for comparison, exhibits thermal expansion coefficients on the order of 13E-6/°C to 15E-6/°C at 300°C.  Figure 64 shows the sapphire rod and invar assembly after it was brazed.    Figure 64 - Sapphire and invar holder brazed assembly.  Brazing was performed by Bodycote in California.   125  Though the pyrometry probe was designed for use in the DIDF head, procurement timelines delayed the implementation of the probe.  Instead, the DIDF head was used solely for ICOS studies, and the probe was reserved for use in a newly designed head [24].  Due to the optical-first design priority of the new head, mounting configurations were much more flexible.  The probe design remained the same, and was positioned in a radius along with 2 other probe bores (currently reserved for the ICOS and pressure transducer) centered on the piston bowl.  These probe bores are pointed directly into the piston bowl at an angle and the injector was repositioned to an upright position central to the bowl.  This configuration was specifically chosen to pair with 2D spatially resolved measurements through the Bowditch piston and Figure 65 shows the probe bore geometry in the new head.  Figure 24 provides an overview of the probes in the new head.    Figure 65 - As designed probe bore dimensions on the custom head used for HPDI studies.  The bore is sealed from the coolant channels.  Dimensions are in millimeters.  126  While the design is based heavily on Yan and Borman’s work [29], [56], the experimental test facility proved to have significant impact and constraint on the overall form.  For instance, the cylinder heads (DIDF and HPDI) only have 2 valves which prevents the installation of the probe in an intake or exhaust port as in Yan and Borman’s design for a 4-valve head, where an exhaust valve was removed to install the probe.  Instead, the probe was installed into a custom location through the head cooling channels for both DIDF and HPDI configurations.  For the DIDF head (original diesel head), a mounting location was machined through the head and cooling channel.  For HPDI, diagnostic probe bores were included in the custom head design.  The fully assembled pyrometry probe is shown in Figure 66. 127   Figure 66 - Fully assembled pyrometry probe and photodetector assemblies.  The ICOS probe is also shown with its mounting clamp.  5.1.1 Flow Restriction Analysis The sapphire rod acts as a window and seal to the combustion chamber and is brazed in an invar steel holder.  The front geometry of the probe sleeve interacts with the sapphire rod to create a restriction to the cavity surrounding the rod.  This restriction is such that the flow of combustion gases in and out of the chamber will flush the front surface of the window to prevent 128  thermophoretic deposition of soot and other combustion products [29], [56] and is shown in Figure 67.    Figure 67 - Sketch of the probe restriction showing the flow of hot gases across the front face of the sapphire rod.  To substantiate the claim that the restriction forces high velocity flow of hot gases across the front face of the sapphire rod, a flow model was developed.  The restriction is treated with a mass flow correlation based on one-dimensional isentropic flow analysis [57] and is used here along with geometry information per Appendix F:   ?̇?(𝑡) = 𝐶𝑑 ∙ 𝐴 ∙𝑝𝑖𝑛(𝑡)√𝑅 ∙ 𝑇𝑐(𝑡)∙ 𝜓 (𝑝𝑖𝑛(𝑡)𝑝𝑜𝑢𝑡(𝑡)) (18) 129   𝑝𝑐𝑟 = [2𝜅 + 1]−𝜅𝜅−1 (19)  𝜓(𝑝𝑖𝑛(𝑡)𝑝𝑜𝑢𝑡(𝑡))={      √𝜅 [2𝜅 + 1]𝜅+1𝜅−1, 𝑝𝑖𝑛/𝑝𝑜𝑢𝑡 ≥ 𝑝𝑐𝑟[𝑝𝑜𝑢𝑡𝑝𝑖𝑛]1𝜅∙ √2𝜅𝜅 − 1∙ [1 − (𝑝𝑜𝑢𝑡𝑝𝑖𝑛)𝜅−1𝜅] , 𝑝𝑖𝑛/𝑝𝑜𝑢𝑡 < 𝑝𝑐𝑟 (20) Here, Cd represents an experimentally determined discharge coefficient, A is the restriction area, Tc is the cylinder temperature, and variables pin and pout represent the inlet and outlet pressures depending on flow direction (driven by differential pressure between the cylinder and cavity due to combustion or piston motion).  κ is the heat capacity ratio of the gas, R the universal gas constant, and pcr is the critical pressure ratio for choked flow.  Using the above equations for valve mass flow rate, several assumptions were made regarding the in-cylinder conditions and the cavity conditions.  In-cylinder pressure data was taken from a real operating point, with initial cylinder temperature Tc set to 300K at bottom dead center.  Tc for the rest of the cycle is calculated using the ideal gas law with real pressure data and calculated in-cylinder mass at the initial time step (bottom dead center).  At all points during the cycle, Tc is also assumed to be equal to Tcavity, the cavity temperature (i.e., no heat transfer).  Finally, Cd was arbitrarily varied as part of a sensitivity analysis.  The result is an ODE system that is solved using 130  MATLAB ODE45 solver and provides the velocity and pressure drop through the restriction area.  The model and assumptions, while rough, provides an approximate indication of the flow velocity around the front of the sapphire rod window and is compared to Yan and Borman’s results.  Results are plotted in Figure 68 and Figure 69 and show significant flow velocity across the restriction.  Figure 68 - Modelled gas pressure drop between cylinder and probe cavity volume for varying Cd. -200 -150 -100 -50 0 50 100 150 200-0.500.5CADPressure Drop (bar)  Cd = 0.3Cd = 0.5Cd = 0.7Cd = 0.9131   Figure 69 - Modelled gas velocity through restriction area between cylinder and probe cavity volume for varying Cd.  From the results, the maximum velocity through the flow restriction is approximately 20 m/s for this test case.  The velocity will change with in-cylinder pressure, and so this only provides an estimation of the flow performance.  Estimated pressure drop ranges from 0.05 bar to 0.5 bar based on Cd selection.  Yan’s work [56] also performs similar modelling with additional consideration for heat transfer, and also results in similar flow velocities and pressure drops across the interface: 40, 20, and 10m/s and 0.35, 0.1, and 0.02bar for large, medium and small cavity volumes respectively.  Yan & Borman reported that the high velocity flow (estimated 40m/s) for their large cavity volume maintained a clean window surface for 7 hours continuous operation at ϕdiesel = 0.5.  Based on this analysis and the overall similarities between designs, the probe is expected to perform like the Yan and Borman design and exhibit self-cleaning capabilities.  -200 -150 -100 -50 0 50 100 150 200-20-15-10-50510152025CADVelocity (m/s)  Cd = 0.3Cd = 0.5Cd = 0.7Cd = 0.9132  5.1.2 Pyrometry Probe Optical Components The optical components that are housed within the probe include the sapphire rod, apertures, fiber optic adapter, and finally, the trifurcated fiber cable.  The fiber optic cable is split into 3 ends, though only 2 are used for pyrometry.  Each end is terminated with a collimator and narrow bandpass filter (700 or 800nm) prior to a photodetector.  A simplified schematic of the optical components is shown in Figure 70.  Figure 70 - Pyrometry probe connection diagram  Within the probe sleeve and immediately following the front sapphire rod and invar holder combustion seal interface are a series of apertures.  While the design allows the implementation of these apertures to limit the field of view (FOV), the current investigation has these removed (i.e., only a fiber optic cable adapter follows the sapphire rod and invar holder assembly).  This maximizes the field of view, and renders the probe less sensitive to the probe installation location 133  relative to the fuel jets.  The apertures and the fiber optic adapter are designed with conical features to promote self-alignment when axial force through the assembly is applied via the clamp.  Figure 71 shows the probe assembly that was implemented for this study.     Figure 71 - Pyrometry probe configuration as used in HPDI experiments.  Apertures were removed for the maximum FOV, estimated to be approximately 50° cone angle based on single-reflection ray tracing analysis.  A crucial aspect of the pyrometry probe is its FOV.  To analyze this, the sapphire rod was treated similarly to a fiber optic cable, though the acceptance angle was limited to those within the critical angle of total internal reflection.  This was calculated by manipulating Snell’s law (used for calculating angle of light refraction between two different mediums) in eq. (21) such that θ2 = 90°.  The critical angle of total internal reflection is calculated in eq. (22). 134   𝑛1𝑠𝑖𝑛𝜃1 = 𝑛2𝑠𝑖𝑛𝜃2 (21)  𝜃𝑐 = arcsin (𝑛2𝑛1) (22) In eq. (21) and (22), subscripts 1 and 2 refer to the medium through which the light passes.  θc is the critical angle of total internal reflection and n is the index of refraction for the medium.  For this application, the mediums that the light passes through are the combustion chamber (air), sapphire rod, and finally air again, before the fiber optic receiving interface.    Assuming the light that reaches the fiber optic interface is either through a direct path or a single reflection within the sapphire rod, the maximum FOV under these assumptions can be calculated by using eq. (21) and by tracing a pathway through the sapphire rod to the fiber receiving interface.  The projected maximum FOV (50° cone angle) using this method is shown in Figure 72.  While the sapphire rod is capable of full internal reflection at much steeper angles (i.e., multiple bounces within the sapphire rod), these are not considered as the fiber receiving interface acts as an aperture to these higher angle reflections. 135   Figure 72 - Projected maximum FOV under experimental setup geometry (50° cone angle).  An arbitrary ray starts at the extremities of the sapphire rod FOV and ends at fiber optic receiving interface.    The receiving fiber optic cable is a randomized bundle consisting of 19 individual fibers.  It is trifurcated and extends out from the pyrometry probe assembly to split into three ends of 6, 6, and 7 fibers.  Each end is terminated with a collimator, narrow bandpass filter, and photodetector.  In this study, 700nm and 800nm narrow bandpass filters are selected for the 6-fiber branches and are used for two-colour pyrometry.  The remaining branch consisting of 7 fibers is reserved for future use, possibly in the UV wavelength of 310nm for comparison with 2D high-speed UV imaging of the engine in optical configuration.  The 19 bundled fibers are randomized such as to minimize the effect of seeing different field of views, and are terminated with SMA connections at each end.    A detailed look at the photodetector assembly and connection to the fiber is presented in Figure 73.  With the use of a randomized fiber optic cable, a collimating lens setup is required for each 136  measurement leg of the cable to focus the light onto the photodetectors.  For the 700nm and 800nm wavelengths, a Thorlabs F240SMA-780 fiber collimation package was used.  The divergence in this case was negligible (0.05° @ 700nm, 0.04° @ 800nm) for the short focal lengths involved and therefore a 780nm center wavelength (CWL) collimation package was deemed suitable.  These collimators were paired with 12.5mm diameter bandpass filters from Edmund Optics (#87-888 700nm CWL, #65-118 800nm CWL, OD4 10nm FWHM bandpass filters).  The photodetectors used are Thorlabs PDA36A switchable gain detectors and are directly connected to the DAQ through a NI 9215 Simultaneous Analog Input module.  The UV leg of the cable was not used for this investigation.    Figure 73 - Detailed pyrometry probe photodetector setup  5.2 Pyrometry Probe Calibration The calibration of the probe links the voltage output of the photodetectors to a physical quantity.  The physical quantity in this case is the apparent soot temperature which is required for pyrometry calculations and relates in-cylinder soot concentration and temperature as discussed in detail in 137  §2.2.2.  The voltage measured is directly proportional to the radiation intensity due to combustion, and subsequently, the temperature of the radiating soot.    A well characterized light source (Labsphere Custom USS-800 Integrating Sphere) was used to calibrate the spectral response of the probe, optical fiber, and photodetector assembly at 700nm and 800nm.  Specifically, the integrating sphere is calibrated for a range of wavelengths at which the uniform output intensity is known.  The voltage measured at the photodetectors can then be tracked against the intensity of the sphere, and subsequently, the apparent temperature Ta is measured.  A photo of the benchtop calibration setup is shown in Figure 74.  Figure 74 - Pyrometry probe benchtop calibration setup with Labsphere.  Neutral density filters of various optical densities (OD = 0.1 - 0.6) were placed between the probe and integrating sphere such that a range of known intensities can be measured.  This provides a calibration curve from which the combustion intensity can be extrapolated based on the measured 138  voltage.  Care must be taken in ensuring the benchtop test is as repeatable as possible to reduce errors in the calibration.  To reduce uncertainties in calibration, several steps were taken to ensure good agreement between separate instances of calibrations and tests.  These steps are documented in Appendix G.  In general, a “clean” and “dirty” calibration must be performed for each optical engine test such that the attenuation of the sapphire rod is tracked for each test as it becomes contaminated.  Clean and dirty refers to the state of the sapphire rod – the sapphire rod is considered dirty after any test.  A sample clean and dirty calibration for 700nm is shown in Figure 75 below.      Figure 75 - Sample 700nm calibration after a set of optical tests.  Clean and dirty calibration refers to the state of the sapphire rod before and after testing.  139  Throughout all tests, the clean calibration was used in the calculations of soot temperature and concentration.  A separate evaluation on the effects of window fouling is performed here to characterize the self-cleaning characteristics of the probe design.  Figure 76 below shows the difference in calibration before and after a set of thermodynamic tests (~3 hours).  Figure 76 - Comparison calibration before and after 3 hours of continuous thermodynamic operation.  In Figure 76, the maximum drop in voltage between a corresponding clean and dirty calibration is approximately 0.5V at the highest intensities.  To show the effect of this difference in calibration, a sample diesel-only test point (Figure 77) was processed with both calibrations to examine changes in the soot temperature and KL.  The dirty calibration was found to overestimate T and 140  KL – a 6K increase in temperature and 8% increase in KL was observed at their corresponding peak values.  For this specific clean-dirty calibration pair (~3 hours run time), the self-cleaning feature of the probe appears to result minor changes in the temperature while KL exhibits higher sensitivity.    Figure 77 - Diesel-only HPDI test.  Two different calibration profiles were applied to showcase the change in temperature and KL.  141  5.3 Pyrometry Probe HPDI Experimental Results and Discussion To assess the utility of the pyrometric probe under a range of engine operating conditions, several HPDI injection strategies and injection parameter variations were considered.  Relative to a baseline operating point with a single diesel pilot and main natural gas injection, the gas pulse width (injection duration), injection pressure, and pilot to main injection dwell time were varied.  In addition, a late post injection of natural gas was also considered, as this has been noted in previous HPDI works as having a significant impact on the soot emissions [21].  Finally, an operating mode with only diesel injection (pilot and main) was also considered.  The specific timings and durations for each of the injection strategies are summarized in Table 10.  All measurements were carried out at 600 rpm.  It should be noted that the Westport injector employed here is not representative of a standard diesel common-rail injection system, nor is it representative of production HPDI injectors. Table 10 - Selected HPDI operating points.  GPW: Gas Pulse Width; DRP: Diesel Rail Pressure; PSEP: Pilot Separation; LPI: Late Post Injection.  These points were run in both thermodynamic and optical configurations with the ICOS and pyrometry probe installed.  High-speed spatially resolved NL and OH* videos were recorded for test points in optical configuration. Measurement Point Injection Pressure Pilot Injection (diesel) Main Injection (diesel) Main Injection (NG) Post Injection (NG) (bar) Timing (aTDC) Duration (ms) Timing (aTDC) Duration (ms) Timing (aTDC) Duration (ms) Timing (aTDC) Duration (ms) Baseline 200 -7°  0.7 - - 3° 1.2 - - Long GPW 200 -7° 0.7 - - 3° 1.5 - - High DRP 220 -7° 0.7 - - 3° 1.2 - - Short PSEP 200 -7° 0.7 - - -3° 1.2 - - LPI 200 -7° 0.7 - - 3° 1.2 11° 0.9 Diesel-Only 200 -7° 0.7 3° 1.5 - - - -  Baseline Long GPW High DRP Short PSEP LPI Diesel-Only Equivalence Ratio 0.59 0.97 0.79 0.53 1.13 0.12 GIMEP (bar) Optical / Thermo 4.19 / 5.81 9.11 / 8.47 8.50 / 7.54 2.94 / 5.32 8.81 / 9.00 0.52 / 1.39  142  The pyrometry probe was used to characterize the soot concentration and temperature for the operating conditions outlined in Table 10, and optical results are shown in Figure 78 and Figure 79 along with the apparent heat release rates.  The heat release rate, soot temperature and KL factors for all optical HPDI test points were averaged across 15 fired cycles.  The soot temperature and KL are only relevant for regions of the cycle when soot radiation can be detected, and are indicated as zero otherwise.  Mean gas temperature (Tmean-gas) is calculated via eq. (17).  Figure 78 - Optical heat release rate, soot temperature, and KL for Baseline, High DRP, and Short PSEP test points.  143   Figure 79 - Optical heat release rate, soot temperature, and KL for Long GPW, LPI, and Diesel test points.  The pyrometry probe can resolve differences between operating points for in-cylinder soot temperature and concentration.  For example, higher temperatures are noted for the short PSEP (~2400K compared to 2000-2200K for the other test points), as expected with the higher rate of heat release, as well as combustion phasing closer to piston top dead center.  For the LPI test point, a second peak appears to correspond to the second gas injection.  The short PSEP and baseline test points produced little soot whereas the long GPW produced the most soot and is evidenced by the 144  strong KL peak.  This trend was expected due to the higher global equivalence ratio [33].  Also in agreement with expectations from diesel pyrometry studies [33], the higher injection pressure (high DRP) resulted in higher peak KL values, relative to the baseline.  It should be noted that peak KL does not necessarily correspond to higher engine out soot emissions as the increased turbulence from higher injection pressure also promotes oxidation.  The diesel and late post injection (LPI) points show two peaks in the soot formation/oxidation shape.  In the diesel case, the two peaks are associated with the diffusion flame prior to the flame impinging on the wall, and the subsequent mixing after flame impingement, as shown in Figure 80.  For LPI, both gas combustion events (‘main’ and ‘late’) are well mixed, similar to the post-wall impingement phase of the diesel combustion.  Thus, the second peak noted during LPI is due to the second gas injection, as shown in Figure 81.    145    Figure 80 - Comparison between high-speed natural luminosity imaging and KL factor measured by the probe for diesel-only operation.  Freeze-frame images from the spatially resolved NL videos are flipped vertically due to a mirror.   146    Figure 81 - Comparison between high-speed natural luminosity imaging and KL factor measured by the probe for HPDI late post injection (LPI) operation.  Freeze-frame images from the spatially resolved NL videos are flipped vertically due to a mirror.   147  A set of supplementary thermodynamic tests at the same points were also performed.  The same plots for heat release rate, soot temperature, and KL are shown in Figure 82 and Figure 83 and discussed in detail.  Figure 82 - Thermodynamic heat release rate, soot temperature, and KL for Baseline, High DRP, and Short PSEP test points.  148   Figure 83 - Thermodynamic heat release rate, soot temperature, and KL for Long GPW, LPI, and Diesel test points.  The rising KL trend at ~42° aTDC for the LPI test is likely due to an incorrect numerical solution to eq. (9).  In general, the same trends noted earlier for the optical tests (i.e., higher temperature for short PSEP, second peak for LPI due to second gas injection, etc.) hold true in thermodynamic as well.  However, the relative magnitudes for KL between tests do not match between optical and thermodynamic.  This is hypothesized to be due to the different piston bowl shape resulting in 149  different mixing characteristics.  For the long GPW and LPI tests, the pilot ignition was clearly detected by the pyrometry probe (Figure 83, temperature and KL) and is another indicator of the variation in mixing characteristics between the two configurations.  Differences in KL could also be due to the continuous operation in thermodynamic tests – in-cylinder temperatures are higher and can contribute to enhanced soot oxidation.  Between the two configurations, there is a phase shift.  This is due to the higher compression ratio in thermodynamic configuration, resulting in slightly earlier combustion phasing.    From the comparisons between the natural luminosity still frames and the KL factor, it is clear that the spatial positioning of the probe is crucial for interpretation of the results.  This is further reinforced by the thermodynamic tests where differences in observed trends (i.e., pilot ignition detected, different KL magnitudes) are attributed to the piston bowl shape and mixing characteristics.  The probe offers limited field of view and is only a line of sight measurement.  Subsequently, the measurement is highly subject to non-uniform combustion such as in Figure 80 where diffusion flames can dictate the resultant signal.  Increased field of view can improve the link between the measured KL and the actual soot concentration across the entire cylinder and will help with interpretation between optical and thermodynamic results.    5.4 Pyrometry Probe Summary An in-cylinder, line of sight pyrometry probe was developed to provide insight to the in-cylinder soot temperature and concentration in both optical and thermodynamic (metal) engine configurations and was used here to evaluate HPDI combustion processes.  The effects of the injection strategy on the in-cylinder soot temperature and concentration were noted to follow 150  trends from prior diesel investigations [33].  Comparisons between spatially resolved natural luminosity imaging and probe indicated that the output of the probe is very sensitive to the charge motion and location of the radiating soot.  This was found to be particularly influenced by impingement of the flame on the piston bowl.  Thermodynamic tests revealed differences in the in-cylinder mixing characteristics and it was noted that continuous operation in these tests may also contribute to changes in the recirculated gas composition, as well as an overall increase in cylinder temperatures.  In future work, the probe will be utilized along with spatially resolved, high-speed imaging to provide insight to in-cylinder soot formation behaviour for various combustion strategies and provide a guide for interpretation between optical and thermodynamic results.  151  Chapter 6: Conclusions and Recommendations The overarching objective of this investigation is to develop the foundations for a new thermo-optical approach to ICE research.  The methodology combines conventional thermodynamic analysis (i.e., in-cylinder pressure, HRR) with insight provided by a combination of probe-based and high-speed spatially resolved optical diagnostic tools.  These currently include, but are not limited to, high-speed spatially resolved NL and OH* imaging, an in-cylinder local infrared absorption sensor for fuel concentration, and a two-colour pyrometry probe for soot temperature and concentration.  The intent behind this methodology is to address the shortcomings of optical research engines, which may not necessarily represent an equivalent thermodynamic all-metal engine [7].  Differences in materials, heat transfer characteristics, load limitations, and cylinder geometry, are just a few of the key distinctions.  As such, the thermo-optical approach proposes the use of less intrusive optical measurement probes capable of continuous and high-load thermodynamic operation.  These probes are initially evaluated under optical conditions and compared to high-speed spatially resolved imaging to provide insight to in-cylinder combustion phenomena.  The knowledge gained from this exercise can then be applied to thermodynamic testing under continuous operating conditions as an indication of the parallels or dissimilarities between the two configurations.  Simply put, the thermo-optical methodology utilizes an optical engine with high-speed spatially resolved imaging to calibrate probe-based optical measurement tools.  The implementation of the methodology can provide new perspectives in an all-metal thermodynamic engine running under real-world operating conditions.    This work details the development and application of in-cylinder fuel concentration and pyrometry optical diagnostic tools in a diesel-ignited dual-fuel natural gas engine as part of on-going 152  development of the thermo-optical methodology.  The probes were successfully commissioned in both optical and thermodynamic configurations, a key feature required in the thermo-optical methodology.  It was shown that the probes could operate under continuous thermodynamic operation with higher temperatures and without significant fouling of optical components.  This chapter summarizes the major conclusions from the application of these probes and provides a roadmap for future application and development of the methodology.  6.1 ICOS The local in-cylinder fuel concentration was measured using the Internal Combustion Optical Sensor (ICOS), a commercially available system from LaVision Inc.  The ICOS was operated under DIDF and HPDI combustion strategies, in part to understand how the sensor reacts under premixed and stratified conditions, but also to provide a point of comparison between thermodynamic and optical engine configurations.    For premixed DIDF operation, the ICOS was only implemented under thermodynamic conditions.  Resulting in-cylinder fuel concentration measurements were compared against previous optical measurement campaigns and it was demonstrated that the relation between fuel conversion mechanisms and in-cylinder fuel concentration histories could be elucidated in an all-metal engine, which prior to this work, required optical access.  DIDF test points previously observed to have a fuel conversion mechanism associated with ignition at the piston bowl periphery followed by flame propagation towards the central bowl region were compared to tests where ignition started in the central bowl region before propagating outwards (shown in Figure 10).  The onset of decreasing fuel concentration indicated the combustion progress through the cylinder relative to the ICOS 153  position and identified the mode of combustion.  A sensitivity analysis of the results evaluated cylinder volume, reference intensity (I0), and air mass for their contribution to errors, and found the AFR results to be most sensitive to I0 because of the lack of a “zero-fuel” reference point in DIDF.  The ICOS was also used under HPDI operating conditions to evaluate its suitability for characterizing the local fuel concentration in stratified charges.  While the test points evaluated were selected to provide the most utility to pyrometry probe measurements, these tests served to distinguish the fuel concentration differences in premixed vs. stratified mixtures of NG and air.  Tests were repeated in thermodynamic and optical configurations for comparison.  For the tested HPDI conditions, the in-cylinder fuel concentration signal is largely negligible.  This is attributed to the stratified nature and rapid consumption of NG immediately after its injection, as well as the spray axes pointing away from the ICOS position.  A series of non-fired tests with the diesel pilot disabled were performed for the same NG injection parameters to ascertain the effect of combustion on the fuel concentration signal.  From these evaluations, the initial increase in fuel concentration is linked to the gas jet divergence from the spray axes followed a second, larger peak due to the reflection of the gas jets from the piston bowl walls.  Externally illuminated high-speed videos were used to substantiate the gas jet position relative to the temporal fuel concentration signal.  For the small region of HPDI operating space considered, the ICOS does not provide significant value in identifying combustion modes, but instead, can provide insight to gas jet structures and in-cylinder mixing phenomena.  Future tests with SPC conditions may present more opportunities to link in-cylinder fuel concentration to different HPDI combustion modes.    154  A preliminary comparison between optical and thermodynamic configurations was possible through assessment of the in-cylinder fuel concentration signal.  Owing to the rapid consumption of NG in fired tests, the non-fired tests provide the best point of comparison between the two setups.  Most notably, it was shown that the thermodynamic configuration has a more gradual increase in local fuel-concentration relative to the equivalent optical runs, which is attributed to the differences in mixing characteristics from the different piston bowl shapes.  Overall, a higher peak fuel concentration is also observed in the thermodynamic configuration from the increased compression ratio.    Regardless of engine configuration or combustion strategy employed, the ICOS was found to be sensitive to the spray orientation of the injectors.  Fouling is a potential problem, especially in the case of diesel spray impingement leading to rapid deterioration of the mirror surface.  The ICOS is particularly well-suited to premixed fuel concentration measurements, and can provide insight to fuel conversion mechanisms without the need for optical access.  HPDI proves to be more of a challenge in terms of interpretation for the combustion mode but the ICOS was still found to be useful for non-combusting stratified measurements.  The operating regime of HPDI that was explored is extremely small, and better experimental design may supplement the current results.  In future thermo-optical tests, the ICOS can potentially be used to characterize in-cylinder mixing characteristics for stratified conditions.  6.2 Pyrometry Probe The two-colour pyrometry probe was designed and built for the optical engine test facility at UBC.  Line of sight measurements of combustion intensities at 700 and 800nm wavelengths were used 155  to calculate in-cylinder soot temperature and concentration.  The probe was implemented for HPDI thermodynamic and optical configurations only.  A baseline HPDI test point provided a reference for comparison and changes in injection parameters from the baseline were selected based on previous work by Ehsan Faghani [55] to provide a range of tests with different PM formation tendencies.  It was demonstrated that the soot formation and oxidation trends were dependent on the equivalence ratio, injection pressure, as well as split injection strategies.  These generally agreed with observations from prior diesel investigations [33] and can also be used to identify the injection strategy employed based on the measured KL trends for some test points.  Comparison of the KL signal with high-speed NL videos identified the sensitivity of the probe output to charge motion and the location of the radiating soot.  This is particularly influenced by the location of the diesel jets during the pilot auto-ignition event and the impingement of the flame and/or gas jet on the piston bowl and subsequent turbulent mixing.    In comparing the optical and thermodynamic configurations, several differences were noted.  Combustion timing was advanced for thermodynamic operation due to the compression ratio and a comparison of cylinder temperatures at TDC consistently show higher temperatures for the thermodynamic configuration.  It was also observed that the pilot combustion events were sometimes detectable in the thermodynamic soot temperature and KL results.  This phenomenon is ascribed to the bowl shape in thermodynamic configuration resulting in slightly different jet structures. Soot formation and oxidation trends generally matched between the two configurations, though the magnitude varied inconsistently between equivalent optical and thermodynamic tests.  156  This may be due to higher thermodynamic cylinder temperatures or changes to in-cylinder mixing from the bowl shape.  Development of the thermo-optical methodology will have to take each of these aspects into consideration in future testing.  6.3 Recommendations The objectives of this investigation are largely addressed through the development of the two optical probes.  However, the work in commissioning the ICOS and pyrometry probe is still ongoing and there is significant potential for new developments.  Open questions still revolve around the calibration of the probes, interpretation of the signals, as well as methodology for thermo-optical work.  Recommendations for the direction of future work are discussed below.  6.3.1 Future ICOS Work Quantifiable ICOS measurements requires calibration of the fuel concentration signal.  The effect of diesel vapor or droplets on the absorption signal is not well understood, and a test plan focused on studying the effect is recommended.  Further to this, the effect of soot on the absorption signal has not been addressed.  The LaVision sensor server records a secondary channel filtered for 3μm that could potentially be used as a “background” signal to correct fuel concentration signals for soot scattering.  Finally, the methodology of selecting reference intensity I0 should be reviewed and implemented in such a way that the effect of residual gases or premixed intake charge on fuel concentration calculations is minimized.  A list of specific suggestions for future work is provided: • Advanced injection of diesel (e.g., -50° aTDC) should be explored to allow diesel premixing.  This may clarify the effect of diesel vapor in DIDF / HPDI strategies with less risk of rapid fouling compared to aligning the injector sprays such that the diesel impinges 157  directly on the sensor.  Depending on the sensitivity of the ICOS to diesel vapor, interpretation of the results may not be as straightforward as declaring the signal to be NG only. • Soot can result in an absorption signal if significant amounts are found in the measurement volume.  LaVision has implemented a background channel that can be used to correct the signal for these effects, and it is recommended that test points with high soot production (i.e., long GPW, LPI or similar) be used to evaluate the signal vs. the background channels.  Alternatively, the ICOS may be used as a local soot concentration sensor rather than fuel concentration signal using the 3μm channel. • The methodology of I0 selection is crucial to quantitative measurements.  DIDF results in high fuel concentration during the intake stroke, and the exhaust stroke has variable residual gas concentration, rendering these two phases inappropriate for I0 determination.  Similarly, HPDI has variable residual gas concentrations in both intake and exhaust strokes.  One solution may be to record motored intensity signals between fired tests as a method of monitoring I0 as well as the fouling on the ICOS probe.  Linear interpolation between subsequent tests can then be used to trace the magnitude of I0 without concern for residual gases or premixed charge.   • The ICOS should be operated in a well-defined fuel concentration mixture to determine the fuel absorption strength and to quantify the results.  Currently, fuel concentration is reported in arbitrary units, and calibrating to a known value will be useful in future comparisons against CFD results or even against the different engine configurations.  Through this exercise, the effects of temperature and pressure on the fuel absorption strength should also be evaluated.  158  6.3.2 Future Pyrometry Probe Work The pyrometry probe is early in its development and has many avenues open for further exploration.  Interpretation of the results depends on numerical solutions for true temperature T and soot concentration KL in eq. (9) and solution criteria must be developed such that solutions for the temperatures and KL are not prone to solver errors (i.e., incorrect roots).  An example of an incorrect root is in Figure 79 for the Long GPW test – the KL signal is shown to increase during the expansion stroke.  Another aspect of the probe that requires additional investigation is its self-cleaning characteristics.  Though the results only show minor fouling over a period of 3 hours of continuous thermodynamic testing, the range of test points and loads considered have been small.  Fouling of the window can significantly affect any temperature/KL measurements, and as such, a reliable calibration method that considers window fouling should be developed.  Once calibration issues have been addressed, future work can consider comparisons against other diagnostic tools such as a fast exhaust nephelometer [58] for tailpipe PM concentration, 2D spatially resolved pyrometry through a Bowditch piston, and even benchtop tests against a well-characterized soot generator.  Below is a list of specific suggestions for future development work on the pyrometry probe:  • The Hottel-Broughton correlation solver criteria can include a lower limit to the soot temperature beyond which the solutions are invalid.  There is currently on-going work in developing best-practices for selecting the proper solution.  Another solution is to implement a third wavelength from which up to three pairs of colours can be used to solve the Hottel-Broughton correlation.  Agreement between the different colour pairs can provide a partial validation method for the temperature and KL. 159  • Calibration of the probe is currently performed before and after a test (i.e., clean and dirty).  This provides some measure of the effect of fouling against the temperature and KL, but is not truly a cycle to cycle calibration.  One option is to implement a stroboscopic light source to provide a known light intensity for each cycle.  However, this would require optical access and would presumably be limited to the optical configuration.  Alternatively, the self-cleaning function can be evaluated over long duration tests in thermodynamic configuration – if the measured intensity does not change significantly after some threshold “fouling” has occurred, it may be beneficial to calibrate the probe under dirty conditions.   • The end KL value from pyrometry measurements may be useful in characterizing cycle-to-cycle engine-out soot emissions.  A fast exhaust PM concentration measurement would be necessary for validation of this, but will provide the link between exhaust-stream emissions and the soot formation and oxidation trends measured by the probe.  To further the understanding between in-cylinder and exhaust stream emissions, 2D spatially resolved pyrometry can be included, in addition to the fast exhaust PM measurement. An increase to the sensitivity of the probe should also be considered through improving the light collection, expanding the FOV, and increasing detector sensitivity.  These improvements can increase the signal to noise ratio at the end of combustion when soot radiation is low. • Expansion of the probe FOV may provide the opportunity to measure a larger area within the cylinder.  This would reduce the effect of localized phenomena, but will require slight modification to the probe.  Options to expand the FOV include shortening the sapphire rod (and accompanying probe sleeve modifications), increasing the diameter of the sapphire rod and probe sleeve opening, and perhaps incorporating a convex shape to the end of the sapphire rod. 160  References [1] Environmental Protection Agency, “Climate Change Indicators in the United States, Fourth Edition,” 2016. [2] Environmental Protection Agency, National Highway Traffic Safety Administration, and Department of Transportation, Greenhouse Gas Emissions and Fuel Efficiency Standards for Medium- and Heavy-Duty Engines and Vehicles - Phase 2. United States, 2016. [3] Environmental Protection Agency, Heavy-Duty Highway Compression-Ignition Engines And Urban Buses: Exhaust Emission Standards. United States, 2016. [4] European Parliament And The Council Of The European Union, “Regulation (EC) No 595/2009 on type-approval of motor vehicles and engines with respect to emissions from heavy duty vehicles (Euro VI) and on access to vehicle repair and maintenance information and amending Regulation (EC) No 715/2007 and Directive 2007,” Off. J. Eur. Union, vol. L 188, pp. 1–13, 2009. [5] D. W. Stanton, “Systematic Development of Highly Efficient and Clean Engines to Meet Future Commercial Vehicle Greenhouse Gas Regulations,” SAE Tech. Pap., no. 2013-01–2421, 2013. [6] National Academies of Sciences Engineering and Medicine, Review of the 21st Century Truck Partnership, Third Report. 2015. [7] J. T. Kashdan and B. Thirouard, “A Comparison of Combustion and Emissions Behaviour in Optical and Metal Single-Cylinder Diesel Engines,” SAE Int. J. Engines, vol. 2, no. 1, pp. 1857–1872, 2009. [8] G. A. Karim, Z. Liu, and W. Jones, “Exhaust Emissions from Dual Fuel Engines at Light Load,” SAE Tech. Pap., no. 932822, 1993. 161  [9] B. B. Sahoo, N. Sahoo, and U. K. Saha, “Effect of engine parameters and type of gaseous fuel on the performance of dual-fuel gas diesel engines-A critical review,” Renew. Sustain. Energy Rev., vol. 13, no. 6–7, pp. 1151–1184, 2009. [10] N. Dronniou, J. Kashdan, B. Lecointe, K. Sauve, and D. Soleri, “Optical Investigation of Dual-fuel CNG/Diesel Combustion Strategies to Reduce CO2 Emissions,” SAE Int. J. Engines, vol. 7, no. 2, pp. 873–887, 2014. [11] J. Rochussen, “Thermodynamic and Optical Investigation of the Combustion Mechanisms of Diesel-Ignited Dual-Fuel Natural Gas Combustion,” The University of British Columbia, 2015. [12] A. P. Carlucci, D. Laforgia, R. Saracino, and G. Toto, “Combustion and emissions control in diesel-methane dual fuel engines: The effects of methane supply method combined with variable in-cylinder charge bulk motion,” Energy Convers. Manag., vol. 52, no. 8–9, pp. 3004–3017, 2011. [13] S. Schlatter, B. Schneider, Y. Wright, and K. Boulouchos, “Experimental Study of Ignition and Combustion Characteristics of a Diesel Pilot Spray in a Lean Premixed Methane/Air Charge using a Rapid Compression Expansion Machine,” SAE Tech. Pap., no. 2012-01–0825, 2012. [14] G. A. Karim, K. Ito, M. Abraham, and L. Jensen, “An Examination of the Role of Formaldehyde in the Ignition Processes of Dual Fuel Engine,” SAE Tech. Pap., no. 912367, 1991. [15] Z. Liu and G. A. Karim, “The ignition delay period in dual fuel engines,” SAE Tech. Pap., no. 950466, 1995. [16] J. Rochussen, J. Yeo, and P. Kirchen, “Effect of Fueling Control Parameters on 162  Combustion and Emissions Characteristics of Diesel-Ignited Methane Dual-Fuel Combustion,” SAE Tech. Pap., no. 2016-01–0792, 2016. [17] J. E. Dec, “A conceptual model of DI diesel combustion based on laser-sheet imaging,” SAE Tech. Pap., no. 970873, 1997. [18] B. Patychuk, N. Wu, G. Mctaggart-cowan, P. Hill, and S. Munshi, “Intake and Exhaust Valve Timing Control on a Heavy-Duty , Direct-Injection Natural Gas Engine,” SAE Tech. Pap., no. 2015-01–0864, 2015. [19] D. R. Tree and K. I. Svensson, “Soot processes in compression ignition engines,” Prog. Energy Combust. Sci., vol. 33, no. 3, pp. 272–309, 2007. [20] P. Ouellette, D. Goudie, and G. McTaggart-Cowan, “Progress in the development of natural gas high pressure direct injection for Euro VI heavy-duty trucks,” in Internationaler Motorenkongress, 2016, pp. 591–607. [21] E. Faghani, P. Kheirkhah, C. W. J. Mabson, P. Kirchen, G. McTaggart-Cowan, and S. N. Rogak, “Effect of Injection Strategies on Emissions from a Pilot-Ignited Direct-Injection Natural-Gas Engine- Part I: Late Post Injection,” SAE Tech. Pap., no. 2017-01–0774, 2017. [22] G. McTaggart-Cowan, K. Mann, J. Huang, A. Singh, B. Patychuk, Z. X. Zheng, and S. Munshi, “Direct Injection of Natural Gas at up to 600 Bar in a Pilot-Ignited Heavy-Duty Engine,” SAE Int. J. Engines, vol. 8, no. 3, pp. 981–996, 2015. [23] D. Goudie, M. Dunn, and S. R. Munshi, “Development of a Compression Ignition Heavy Duty Pilot-Ignited Natural Gas Fuelled Engine for Low NOx Emissions,” SAE Tech. Pap., no. 2004-01–2954, 2004. [24] J. Rochussen, J. Son, J. Yeo, M. Khosravi, P. Kirchen, and G. McTaggart-Cowan, 163  “Development of a Research-Oriented Cylinder Head with Modular Injector Mounting and Access for Multiple In-Cylinder Diagnostics,” SAE Tech. Pap., no. 2017-24–0044, 2017. [25] S. L. Kokjohn, R. D. Reitz, D. Splitter, and M. Musculus, “Investigation of Fuel Reactivity Stratification for Controlling PCI Heat-Release Rates Using High-Speed Chemiluminescence Imaging and Fuel Tracer Fluorescence,” SAE Int. J. Engines, vol. 5, no. 2, pp. 248–269, 2012. [26] M. Khosravi, J. Rochussen, J. Yeo, P. Kirchen, G. McTaggart-Cowan, and N. Wu, “Effect of Fuelling Control Parameters on Combustion Characteristics of Diesel-Ignited Natural Gas Dual-Fuel Combustion in an Optical Engine,” in ASME 2016 Internal Combustion Engine Division Fall Technical Conference, 2016, pp. 1–10. [27] M. P. B. Musculus, “Multiple Simultaneous Optical Diagnostic Imaging of Early-Injection Low-Temperature Combustion in a Heavy-Duty Diesel Engine,” SAE Tech. Pap., no. 2006-01–0079, 2006. [28] Y. Matsui, T. Kamimoto, and S. Matsuoka, “A Study on the Time and Space Resolved Measurement of Flame Temperature and Soot Concentration in a D. I. Diesel Engine by the Two-Color Method,” SAE Tech. Pap., no. 790491, 1979. [29] J. Yan and G. L. Borman, “Analysis and In-Cylinder Measurement of Particulate Radiant Emissions and Temperature in a Direct Injection Diesel Engine,” SAE Tech. Pap., no. 881315, 1988. [30] J. J. Lopez, J. Martin, A. Garcia, D. Villalta, A. Warey, and V. Domenech, “Characterization of In-Cylinder Soot Oxidation Using Two- Color Pyrometry in a Production Light-Duty Diesel Engine,” SAE Tech. Pap., no. 2016-01–0735, 2016. 164  [31] P. Vögelin, P. Obrecht, and K. Boulouchos, “Experimental Investigation of Multi-In-Cylinder Pyrometer Measurements and Exhaust Soot Emissions Under Steady and Transient Operation of a Heavy-Duty Diesel Engine,” SAE Int. J. Engines, vol. 6, no. 3, pp. 1855–1865, 2013. [32] P. Kirchen, P. Obrecht, K. Boulouchos, and A. Bertola, “Exhaust-Stream and In-Cylinder Measurements and Analysis of the Soot Emissions From a Common Rail Diesel Engine Using Two Fuels,” J. Eng. Gas Turbines Power, vol. 132, no. 11, p. 112804, 2010. [33] S. Kunte, A. Bertola, P. Obrecht, and K. Boulouchos, “Temporal soot evolution and diesel engine combustion: influence of fuel composition, injection parameters, and exhaust gas recirculation,” Int. J. Engine Res., vol. 7, no. 6, pp. 459–470, 2006. [34] H. Ma, R. Stevens, and R. Stone, “In-Cylinder Temperature Estimation from an Optical Spray-Guided DISI Engine with Color-Ratio Pyrometry (CRP),” SAE Tech. Pap., no. 2006-01–1198, 2006. [35] K. Zha, X. Yu, R. Florea, and M. Jansons, “Impact of Biodiesel Blends on In-cylinder Soot Temperature and Concentrations in a Small-Bore Optical Diesel Engine,” SAE Tech. Pap., no. 2012-01–1311, 2012. [36] H. Dembinski and H.-E. Angstrom, “Swirl and Injection Pressure Impact on After-Oxidation in Diesel Combustion, Examined with Simultaneous Combustion Image Velocimetry and Two Colour Optical Method,” SAE Tech. Pap., no. 2013-01–0913, 2013. [37] J. Vattulainen, V. Nummela, R. Hernberg, and J. Kytölä, “A system for quantitative imaging diagnostics and its application to pyrometric in-cylinder flame-temperature measurements in large diesel engines,” Meas. Sci. Technol., vol. 11, no. 2, pp. 103–119, 2000. 165  [38] Q. Shi, T. Li, X. Zhang, B. Wang, and M. Zheng, “Measurement of Temperature and Soot ( KL ) Distributions in Spray Flames of Diesel-Butanol Blends by Two-Color Method Using High-Speed RGB Video Camera Theory of Two-Color Method,” SAE Tech. Pap., no. 2016-01–2190, 2016. [39] A. Magno, E. Mancaruso, and B. M. Vaglieco, “Endoscopic Investigation of Combustion Process in a Small Compression Ignition Engine Fuelled with Rapeseed Methyl Ester,” SAE Tech. Pap., no. 2014-01–2649, 2014. [40] S. Skeen, J. Manin, L. Pickett, K. Dalen, and A. Ivarsson, “Quantitative Spatially Resolved Measurements of Total Radiation in High-Pressure Spray Flames,” SAE Tech. Pap., no. 2014-01–1252, 2014. [41] W. Jing, W. Roberts, and T. Fang, “Comparison of Soot Formation For Diesel and Jet-A in a Constant Volume Combustion Chamber Using Two- Color Pyrometry,” SAE Tech. Pap., no. 2014-01–1251, 2014. [42] A. Grosch, V. Beushausen, O. Thiele, and R. Grzeszik, “Crank Angle Resolved Determination of Fuel Concentration and Air/Fuel Ratio in a SI-Internal Combustion Engine Using a Modified Optical Spark Plug,” SAE Tech. Pap., no. 2007-01–0644, 2007. [43] F.-Q. Zhao, M. Taketomi, K. Nishida, and H. Hiroyasu, “Quantitative Imaging of the Fuel Concentration in a SI Engine with Laser Rayleigh Scattering,” SAE Tech. Pap., no. 932641, 1993. [44] J. A. Snyder, R. K. Hanson, R. P. Fitzgerald, and R. R. Steeper, “Dual-Wavelength PLIF Measurements of Temperature and Composition in an Optical HCCI Engine with Negative Valve Overlap,” SAE Int. J. Engines, vol. 2, no. 1, 2009. [45] M. H. Koenig and M. J. Hall, “Cycle-Resolved Measurements of Pre-Combustion Fuel 166  Concentration Near the Spark Plug in a Gasoline SI Engine,” SAE Tech. Pap., no. 981053, 1998. [46] J. Yeo, J. Rochussen, and P. Kirchen, “Application of an In-Cylinder Local Infrared Absorption Fuel Concentration Sensor in a Diesel-Ignited Dual-Fuel Engine,” SAE Tech. Pap., no. 2016-01–2310, 2016. [47] D. D. Brehob and D. B. Kittelson, “A Review of CI Engine In-Cylinder Diagnostics for the Investigation of Soot Loading, Chemical Composition, and Temperature,” SAE Tech. Pap., no. 880515, 1988. [48] H. C. Hottel and F. P. Broughton, “Determination of True Temperature and Total Radiation from Luminous Gas Flames,” Ind. Eng. Chem., vol. 4, no. 2, pp. 166–175, 1932. [49] N. Ladommatos and H. Zhao, “A Guide to Measurement of Flame Temperature and Soot Concentration in Diesel Engines Using the Two-Colour Method Part I: Principles,” SAE Tech. Pap., no. 941956, 1994. [50] N. Ladommatos and H. Zhao, “A Guide to Measurement of Flame Temperature and Soot Concentration in Diesel Engines Using the Two-Colour Method Part 2: Implementation,” SAE Tech. Pap., no. 941957, 1994. [51] L. S. Rothman et al., “The HITRAN2012 molecular spectroscopic database,” J. Quant. Spectrosc. Radiat. Transf., vol. 130, pp. 4–50, 2013. [52] L. A. Dombrovsky, S. S. Sazhin, S. V. Mikhalovsky, R. Wood, and M. R. Heikal, “Spectral properties of diesel fuel droplets,” Fuel, vol. 82, no. 1, pp. 15–22, 2003. [53] P. Dirrenberger, H. Le Gall, R. Bounaceur, O. Herbinet, P.-A. Glaude, A. Konnov, and F. Battin-Leclerc, “Measurements of Laminar Flame Velocity for Components of Natural Gas,” Energy and Fuels, vol. 25, no. 9, pp. 3875–3884, 2011. 167  [54] J. Rochussen, “Thermodynamic and Optical Investigation of the Combustion Mechanisms of Diesel-Ignited Dual-Fuel Natural Gas Combustion,” University of British Columbia, 2015. [55] E. Faghani, “Effect of Injection Strategies on Particulate Matter Emissions From HPDI Natural-Gas Engines,” University of British Columbia, 2015. [56] J. Yan, “Analyses and In-Cylinder Measurements of Local and Hemispherical Particulate Radiant Emissions and Temperatures in a Direct Injection Diesel Engine,” The University of Wisconsin-Madison, 1988. [57] J. B. Heywood, Internal Combustion Engine Fundementals, vol. 21. 1988. [58] P. Kheirkhah, P. Kirchen, and S. Rogak, “Fast Exhaust Nephelometer ( FEN ): A New Instrument for Measuring Cycle-Resolved Engine Particulate Emission,” SAE Tech. Pap., no. 2016-01–2329, 2016.  168  Appendices Appendix A  List of Signals Recorded by NI CAS (Data Acquisition cDAQ-9188 Chassis) Signal Name NI Module Measurement Type Functionality PT_CYL_1 NI-9215 Crank Synchronous In-cylinder pressure transducer MAP_IN_1 NI-9215 Crank Synchronous Intake manifold absolute pressure REC_POS NI-9215 Crank Synchronous 5V TTL signal that signifies the start of imaging from the Photron high-speed camera 700_PROBE NI-9215 Crank Synchronous Pyrometry probe 700nm voltage signal 800_PROBE NI-9215 Crank Synchronous Pyrometry probe 800nm voltage signal 45_FEN NI-9215 Crank Synchronous Fast exhaust nephelometer 45° scattering voltage 135_FEN NI-9215 Crank Synchronous Fast exhaust nephelometer 135° scattering voltage MAF_IN_1 NI-9205 Medium Speed Analog Intake manifold mass air flow PT_DSL_3 NI-9205 Medium Speed Analog Diesel pre-supply pressure PT_EX_1 NI-9205 Medium Speed Analog Exhaust backpressure PT_OIL_1 NI-9205 Medium Speed Analog Engine oil pressure at distribution manifold PT_CNG_1 NI-9205 Medium Speed Analog CNG delivery pressure (DIDF) PT_CNG_2 NI-9205 Medium Speed Analog CNG rail pressure (DIDF) COR_CNG_1 NI-9205 Medium Speed Analog Compressed natural gas flowrate PT_DSL_4 NI-9205 Medium Speed Analog Diesel rail pressure (HPDI) PT_CNG_3 NI-9205 Medium Speed Analog CNG delivery pressure (HPDI) PT_CNG_5 NI-9205 Medium Speed Analog CNG rail pressure (HPDI) TQ_DYNO_1 NI-9205 Medium Speed Analog Dynamometer torque readout  169  Signal Name NI Module Measurement Type Functionality TC_OIL_1 NI-9205 Medium Speed Analog Engine oil temperature TC_IN_1 NI-9213 Thermocouple Intake manifold temperature TC_EX_1 NI-9213 Thermocouple Exhaust manifold temperature TC_TMS_1 NI-9213 Thermocouple Thermal management system temperature (coolant temp.) TC_SAM_1 NI-9213 Thermocouple AVL heated sampling line temperature TC_CNG_1 NI-9213 Thermocouple CNG fuel board temperature TC_CNG_2 NI-9213 Thermocouple CNG rail temperature (DIDF) TC_DSL_1 NI-9213 Thermocouple Diesel fuel temperature TC_CNG_3 NI-9213 Thermocouple CNG rail temperature (HPDI)   170  Appendix B  FPGA SubVIs and Functionality SubVI Name Functionality Engine Position Tracker (EPT) Takes encoder inputs from the AD Combination Module and converts it to engine position information for use in other subVIs.  Inputs include 2 hall-effect sensors (100-tooth flywheel encoder, 2-tooth camshaft encoder) and an index from a crankshaft mounted encoder.  The combination of these three signals provide information on crank-angle position relative to a fixed reference (index or top dead center), and whether it is a firing cycle or non-firing. Direct Injector User inputs of injector timing and position is combined with EPT output to actuate the direct injector(s).  Fixed inputs are required based on injector details – namely the voltage drive and current drive waveforms that are desired for actuating the injector needle.  For HPDI, the drive voltage is 50V whereas for common rail, the drive voltage is 75V.  Current waveforms (peak-hold) are specified by the OEMs. Port Fuel Injector & Low Side Driver User inputs of injector timing and position is combined with EPT output to actuate the port fuel injector.  Fixed inputs are required based on injector details – namely the voltage drive and current drive waveforms that are desired for actuating the injector needle.  In addition, the module provides functionality for diesel common-rail solenoids for controlling rail delivery pressure. ETTL Module User inputs for specific camera trigger positions and durations used to trigger optical measurement equipment.     171  Appendix C  Operating PC Virtual Instrument Descriptors  Figure C.1 - Emissions (AVL bench) and fuel measurement logging screens.  Figure C.2 - CAS live sensor values.  These signals are communicated directly from CAS and are only for display purposes.  The green indicators change to yellow or flash when the process value has exceeded its critical threshold. 172   Figure C.3 - Skip firing sequence parameter controls.  Figure C.4 - ECU control parameters.  Diesel common rail, master injection controls, O2 wideband, and PFI controls are shown. 173   Figure C.5 - Single cycle controls within Sequence Parameters array.   Table C.1 - Operator adjustable parameters for ECU operation.   Operating Parameter Description Sequence Parameters • An array that contains controls for camera triggers, port fuel injection, and direct injection events • Used in combination with Total Sequences and Sequence Length parameters to specify an engine firing/operating pattern Injector Timing (common rail, PFI, HPDI gas, HPDI diesel)  • Within Sequence Parameters • Specifies the commanded start of injection (CSOI) for the injector and is set by engine position aTDC. • Up to 5 injections per cycle is possible; software currently limits maximum injections per cycle to 2 Injector Duration (common rail, PFI, HPDI gas, HPDI diesel) • Within Sequence Parameters • Specifies the duration in milliseconds that the injector needle is commanded open   174  Operating Parameter Description Injector Enable • Within Sequence Parameters • Enables each specific injection event • DSL_Inj1/2, CNG_Inj1/2 are direct injection events corresponding to first or second injection within a cycle • PFI1/2 Enable are port fuel injection events corresponding to first or second injector in the manifold; timing and duration are not-cycle specific as with direct injection events • Injection Master Enable, PFI Enable, and Enable Camera & Fuel Injection commands must be on to allow injection per the specified operation pattern Camera Trigger Timing • Specifies the camera start trigger and is set by engine crank position aTDC • CameraTrigStart is tied to the Sequence Parameters Camera Trigger button • Camera2TrigStart operates continuously if the software detects that the engine is motoring and is currently used to provide a TDC reference for the Internal Combustion Optical Sensor Total Sequences • Specifies the total number of repetitions of the Sequence Parameter array pattern Sequence Length • Specifies the length of the array pattern to follow within the Sequence Parameters array • The length corresponds to number of cycles IMV / Rail Valve Enable • Controls the inlet metering valve and common rail valve • The IMV is set to a fixed duty cycle PWM value • The common rail valve is controlled by a software PID controller to achieve the common rail pressure set by the user Rail Pressure Set • Specifies the desired diesel common rail set pressure; PID control is used to determine the rail pressure. Continuous Mode • Specifies if engine operation is to be continuous or to follow a user specified skip firing pattern • Under continuous mode, the software loops a single cycle (i.e., only the first cycle in the Sequence Parameter array is considered) PFI Controls • Toggles whether the port fuel injectors are enabled; PFI1/2 Enable, Injection Master Enable, and Enable Camera & Fuel Injection commands must be on to allow injection per the specified operation pattern   175  Operating Parameter Description Injection Master Enable • A injection specific master control  • Mostly used to disable injectors when testing cameras prior to an optical test HPDI / Diesel Toggle • Sets the appropriate direct injector drive voltage and current waveform depending on common rail or HPDI operation • The appropriate injectors must be connected to the direct injector drivers when swapping between common rail and HPDI UEGO1 Enable • Turns on the exhaust wideband O2 sensor Sim Enable EPT Error Clear • Part of the engine position tracking functionality • Sim Enable artificially sends engine position signals and can be used for testing cameras or injectors off-engine • EPT Error Clear is used to clear engine position tracking faults Enable Camera & Fuel Injection • Master control for all injection and camera triggers and is used to start engine operation   176  Appendix D  Safety Process Values Table D.1 - Engine process values and their respective critical process values. Process Values Critical Value Description Action Intake Temperature >200°C Intake heater can heat up the intake higher than 100°C.  Higher than this may indicate a system fire. Shut down any fueling activity and flash warning. Oil Temperature >110°C Protect against over-temperature in the oil. Shut down any fueling activity and flash warning. Oil Pressure <18 psi Low oil pressure can indicate poor lubrication flow around the engine. Shut down any fueling activity and flash warning. Exhaust Temperature >900°C Protect exhaust system against extreme high temperatures. Shut down any fueling activity and flash warning. Exhaust Back Pressure >30 psi High back pressure may indicate a blockage in the exhaust. Shut down any fueling activity and flash warning. Diesel Pre-Supply Pressure <5 psi Low pre-supply pressure indicates significant vacuum being drawn by the CP3 high pressure pump. Shut off common rail solenoids and flash warning. In-Cylinder Pressure >9500 kPa 9.5 bar maximum cylinder pressure set to protect engine head.  Can be adjusted as experiments evolve to higher loads. Shut down any fueling activity and flash warning. Common Rail Pressure >1850 bar Maximum pressure the common rail is rated for. Shut off common rail solenoids and flash warning. CNG Rail Pressure (DIDF) >250 psi May indicate a pressure regulator failure upstream and risk injecting higher than expected fuel quantities into the intake. Shut down any fueling activity and flash warning. Coriolis CNG Flow Rate >4 kg/hr  (>100 kg/hr) May indicate a leak in the CNG fuel board.  Optical HPDI operation can trigger surges in the fuel board and an artificially high setting is required for the short test. Shut down any fueling activity and flash warning.  177  Process Values Critical Value Description Action CNG Rail Pressure (HPDI) >350 bar May indicate a pressure regulator failure upstream and risk injecting higher than expected fuel quantities into the engine. Shut down any fueling activity and flash warning. Diesel Rail Pressure (HPDI) >350 bar May indicate a pressure regulator failure upstream and risk injecting higher than expected fuel quantities into the engine. Shut down any fueling activity and flash warning.   178  Appendix E  Calculation of Mixture Heat Capacity Ratio The mixture heat capacity ratio was calculated for each DIDF test point and considers O2, N2, and CH4 species only.  Equivalence ratio data from each test point was used to calculate the relative molar fraction for each species.  Cp is calculated at each point and then converted to κ heat capacity ratio.  The MATLAB  code used to calculate the κ value for each test point is shown below:  % uses NASA / CHEMKIN database for Cp polynomials % initialize variables Ru = 8.3144598; % [kJ/K/kmol] T1 = 776; % [k] calculated based on isentropic compression from intake conditions   % Cp coefficients O2 o2_a1 = 3.78245636; o2_a2 = -2.99673415E-3; o2_a3 = 9.84730200E-6; o2_a4 = -9.68129508E-9; o2_a5 = 3.24372836E-12; o2_molmass = 32; % [g/mol] o2_cp = (o2_a1+o2_a2*T1+o2_a3*T1^2+o2_a4*T1^3+o2_a5*T1^4)*Ru / o2_molmass;... % [kJ/kg/K]   % Cp coefficients N2 n2_a1 = 3.53100528; n2_a2 = -1.23660987E-4; n2_a3 = -5.02999437E-7; n2_a4 = 2.43530612E-9; n2_a5 = -1.40881235E-12; n2_molmass = 28; % [g/mol] n2_cp = (n2_a1+n2_a2*T1+n2_a3*T1^2+n2_a4*T1^3+n2_a5*T1^4)*Ru / n2_molmass;... % [kJ/kg/K]   % Cp coefficients CH4 ch4_a1 = 5.14987613; ch4_a2 = -1.36709788E-2; ch4_a3 = 4.91800599E-5; ch4_a4 = -4.84743026E-8; ch4_a5 = 1.66693956E-11; ch4_molmass = 16; % [g/mol] ch4_cp = (ch4_a1+ch4_a2*T1+ch4_a3*T1^2+ch4_a4*T1^3+ch4_a5*T1^4)*Ru /... ch4_molmass; % [kJ/kg/K]   179  for i = 1:NumberDFPoints     % calculates the mixed cp     N_o2 = 2*output(i).Phi.CNG_lambda; % number of moles o2 (EQR)     N_n2 = 2*3.76*output(i).Phi.CNG_lambda; % number of moles n2 (EQR)     N_ch4 = 1; % number of moles ch4     N_total = N_o2 + N_n2 + N_ch4;     output(i).mix.cp = N_ch4/N_total*ch4_cp + N_o2/N_total*o2_cp + N_n2/N_total*n2_cp; % [kJ/kg/K]          % calculates the mixed molecular weights, mixed R, mixed cv, and k     output(i).mix.M = N_ch4/N_total*ch4_molmass + N_o2/N_total*o2_molmass + N_n2/N_total*n2_molmass; % [kg/kmol]     output(i).mix.R = Ru / output(i).mix.M; % [kJ/kg/K]     output(i).mix.cv = output(i).mix.cp - output(i).mix.R;     output(i).mix.k = output(i).mix.cp / output(i).mix.cv; % k value for use in ideal gas relations end   180  Appendix F  Geometry Information for Calculating Mass Flow Across a Restriction  Figure F.1 – Pyrometry probe geometry.  Values are displayed in inches.   181  Appendix G  Pyrometry Probe Calibration Best Practices This appendix lists some of the best practices for consistent calibration of the pyrometry probe using the Labsphere.  These steps are based on the preliminary investigation documented by this thesis and are by no means the only way to calibrate the probe.  It is highly suggested that the reader apply these steps with consideration to any new changes or improvements that may be possible to the current setup. 1. Maintain the same photodetector gain settings as appropriate to the testing that is being performed.  Generally, diesel-only combustion is significantly brighter, and it may be necessary to reduce gain to avoid saturating the photodetectors.  In any case, the same gain must be used between corresponding calibration and test pairs. 2. Maintain standoff distances between the Labsphere output and the probe receiving end (Figure 74).  In this thesis, it was maintained at 23mm between the front of the Labsphere (opening ring) and the front of the probe sleeve.  This is done as the probe appears to be sensitive to the distance away from the Labsphere.  It is not clear whether this is due to non-uniform output by the sphere, or if the probe’s optics are playing a role. 3. The geometry (i.e., spacing) between the collimating lens, narrow bandpass filter, and photodetector diode should be maintained.  The collimation is not perfect, and so it is possible that rotation of the fiber or collimator may result in some of the light spilling over the detectable area of the photodiode.  The fibers should be threaded in the same orientation to the collimating lens and photodetector between calibration and test.  

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.24.1-0354490/manifest

Comment

Related Items