Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Numerical prediction and pyrometric imaging of a hot-surface ignition-assist application Zepeda Gutiérrez, René 2019

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

Item Metadata

Download

Media
24-ubc_2019_may_zepedagutierrez_rene.pdf [ 4.81MB ]
Metadata
JSON: 24-1.0378384.json
JSON-LD: 24-1.0378384-ld.json
RDF/XML (Pretty): 24-1.0378384-rdf.xml
RDF/JSON: 24-1.0378384-rdf.json
Turtle: 24-1.0378384-turtle.txt
N-Triples: 24-1.0378384-rdf-ntriples.txt
Original Record: 24-1.0378384-source.json
Full Text
24-1.0378384-fulltext.txt
Citation
24-1.0378384.ris

Full Text

NUMERICAL PREDICTION AND PYROMETRIC IMAGING OF A HOT-SURFACE IGNITION-ASSIST APPLICATION by  René Zepeda Gutiérrez  B.Eng., Instituto Tecnológico y de Estudios Superiores de Occidente, 2014  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)  April 2019  © René Zepeda Gutiérrez, 2019  ii  The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, a thesis/dissertation entitled:  Numerical prediction and pyrometric imaging of a hot-surface ignition-assist application  submitted by René Zepeda Gutiérrez in partial fulfillment of the requirements for the degree of Master of Applied Science in Mechanical Engineering  Examining Committee: Steven Rogak, Mechanical Engineering  Supervisor  Patrick Kirchen, Mechanical Engineering Supervisory Committee Member   Supervisory Committee Member Gordon McTaggart-Cowan, Engine Applications R&D Lead, Westport Fuel Systems, Inc. Additional Examiner   Additional Supervisory Committee Members:  Supervisory Committee Member  Supervisory Committee Member iii  Abstract The combustion of natural gas is an interesting alternative to liquid fossil fuels due to its competitive price and lower CO2 emissions. One technique to burn natural gas inside direct-injection engines is the hot-surface ignition-assist method. The natural gas jet impinges on the hot surface, which acts as an ignition source. As a constant high temperature is required in the hot surface to have quick and consistent ignition events, a numerical prediction of the temperature of an application of the hot surface technique was done, and a method to study the temperature of the hot surface was developed.  One proposed application of the hot-surface method consists in a fuel injector equipped with a heater ring. The high temperature of the heater could produce an excessive temperature in the injector, affecting its functioning or the fuel. To study this, heat transfer simulations were performed. A sensitivity analysis revealed a large effect of the coolant temperature in the temperature of the injector, and a large effect of the input power and surface emissivity on the temperature of the heater. With an input power of 100 W, the injector temperature is expected to remain within 200 °C, while the heater reaches a temperature beyond 1300 °C.  During the injection of the fuel, the hot surface experiences a rapid cooling event. This can affect the ignition delay, hampering the combustion efficiency. To study this, a pyrometer method with high spatial and temporal resolution was developed. A hot surface was subject to a series of cooling jets. The pyrometer method revealed that the jets with a direct orientation produced a larger temperature drop compared to jets with a side orientation, regardless of their pressure or duration. The thermometer method developed has the potential to be used in different applications where rapid changes in temperature are expected, allowing to calculate the temperature of a surface in transient-state using a digital camera and the radiation intensity at a single wavelength.   iv  Lay Summary The hot-surface ignition-assist method is a technology under development to burn natural gas inside direct-injection engines, which could help to maximize its benefits and increase its use. The ignition happens by impinging the hot surface with fuel jets, which can decrease the temperature of the surface, affecting the ignition reliability and timing. An application of this technology consists in a fuel injector equipped with a heater ring. As the high temperature required in the heater could affect the injector and fuel, a series of heat transfer simulations were done. To analyze the cooling effect of the fuel jets in a hot surface, a remote temperature measurement method was developed. This method was used to assess the different cooling effects that gaseous jets produced on a heater ring, showing its potential to be used in other applications.   v  Preface The research objective of this thesis was outlined by Steven Rogak and Patrick Kirchen from The University of British Columbia, and Gordon McTaggart-Cowan from Westport Fuel Systems. Steven Rogak and Patrick Kirchen also provided close guidance for the experiments and analyses presented in this thesis. I also received feedback and different contributions from various members of Westport, including Sandeep Munshi, Hamed Karimi Sharif, Ashish Singh, Gavin Hartnett, Riley Cahill and, specially, Gordon McTaggart-Cowan. The heater ring and insulation of the test object used in section 3.3.1 were installed by the UBC Materials Engineering department. Westport designed and conducted the experimental tests of the HSI injector prototype shown in section 3.3.2 and provided me with the design of the injector prototype for its computational modelling. The rest of experiments from Chapter 3 were designed, built and conducted by me. All the modelling and heat transfer simulations from Chapter 3 were done by me using ANSYS Icepak. The pyrometer method discussed in Chapter 4 was developed by me, Steven Rogak and Patrick Kirchen. The MATLAB code used in the pyrometer method was based on a code wrote by Mahdiar Khosravi. I wrote the final version of the code used. Mahdiar also provided guidance for the data collection needed for the pyrometer method. The test object used in section 4.5 was designed and built by Westport and the UBC Materials Engineering department. The 0.7 mm nozzle for the cooling tests was built by Riley Cahill. The rest of components and test objects used for the pyrometer tests were designed and machined/built by me.  Between December 2017 and May 2018 I participated in an internship at Westport as part of the Mitacs Accelerate program. The project of the internship was entitled “Emissions control and reduction for natural gas engines”. vi  Table of Contents Abstract ......................................................................................................................................... iii Lay Summary ............................................................................................................................... iv Preface .............................................................................................................................................v Table of Contents ......................................................................................................................... vi List of Tables ................................................................................................................................ ix List of Figures .................................................................................................................................x List of Symbols ........................................................................................................................... xiii List of Abbreviations ...................................................................................................................xv Acknowledgements .................................................................................................................... xvi Dedication .................................................................................................................................. xvii Chapter 1: Motivation and Objectives .........................................................................................1 1.1 Introduction ..................................................................................................................... 1 1.2 Research motivation........................................................................................................ 2 1.3 Research objectives and thesis structure ......................................................................... 3 Chapter 2: Background Information ...........................................................................................5 2.1 Introduction ..................................................................................................................... 5 2.2 Use of glow plugs to ignite natural gas inside direct injection engines .......................... 6 2.3 Westport’s Hot-surface ignition fuel injector ................................................................. 7 2.4 Two-colour pyrometer method ....................................................................................... 9 Chapter 3: Modeling of the Heat Transfer in the Hot-Surface Injector ................................12 3.1 Introduction ................................................................................................................... 12 3.2 Methods......................................................................................................................... 13 vii  3.2.1 CFD simulations using Ansys Icepak ....................................................................... 13 3.2.2 Comparison of results between experimental tests and simulations ......................... 15 3.3 Experimental tests to develop and validate the injector heat transfer simulation ......... 16 3.3.1 Stage one of validation tests: Heater ring ................................................................. 16 3.3.2 Stage two of validation tests: HSI injector under combustion vessel conditions ..... 25 3.4 Heat transfer simulation of the HSI injector ................................................................. 37 3.4.1 Temperature prediction of the HSI injector under combustion engine conditions ... 37 3.4.2 Sensitivity analysis of the main HSI injector components ....................................... 45 3.5 Design recommendations .............................................................................................. 49 Chapter 4: Development and Application of a Pyrometer Method to Measure the Surface Temperature of a Hot Surface ....................................................................................................52 4.1 Introduction ................................................................................................................... 52 4.2 Pyrometry theory .......................................................................................................... 53 4.3 Apparatus and procedure .............................................................................................. 56 4.4 Development and validation of the hybrid pyrometer method ..................................... 58 4.4.1 Two-colour thermal imaging .................................................................................... 59 4.4.2 Single-colour thermal imaging ................................................................................. 66 4.5 Temperature measurement of the HSI heater ring ........................................................ 72 Chapter 5: Main Conclusions, limitations and Future Work ..................................................91 5.1 Summary and conclusions of Chapter 3 ....................................................................... 91 5.2 Summary and conclusions of Chapter 4 ....................................................................... 93 5.3 Future work ................................................................................................................... 96 Bibliography ...............................................................................................................................100 viii  Appendices ..................................................................................................................................105 Appendix A Steps followed to generate the heat transfer models in Icepak .......................... 105 Appendix B Preliminary stage one validation tests: Heat transfer using a cartridge heater ... 107 Appendix C Preliminary stage two validation tests: Main body insulation ............................ 113  ix  List of Tables Table 3.1. Materials’ Properties Used in the Model of Stage One. .............................................. 19 Table 3.2. Parameter Values Used to Generate the Simulations of Stage One. ........................... 20 Table 3.3. Temperatures Recorded in the Experimental Tests of Stage Two. .............................. 27 Table 3.4. Materials’ Properties Used in the Model of Stage Two............................................... 30 Table 3.5. Materials Used for the Added Components in the Injector Model of the Combustion Engine Simulations. ...................................................................................................................... 39 Table 3.6. Results from the Combustion Engine Simulation. ....................................................... 44 Table 3.7. Variables Adjusted for the Sensitivity Analysis of the HSI Injector Model ............... 46 Table 3.8. Dimensionless Sensitivity of the HSI Injector Model Variables on the Temperature Study Points. ................................................................................................................................. 47 Table 4.1. Variables Tested in the HSI Cooling Tests. ................................................................. 80 Table 4.2. Mass Expelled According to Injection Pressure and Duration. ................................... 80  x  List of Figures Figure 3.1. Multiview Projection of the SS Object Used for Stage One. ..................................... 17 Figure 3.2. Detail of the Heater Embedded in the Insulation Layer Used for Stage One. ............ 18 Figure 3.3. Cross-section View of the Change in the Heater Ring Profile. .................................. 21 Figure 3.4. Views of the Mesh for Model of Stage One Focused on the SS Object .................... 22 Figure 3.5. Temperature Map of the Stage One Simulation, Nominal Case. ............................... 23 Figure 3.6. Comparison of Results between Experiment and Simulation of Stage One. ............. 24 Figure 3.7. Detail of the HSI Injector Tip Design Used for Stage Two. ...................................... 26 Figure 3.8. Cross-section View of the Test Object Used for Stage Two. ..................................... 28 Figure 3.9. Cross-section View with the Dimensions of the Test Object Used for Stage Two. ... 29 Figure 3.10. Detail of the Heater Ring Model Used for the Simulation of Stage Two. ............... 31 Figure 3.11. Views of the Mesh for the Model of Stage Two Focused on the HSI Tip. .............. 32 Figure 3.12. Temperature Map of the Stage Two Simulation, 38.5 W Nominal Case. ................ 33 Figure 3.13. Charts with the Comparison of Results between Experiments and Simulations of Stage Two. .................................................................................................................................... 35 Figure 3.14. Cross-section View of the HSI Injector Model Used for the Combustion Engine Simulations. .................................................................................................................................. 38 Figure 3.15. Surface Conditions Used in the Model of the Combustion Engine Simulations. .... 40 Figure 3.16. Views of the Mesh for the Model of the Combustion Engine Simulations Focused on the HSI tip. ............................................................................................................................... 42 Figure 3.17. Temperature Map of the Combustion Engine Simulation. ....................................... 43 Figure 4.1. Integrating Sphere Used for the Calibration Process. ................................................ 57 Figure 4.2. Flowchart of the Hybrid Pyrometer Method .............................................................. 59 xi  Figure 4.3. Multiview Projection of the SS Prism Object. ........................................................... 60 Figure 4.4. Test Rig for the SS Prism Tests. ................................................................................. 61 Figure 4.5. Raw Image of the SS Prism Taken by the Camera .................................................... 62 Figure 4.6. Temperature Map of the SS Prism when Applying a Voltage of 115 V to the Heater....................................................................................................................................................... 63 Figure 4.7. Test Rig Setup of the SS Prism Tests Using a Commercial Pyrometer ..................... 63 Figure 4.8. Results Comparison of the SS Prism Tests. ............................................................... 65 Figure 4.9. Plot of Temperature vs Radiation Intensity Calculated from the Results of the 115 V SS Tests Case ................................................................................................................................ 67 Figure 4.10. Plot of Temperature vs Radiation Intensity According to Pixel Location ............... 68 Figure 4.11. Temperature Maps Comparison when Applying a Voltage of 115 V in the SS Prism Tests. ............................................................................................................................................. 70 Figure 4.12. Histograms with the Comparison of the SS prism Temperature Maps. ................... 71 Figure 4.13. Test Rig of the HSI tip on the Rotary Table. ............................................................ 73 Figure 4.14. Two-colour Method Temperature Maps for the 17 V and 20 V HSI Test Cases. .... 74 Figure 4.15. Plot of Temperature vs Radiation Intensity Calculated from the HSI Tests Results........................................................................................................................................................ 74 Figure 4.16. Temperature Map of the HSI Heater at Various Rotation Angles. .......................... 75 Figure 4.17. Test Rig Detail for the HSI Cooling Tests. .............................................................. 77 Figure 4.18. P&ID Drawing of Components Connected to the Solenoid Valve. ......................... 78 Figure 4.19. Top-view of the Nozzle Orientations Tested in the HSI Cooling Tests. .................. 79 Figure 4.20. Temperature Map of the HSI heater during a Cooling Event Using the Direct Nozzle Orientation. ................................................................................................................................... 81 xii  Figure 4.21. Temperature Map of the HSI Heater during a Cooling Event Using the Side Nozzle Orientation. ................................................................................................................................... 82 Figure 4.22. Region of the HSI Heater Used for the Temperature Plots. ..................................... 83 Figure 4.23. Maximum and Minimum Temperatures from the "40 psi, 60 ms, Direct" case. ..... 84 Figure 4.24. Maximum and Minimum Temperatures from the "90 psi, 95 ms, Side" case. ........ 84 Figure 4.25. Comparison of Direct and Side Orientation Cases Using 40 psi as Injection Pressure. ........................................................................................................................................ 86 Figure 4.26. Comparison of Injection Pressures Using the Direct Nozzle Orientation ................ 87 Figure 4.27. Comparison of Injection Durations Using 90 psi as Injection Pressure ................... 88  xiii  List of Symbols CO2 Carbon dioxide  b Black body C1 Constant used for Plank’s law (3.74177 ✕ 108 W·μm4 /m2) C2 Constant used for Plank’s law (1.43878 ✕ 104 μm·K) ∇ Nabla operator  E Emissive power F External forces and source terms h Sensible enthalpy I Radiation intensity i Input IT Unit tensor k Thermal conductivity m Mass NOX Nitrogen oxides o Output P Pressure P1 Initial pressure P2 Final pressure RS Nitrogen gas constant SiC Silicon carbide T Temperature t Time xiv  Ta Apparent temperature UB Upper bound LB Lower bound V Volume v Velocity ZrO2 Zirconium dioxide (Zirconia) ελ Spectral hemispherical emissivity λ Wavelength μ Dynamic viscosity ρ Density ρg gravitational body force τ Stress tensor xv  List of Abbreviations AISI American iron and steel institute  AWG American wire gauge  CAE Computer-aided engineering CFD Computational fluid dynamics CMOS Complementary metal-oxide-semiconductor DAS Data acquisition system DI Direct injection GDI Gasoline direct injection GP Glow plug HS Hot surface HSI Hot-surface ignition HT Heat transfer ID Ignition delay NG Natural gas NPT National pipe thread P&ID Process and instrumentation diagram PM Particulate matter  PZF Lead zirconate titanate  SAE Society of automotive engineers  SCRE Single cylinder research engine SS Stainless steel TC Thermocouple xvi  Acknowledgements I would like to begin by thanking my supervisor, Steven Rogak, for the wonderful opportunity that he gave me by making me part of this project. His guidance and teachings helped me to become a better version of myself, both professionally and personally. Just as much recognition deserves Patrick Kirchen. I’m thankful for his amazing support and for always treating me as one of his supervised students. I would also like to thank Gordon McTaggart-Cowan for being so accessible and supportive with me. I’m a big admirer of their dedication, knowledge, and engineering skills, and I was very lucky to have learnt from them.  I was also fortunate to collaborate with members of Westport during my time at UBC. I would like to thank Sandeep Munshi, Hamed Karimi Sharif, Ashish Singh, Gavin Hartnett, and Riley Cahill for all their help and input expertise in the development of this project.  I would like to acknowledge the financial support of this project. Funding was received from the Natural Sciences and Engineering Research Council of Canada (NSERC) Collaborative Research and Development (CRD) grant in conjunction with Westport Fuel Systems Inc., the Clean Energy Research Centre (CERC) at The University of British Columbia, the NSERC Collaborative Research and Training Experience (CREATE) program, and the Mitacs organization.   I’m grateful for having had the opportunity to share the lab space with such an amazing group of grad students. It was a pleasure to work with them. I would like to thank Mike Karpinski-Leydier, Pooyan Kheirkhah, Mahdiar Khosravi, Jeff Meiklejohn, Jeremy Rochussen, Aditya Singh, and Dave Sommer for their help and friendship. Special thanks go to Mahdiar for his help with the pyrometer method and to Aditya, with whom I spent most of my time in the lab. I would like to acknowledge two family members that also helped me during this time. Thanks to my grandma Rosy Hernández, for her love, support, and for being an inspiration to me and the Gutiérrez family. And thanks to my uncle Edy Gutiérrez, whose persistence in life encouraged me to persevere during the highs and lows of experimental research.    Finally, I would like to thank my parents, René Zepeda Sr. and Paty Gutiérrez, and my sister, Alina Zepeda. Your steadfast love, support and belief in me throughout my life is the reason why I am here today. Thanks for always being there for me. I’m forever grateful to have you. xvii  Dedication   Dedicated to my parents, René Zepeda Zaragoza and Patricia Gutiérrez Hernández.  1  Chapter 1: Motivation and Objectives Air pollution is a serious problem affecting human health and climate change. Traffic of vehicles is identified as one of the main producers of air pollution, the majority of which use diesel or gasoline to operate. One strategy to reduce the pollutant emissions is to use using natural gas inside direct-injection engines. A method to do this involves using a hot surface, which provides an ignition point for the natural gas. The high temperature required on the hot surface suggests the need to develop methods to characterize this device.   1.1 Introduction Air pollutants, such as particulate matter emissions, are now believed to be the number one cause of death in the world [1], surpassing other death causes, such as cigarette smoking and vehicle road injuries. In Europe, it has been estimated that the mean life expectancy is reduced by 2.2 years as a result of exposure to air pollution. [2]  The internal combustion engine used for motorized transportation is one of the top contributors of air pollution. A recent study [3] identified the traffic of vehicles as the number one source of PM2.5 emissions, a damaging pollutant to the human cardiovascular system. Pollutant emissions from a combustion engine include nitrogen oxides, sulfur oxides, carbon monoxide and dioxide, unburned hydrocarbons and the aforementioned PM. Among them, CO2, methane and PM are given special attention due to their impact on climate change. As such, engine manufacturers have turned their attention to decrease the levels of pollutant emissions by exploring different combustion strategies and alternative fuels.   One alternative fuel that has gained attention lately is natural gas, due in part to its availability and competitive price. [4] In terms of production of pollutant emissions, the 2  combustion of natural gas produces less CO2 emissions than other fossil fuels, owing to the lower carbon to hydrogen ratio of methane, its main constituent. [5] Thermal formation of NOx is also lower compared to diesel or gasoline, due to the lower adiabatic flame temperature reached when burning this gaseous fuel. On the other hand, unburned methane emissions from the combustion of NG represent a problem due to methane’s high global warming potential.   Direct-injection (DI) engines present certain advantages against port fuel injection engines. For example, energy losses associated with air throttling are absent, as engine torque is only controlled by the amount of fuel injected. In addition, in late-cycle DI engines, the fuel is injected close to the end of the compression stroke, avoiding engine knocking. This allows higher engine compression ratios, increasing the fuel conversion efficiency. [6] Traditional late-cycle DI engines work with diesel. The diesel is injected near the end of the compression stroke so that it ignites by mixing with the high-temperature in-cylinder air of the combustion chamber. A different combustion strategy in DI engines consists in injecting the fuel at the intake stroke and using a spark plug to initiate the combustion event. Gasoline engines using this technology (GDI) have become very popular the last few years. However, GDI engines can generate higher black carbon emissions compared to gasoline port fuel injection engines, which could offset some of the benefits related to the use of this technology. [7] The higher efficiencies that can be achieved with DI engines have encouraged the development of combustion strategies to burn alternative fuels, such as natural gas.  1.2 Research motivation One method to use natural gas inside late-cycle DI engines consists in using a hot surface to initiate the combustion process. For the combustion to occur quickly and repeatedly, the surface 3  must be operated above 1200 K. [8] This can be a challenge, as the fuel jets, which should come in contact with the HS, produce a cooling effect on the surface, retarding the ignition. In addition, if the design of the hot surface requires the injector and HS to be positioned at a close distance, the high temperature of the HS could bring detrimental effects, such as damaging the mechanical strength of the injector or altering the fuel by pyrolysis. Thus, a characterization of the temperature of the HS during normal operation could bring a better design and operating strategies that result in the propagation of this technology.   1.3 Research objectives and thesis structure The main goal of this thesis was to develop and validate two temperature characterization methods for the hot-surface ignition-assist technology. The first one consisted in the characterization of the temperature of a fuel injector equipped with a heater ring. This injector, also called hot-surface ignition (HSI) injector, is an application of the HS technology conceived by Westport Fuel Systems. The second characterization consisted in the temperature of a hot surface during the fuel jet impingement event. This, as the cooling effect of the jet produces a localized rapid drop in temperature. To accomplish this, the following objectives were set: 1. Prediction of the steady-state temperature of the HSI injector under combustion engine conditions. To accomplish this, a numerical model was developed and validated using a series of heat transfer simulations.  2. Measurement of the temperature drop of a HS application as a consequence of gas jets impinging on its surface. To accomplish this, a pyrometer method with high spatial and temporal resolution was developed and validated, which was used to 4  analyze the cooling effect of a series of jets with different injection pressures, durations and orientations.  This thesis was split into five chapters, with Chapter 3 and Chapter 4 containing the main analyses.  In Chapter 3: Modeling of the Heat Transfer in the Hot-Surface Injector, the temperature of the HSI injector under combustion engine conditions was predicted using heat transfer simulations. A series of experiments were performed to validate the model. The most sensitive components affecting the internal temperature of the injector and the temperature of the heater ring were identified by doing a sensitivity analysis. A series of design recommendations are presented at the end of the chapter.  In Chapter 4: Development and Application of a Pyrometer Method to Measure the Surface Temperature of a Hot Surface, a contactless temperature measurement method with high spatial and temporal resolution is presented and validated to measure the temperature of glowing surfaces. The method is used to analyze the steady-state temperature of a heater ring, as well as the cooling effect of a series of jets with different injection pressures, durations and orientations impinging on the heater.      5  Chapter 2: Background Information The high compression ratio required to auto-ignite natural gas in a timely manner inside late-cycle direct injection engines necessitates the use of ignition assist methods, such as the hot surface method. A glow plug is a device that has been used as a hot surface, providing an ignition point for the jets of natural gas. Another application of a hot surface consists in a fuel injector equipped with a heater ring close to the injector nozzle. As the temperature of the hot surface is crucial to achieve high combustion efficiencies, and considering the rapid cooling event induced by the natural gas jets, developing a pyrometer method to measure its temperature during operating conditions could help to improve its design.    2.1 Introduction The attractiveness of natural gas as an alternative fuel have encouraged the development of ignition technologies able to maximize its benefits and increase its use in the motorized transportation industry. One of these strategies involves the combustion of the NG inside late-cycle direct injection engines, which are able to reach higher efficiencies than their port fuel injection engine counterparts. [9] As methane, the main constituent of NG, has a research octane number of more than 120 [10], its auto-ignition temperature is higher than that of other fossil fuels used in DI engines, preventing its ignition in a timely manner. [11] To overcome this, ignition assist methods are used, providing an ignition source for the NG. An ignition assist method that has been successfully commercialized concerns the use of a pilot quantity of diesel injected directly into the cylinder [5], which produces a flame that is used to start the combustion of the NG. Engines using this technique are able to achieve a performance and efficiency similar to DI diesel engines while reducing the production of pollutant emissions. [12] However, it also implies the need to 6  add a secondary fuel system which relies in a liquid fossil fuel. To avoid this, an alternative ignition assist method consists in using a hot surface device to increase the local temperature of the natural gas-air mixture and produce a flame. Usually, these devices are commercially available GP with some modifications. In this chapter, an overview of studies using GP to ignite NG will be presented, with the idea to provide the reader with background knowledge about this technology. Then, a novel HS device will be introduced. This device, conceived by Westport Fuel Systems, is a NG injector equipped with a HS ring at the periphery of its tip [13], which looks to address some of the challenges associated with the use of GP. Lastly, as it has been demonstrated that the temperature of the HS is fundamental to achieve low ignition delays [14], methods that may be used to measure its temperature will be presented.  2.2 Use of glow plugs to ignite natural gas inside direct injection engines A common device used as the ignition source for NG inside DI engines is a GP. These devices are designed to be used in diesel-cycle engines to aid in the engine start when the local temperature of the compressed air is too low to ignite the fuel. When used as ignition assist for combustion of NG, the GP must provide enough heat to a portion of the ignitable mixture of air and fuel so that a flame can be produced, burning the rest of the mixture inside the combustion chamber. This ignition process should be sufficiently quick to ensure an ignition delay within 2 ms to achieve high combustion efficiencies. A test conducted by Aesoy and Valland found that the GP should keep a surface temperature above 1200 K to accomplish this ignition delay timing. [8] Maintaining a high temperature on the GP is one of the main challenges associated with this technology. As the air-fuel mixture should come close to the HS in order to reach its ignitable temperature, a common strategy is to impinge the GP with the fuel jets coming from the fuel 7  injector. This, however, produces a cooling effect on the surface of the GP, prolonging the ignition timing. To reduce the cooling effect, studies have determined that the use of a shield is critical to protect the GP from the fuel jets [15,16], which also has the advantage of protecting the GP from the in-cylinder air flows. The shield should be designed in such a way that only a portion of the jet enters and mixes with air to form an ignitable mixture that resides for a moment in the close vicinity of the GP, starting the flame. The flame, then, should come out of the shield to burn the rest of the fuel inside the combustion chamber. [17] Besides needing a shield, it was found that, in order to achieve short and repeatable ignitions, the GP must be powered beyond its recommended input voltage [18,19] to reach higher surface temperatures, shortening its lifecycle. During the years 2000 to 2006, an engine using this technology was developed to be used inside a commercial vehicle, with the final product showing overall good results. [20] This project, developed by Westport Fuel Systems and Isuzu, showed good performance and emissions levels. However, issues with reliability and durability of the ignition components (GP) hindered the further development of the project.   2.3 Westport’s Hot-surface ignition fuel injector Considering the shortcomings of the glow plugs, Westport Fuel Systems designed a hot surface (HS) integrated with a NG fuel injector, eliminating the need to add extra components to the combustion chamber. [13] This HS, having the shape of a heater ring, is located close to the injector tip, surrounding the injector at a location close to the nozzle holes. To ignite the NG, a pilot quantity of NG is injected close to the heater ring, which causes the air-fuel mixture to ignite. For this, various pilot jets that emerge from the injector are used. The ignition happens in a semi-8  protected area to shelter the hot surface from the in-cylinder flows and increase the residence time of the ignitable jets. The main fuel jets are then injected inside the combustion chamber, which ignite thanks to the flame already formed by the pilot jets.  A prototype of the HSI injector based on the original patent was recently tested inside a combustion vessel and further evaluated through CFD modelling. [21] In this HSI injector design, the pilot jets impinge on the heater ring, which is protected by a recess. This recess helps the HS to avoid the cooling, and degrading, effect of the main fuel jets, as well as increasing the residence time of the pilot jets in the vicinity of the heater to initiate the combustion process. After the ignition starts, the flame propagates to the rest of the combustion chamber to burn the main jets. The tests using this injector confirmed the importance of running the HS at a high temperature to keep a low ID. The temperature was noted to be non-uniform across the heater. Using eight pilot gas jets resulted in a better combustion process (i.e. faster ignition, stronger flame) than using a single jet, in part due to the eight jets having a more angled orientation when emerging from the injector. These pilot jets should form an ignitable air-fuel mixture close to the HS to ignite, as well as carrying enough momentum to push the flame out of the semi-protected area towards the combustion chamber, where the main gas jets are injected. It can be concluded, both from the GP studies and from Westport’s HSI injector study, that the temperature of the HS is crucial to achieve low ignition delay (ID). However, the need to run the HS at temperatures as high as 1500 K bring technical challenges when tightly integrated with a fuel injector.   9  2.4 Two-colour pyrometer method  The temperature of the HS must be kept constantly at a high temperature in order to achieve high combustion efficiencies. Yet, during normal operation of the HS, the surface is expected to experience a cooling event as a consequence of fuel jets impinging on its surface. As such, analyzing the temperature of the surface during the fuel impingements could help to improve the methods to inject the fuel to decrease the temperature of the HS as least as possible. This motivated the research of a temperature measurement method with high temporal and spatial resolution that could provide a temperature map of the HS. One method with the potential to do this consists in a two-colour pyrometer. This type of pyrometers works by taking the ratio of the radiation intensity emitted at two different wavelength bands to calculate the temperature, which requires the assumption of the surface behaving as an ideal grey surface (constant spectral surface emissivity). Errors in the results will occur if this assumption is incorrect. [22] There are different techniques to develop a two-colour pyrometer. One of these involves the use of an optical fiber and two photodetectors. The optical fiber is positioned close to the target location to capture the radiation emitted according to the numerical aperture of the fiber. The radiation captured is then spectrally filtered to conduct the signal to a pair of photodetectors operating at different wavelength bands. This can be done in different ways, including using detectors in series that can absorb and transmit the incoming radiation according to the wavelength [23,24], using a dichroic beamsplitter [25], using bandpass filters in front of the photodetectors [26], or using a wavelength division multiplexing filter. [27] The optical fiber technique has been used for different applications including machining processes [23,27,28], laser sintering processes [24], and a brake pads study. [26]  10  One of the main drawbacks with the use of fiber optics concerns the small target size (in the order of millimeters) for which the temperature can be calculated. As an alternative, a different technique to apply the two-colour pyrometry method consists in using a digital camera to capture the radiation emitted by a surface. Applications using this technique include combustion processes [29–32], laser-based processes [33,34], and a circuit breaker study. [35] As the measurement target size of the camera depends on its field of view, the radiation of several points can be measured simultaneously, permitting the calculation of the whole temperature map of the target. In addition, if a high-speed camera is used, the temporal resolution can allow for the visualization of rapid changes in temperature. To measure both sets of radiation data, more than one camera may be used simultaneously. [29,34] A beamsplitter is used to separate the light into different wavelength bands and redirect the output signal to the cameras. As the cameras must see the same image, the components involved should be meticulously positioned to avoid erroneous measurements. A single camera has also been used in two-colour pyrometers. One way to do this is by using a collimator and a series of mirrors and filters. [35] Two bandpass filters are positioned so that the filtered radiation from the target reaches a pair of mirrors. The mirrors then reflect the image to a collimator, which redirects the signal to the camera. This results in the camera seeing two images side by side, each of them carrying the filtered radiation according to the bandpass filters used, and each of them showing the target from a slightly different perspective and position, which requires the images to be post-processed to ensure both represent the same pixel at the same location. Another method involves using a colour camera and its colour filter array [30,31,33], which is used to filter the incoming radiation in different colours (i.e. wavelengths). To use the filtered information to calculate the temperatures, the recorded image should go through a series of post-processes to extract the spectral radiation read by each pixel. 11  In summary, a two-colour pyrometer method using a camera should be able to measure the radiation at two wavelength bands in order to calculate the temperature of the target, making sure the radiations measured correspond to the same location (i.e. same pixel) at the same state (i.e. same temperature). In Chapter 4, a modified version of the two-colour pyrometry method using a digital camera is presented. This modified version, referred to as a hybrid pyrometer, consists on the measurement of the temperature of a surface at steady-state using two-colour pyrometry. Then, the results are used to make a correlation of temperature with radiation intensity at a single wavelength, which allows for the calculation of the temperature of the surface at transient-state using the camera and one bandpass filter.  A study by Maun et al [32] investigated a slightly similar method, where the brightness intensity of a SiC fiber, used for flame temperature measurement, was measured by a colour camera. The intensity collected by the camera was correlated with measurements done with thermocouples, allowing to calculate the temperature of the flame by using the intensity read by the colour camera.      12  Chapter 3: Modeling of the Heat Transfer in the Hot-Surface Injector The high temperature required in the HSI injector heater to have a reliable and fast ignition event could potentially damage the rest of the injector components and increase the power consumption of the heater due to heating losses. A series of heat transfer simulations were performed and compared with experimental tests to develop and validate a tool able to predict the temperature of the injector at combustion chamber conditions. Using the results, a sensitivity analysis of the main variables was conducted, and a series of design recommendations for the HSI injector were provided.   3.1 Introduction One of the key aspects in the hot-surface ignition assist method involves keeping the fuel ignition delay timing below 2 ms, as discussed in Chapter 2. To accomplish this, the heated surface must be kept at a temperature of more than 1200 K during engine operation. In Westport’s HSI injector, the hot surface is integrated within the injector, as discussed in Chapter 2, which could affect the injector performance due to their close proximity with each other, damaging its mechanical properties or altering the fuel by pyrolysis. Moreover, unwanted heat losses from the heater ring could increase the power consumption required to maintain a given temperature. In this chapter, the heat transfer interactions at steady-state between the heater ring and the rest of the injector components were analyzed using a computer-aided engineering (CAE) software with the goal of finding the temperature of the injector and heater ring at different locations. To provide confidence in the results generated by the simulations, a validation process is presented here, where certain aspects of the HSI injector model were tested separately. This validation process included the testing of an HSI injector prototype inside a combustion vessel, which allowed to measure the 13  temperature of an object resembling the injector under conditions steadier than those that an actual fuel injector inside a combustion engine would face. Finally, a sensitivity analysis on the main components of the HSI injector model was done to identify the components with the largest effect on the temperature. Based on the results of this analysis, certain design recommendations were provided at the end of this chapter.   3.2 Methods One common approach to analyze the heat transfer interactions in a system involves using a software able to simulate these interactions by solving for differential equations, such as the Navier-Stokes equation. [36] The use of software is usually complemented by experimental tests to validate the model, providing some confidence in the numbers generated. In this section, the methods used to follow this approach are presented, which were used for all the tests that will be described in later sections of Chapter 3.   3.2.1 CFD simulations using Ansys Icepak The heat transfer simulations were done using Ansys Icepak (version 17.2). Icepak is a CAE software that provides tools to build a model using predefined objects and generate a mesh in a semi-automated fashion. [37] To solve for the thermal and fluid-flow calculations, Icepak uses the Ansys Fluent commercial CFD solver. The results can then be post-processed in Icepak to examine and visualize them.  As in any CFD software, Icepak (via Fluent) solves for transport and conservation equations to reach a solution. These equations are applied to each cell of the mesh in an iterative 14  process until the residual of the sum of the terms from the equations reaches a value sufficiently close to zero. The governing equations used in Icepak [37] are: Mass conservation equation:  𝜕𝜌𝜕𝑡+ ∇ ∙ (𝜌 ?⃗?) = 0 (3.1) Where ρ is the density of the fluid, v is its velocity, and t is the time.  Momentum equation:  𝜕𝜕𝑡(𝜌 ?⃗?) + ∇ ∙ (𝜌 ?⃗? ?⃗?) = −∇𝑝 + ∇ ∙ (𝜏̿) + 𝜌?⃗? + ?⃗? (3.2) Where ρ is the density of the fluid, v is its velocity, τ is the stress tensor, p is the static pressure, ρg is the gravitational body force, and the F represents external body forces and source terms.  The stress sensor τ is given by:  𝜏̿ =  𝜇 [(∇ ?⃗? + ∇ ?⃗?  𝑇) −23∇ ∙ ?⃗? 𝐼𝑇] (3.3) Where μ is the dynamic viscosity of the fluid, v is its velocity, IT is the unit tensor, and the ∇ ?⃗?  𝑇 term represents the effect of volume dilation.   Energy equation (for fluid regions):  𝜕𝜕𝑡(𝜌 ℎ) + ∇ ∙ (𝜌 ℎ ?⃗?) = ∇ ∙ [(𝑘 + 𝑘𝑡) ∇ 𝑇] + 𝑆ℎ (3.4) Energy equation (for solid regions):  𝜕𝜕𝑡(𝜌 ℎ) = ∇ ∙ (𝑘 ∇ T) + 𝑆ℎ (3.5) Where t is the time, ρ is the density of the region, h is its sensible enthalpy, v is its velocity, k is the thermal conductivity, kt is the thermal conductivity due to turbulent transport, and the Sh term is a source term that includes any volumetric heat sources defined in the model. 15  In each of the cases simulated, the Fluent solver was ran until the equations described above converged to a solution. Then, the results were analyzed using Icepak’s post-processing tools. These tools were able to generate the temperature maps of the models, as well as registering the temperature at specific locations within the solution domain.  A description of the steps followed to build the simulation models in Icepak can be read in Appendix A.    3.2.2 Comparison of results between experimental tests and simulations To complement the simulations, a series of experimental tests were carried out. These experiments were performed with the idea to validate the temperatures predicted in Icepak by comparing them with the temperatures measured. This comparison was done considering the uncertainty associated with each method.  In the case of the experiments, the uncertainties considered consisted in the accuracy of the thermocouples, the accuracy of the DAS, the accuracy of a commercial pyrometer, the uncertainty of the input emissivity used in the pyrometer, and the standard deviation of the data collected. Then, to calculate the total uncertainty, the propagation of uncertainty method was used. In the case of the simulations, the uncertainty of the results consisted in the uncertainty associated with the parameters of the simulations, such as the heat generation in the heating element and the properties of the materials simulated. The uncertainty of the heat generation parameter came from the accuracy of the electric current and voltage readings used in the experiments.  16  3.3 Experimental tests to develop and validate the injector heat transfer simulation Several elements are involved in the HSI injector model used for the heat transfer simulation. Before a robust simulation can be performed, the main elements should be studied, and their simulation validated to provide confidence in the final simulation results. To do this, a series of heat transfer simulations and experimental tests were carried out, comparing the results from each simulation with the results from their corresponding experimental test. This process was done in stages, increasing the complexity of geometries and conditions at each stage. If the temperature results agreed within the uncertainty of each method, the simulation of the stage was validated, which allowed the validation process to move onto the next stage. This process included two preliminary validation stages. The first preliminary stage was conducted to familiarize with Icepak for this kind of applications, simulating a test with a cartridge heater and an aluminum 6061 rod. The second preliminary stage consisted in the simulation of insulating mechanisms, such as using an enclosed air gap and a layer of insulating material between objects with large temperature differences. These preliminary stages can be read in Appendices B and C, respectively. The actual stages of the validation process consisted in only two stages. The first one consisted in the testing of the HSI injector heater ring and insulation layer, which were both mounted to a specially built SS object, while the second stage consisted in the testing of a prototype of the HSI injector mounted inside a combustion vessel.  3.3.1 Stage one of validation tests: Heater ring  For the first stage of the validation tests, certain components from the HSI injector were used. These components were the injector heater ring and the insulation beneath the heater ring, which is of critical importance to keep the heater at a high temperature by reducing the heat transfer 17  from the heater to the rest of the components in the injector. A correct simulation of these two elements at this stage should provide some confidence in the simulation of the same elements in the HSI injector model. Therefore, the objective of this stage was to validate the simulation of the heater ring and insulation. To simplify the test, a small stainless steel object was machined, so that an object made from a metal able to withstand high temperatures could be used to mount the insulation material and heater ring. The use of this metal piece also facilitated the installation of thermocouples, making possible the comparison of the temperatures predicted in Icepak with the temperatures measured using thermocouples in the experimental tests.  The test rig consisted in a 20 mm long SS 304 object with a 30 mm diameter, which was used to mount the insulation layer, heater ring, and thermocouples (figure 3.1).   Figure 3.1. Multiview Projection of the SS Object Used for Stage One. Dimensions shown are in millimeters. Stainless steel type 304 was used.   18  Two small holes were machined to allow the electrical leads of the heater ring to come out from the SS object. To mount both the insulation and the heater ring1, a 6 mm recess was done in the SS object. The insulation layer consisted in zirconium dioxide (ZrO2), also called zirconia, and the heater ring was made from coiled Kanthal APM heating wire, No. 18 AWG (1.024 mm diameter). The heater ring was partially embedded in the insulation material, with only the top part of certain sections of the heater being exposed (figure 3.2).  Figure 3.2. Detail of the Heater Embedded in the Insulation Layer Used for Stage One.  To supply the power to the heater, a Sorensen XPH 42-10 DC power supply was used. To connect the heater to the power supply, a pair of alligator clips were used. A nominal power of 80 W (± 0.97 W) was supplied to the heater. This was estimated using Ohm’s law, and the measured current and voltage, which were read by the DC power supply (accuracy of 0.3% + 3 counts for voltage and 0.6% + 3 counts for current).                                                   1 Mounting of the insulation layer and heater ring was done by the UBC Materials Engineering department 19  To run the experiment, the power was supplied to the heater while a pair of type-k thermocouple ungrounded 1/16-inch probes (accuracy of 0.4%) measured the inner temperature of the SS object. The locations of the temperature measurement points can be seen in figure 3.6. The test was ran for 45 minutes in order to reach steady-state.  Icepak was used to generate the computational model and perform the heat transfer simulation. The power input of 80 W (± 0.97 W) was set as the constant heat generated by the heater. For the boundary conditions, an ambient temperature of 22 °C was set, and heat transfer modes by convection and by radiation were activated. In table 3.1, the list of the most relevant properties of the materials involved in the simulation are shown. These materials were Kanthal APM for the heater, refractory clay for the fire brick, stainless steel 304 for the SS object, and zirconia for the insulation layer.  Table 3.1. Materials’ Properties Used in the Model of Stage One. Material Property  Kanthal APM Thermal conductivity (W/m·K) 22 – 26 [38] Emissivity 0.7 [38] Refractory clay Thermal conductivity (W/m·K) 0.23 [39] Emissivity 0.65 – 0.75 [40] Stainless steel 304 Thermal conductivity (W/m·K) 16.2 – 21.5 [41]  Emissivity 0.15 – 0.44 [42,43] Zirconia Thermal conductivity (W/m·K) 1.84 – 2.3 [44] Emissivity 0.7 [45]  The relatively large range of properties found in the literature produced different temperature results. This enabled the generation of three sets of results. The first one, called 20  nominal, using the nominal power input and the mean value of the properties shown in table 3.1. The second one, called lower limit, using the lower range of the power input and a combination of properties intended to produce the lowest temperature in the SS object. And the third one, called upper limit, which was produced using the higher range of the power input and a combination of properties intended to produce the highest temperature in the SS object. The exact combination of properties can be appreciated in table 3.2. For the parameters that are not shown in table 3.2, their values are the same as the ones shown in table 3.1.   Table 3.2. Parameter Values Used to Generate the Simulations of Stage One. Parameter Nominal Lower limit Upper limit Power input (W) 80.00 79.03 80.97 Kanthal APM Thermal conductivity (W/m·K) 24 24 24 Refractory clay Emissivity 0.70 0.75 0.65 SS 304 Thermal conductivity (W/m·K) 18.9 16.2 21.5 SS 304 Emissivity 0.30 0.44 0.15 Zirconia Thermal conductivity (W/m·K) 2.07 1.84 2.30 Percentage of surface area exposed in heater ring object (%) 14.6 18.0 10.9  To build the heater ring model, the profile of the heater had to be changed from circular to square, as Icepack does not allow to build objects with torus shapes. Another aspect of the heater ring that could not be replicated was its coiled shape, which could result in a modelled object with different volume and surface area than the actual object. Also, as it was seen in figure 3.2, some sections of the heater were partially exposed, protruding from the insulation layer. To account for 21  this, a small object was modelled just on top of the main heater ring model, keeping most of its surface area inside the insulation. To consider the effect of different levels of exposedness of the heater coil, another parameter was added: the percentage of surface area exposed in the heater ring. The final heater ring model used for the nominal simulation case can be seen in figure 3.3.   Figure 3.3. Cross-section View of the Change in the Heater Ring Profile. The profile was changed from circular to square, as Icepak does not allow to build models with circular profiles.  The mesh used for the nominal case analyzed was composed by 518,804 elements, with the elements having a maximum size of 0.01 m at each axis of the Cartesian coordinate system. The mesh type was set to “hex-dominant”, a mesh composed mostly by hexahedral elements. The mesh used for the model can be seen in figure 3.4.  22   Figure 3.4. Views of the Mesh for Model of Stage One Focused on the SS Object The temperature map generated from the simulation results of the nominal case can be seen in figure 3.5. 23   Figure 3.5. Temperature Map of the Stage One Simulation, Nominal Case. Regions at a temperature above the upper colour scale limit (714.0 °C) appear red. The ambient temperature was set to 22 °C, which is the lower colour scale limit.  As it was expected, the highest temperature in the map was located at the heater ring, where a temperature of close to 700 °C was reached as a consequence of setting the heat generation to 80 W. The effect of the zirconia insulation can be seen in the large temperature gradient across the insulation region, where the temperature drops in almost 125 °C.  To make the comparison of the simulated cases with the measured temperatures in the experiment, the results were plotted in a column chart (Figure 3.6).  24   Figure 3.6. Comparison of Results between Experiment and Simulation of Stage One. On the left side, a drawing of the cross-section of the SS object with the two temperature measurement locations is shown. On the right side, the temperature results collected from each method are shown. Error bars from the “Experiment” are based on the uncertainty of the equipment used. Error bars from the “Simulation” represent the “Upper limit” and “lower limit” cases.  The temperatures measured in the experiments were about 30 °C less than the temperatures simulated in the nominal case. However, for the lower limit, the temperatures from the simulation are in agreement with the experiment (figure 3.6). This suggests that the properties used for the lower limit case should be used for the model, producing results similar to the experiments. This information will be useful when creating the HSI injector model, especially when defining the properties of the zirconia insulation layer. For the heater ring model, it appears that the change in profile, going from circular to square, did not affect negatively the matching of the results with the experiment.   25  3.3.2 Stage two of validation tests: HSI injector under combustion vessel conditions The second stage of the validation tests consisted in the testing of an HSI injector with the internal components removed. This injector, which can be pictured as the “shell” of the HSI injector, was equipped with the same heater and insulation of the actual HSI injector. The heat transfer simulations of this injector were conducted once using Icepak. To run the experimental tests, the injector shell was tested inside a combustion vessel2, which made possible the validation of the modelling of certain key aspects (i.e. heater and insulation) of a real, fully functional, HSI injector. The combustion vessel used for the test is described in detail in [21]. In addition, simulating the steady-state conditions inside a combustion vessel is less challenging than simulating the conditions inside the combustion chamber of an internal combustion engine, due in part to the isochoric nature of the combustion vessel. Also, using a device like a combustion vessel allows to test a prototype of the HSI injector without having to deal with all the complications and risks that are involved in a combustion engine, as well as providing easier access to the injector to mount thermocouples. For these reasons, this test was performed as the last step prior to simulating the temperature of the HSI injector under combustion engine conditions.  To run the tests, the injector shell was equipped with the HSI tip3, which is a removable piece that is used to hold the insulation and the heater ring in place. The HSI tip can be seen in figure 3.7.                                                   2 Tests conducted by Westport Fuel Systems  3 The housing of the HSI tip was designed and built by Westport Fuel Systems. The insulation and heater ring were installed by the UBC Materials Engineering department 26   Figure 3.7. Detail of the HSI Injector Tip Design Used for Stage Two.   To measure the temperature of the injector, a pair of type-k thermocouples were attached to the test objects using high temperature cement. To power up the heater ring, its electrical leads were connected to a Sorensen XHR33-33 DC power supply (accuracy of ±0.43 A for current and ±0.43 V for voltage). A voltage input was set in the power supply (table 3.3), which caused the temperature to rise. As soon as the system reached steady-state, a handheld fluke thermocouple reader was used to record the temperatures. To measure the surface temperature of the heater ring, an Ultimax UX-20P single-colour narrow band (0.96 μm) pyrometer (accuracy of 0.5% + 1 count) was used. This was done by focusing the pyrometer at a single spot within the heater ring surface, and by setting the emissivity compensation in the device to 0.75.   In table 3.3, the temperatures recorded for a series of nominal power inputs can be seen. 27  Table 3.3. Temperatures Recorded in the Experimental Tests of Stage Two. Nominal power input (W) T point 1 (°C) T point 2 (°C) T heater ring (°C) 15.0 (± 5.0) 48.3 37.2 674.0 19.2 (± 5.4) 56.5 44.0 732.5 24.3 (± 6.0) 58.2 42.6 803.9 28.8 (± 6.9) 68.6 48.7 866.3 38.5 (± 8.0) 82.6 57.0 1001.6  The cross-section schematic of the HSI injector shell used for the tests, as well as the locations of the temperature measurement points, can be seen in figure 3.8. 28   Figure 3.8. Cross-section View of the Test Object Used for Stage Two.   The heater ring object modelled had some differences with respect to the actual geometry (figure 3.7). The heater ring was modelled as a solid ring with a square profile, as opposed to the torus and coiled shape of the actual heater ring. This change could impact the temperature results, as the modelled object could have different volume and surface area as the actual object, affecting the heat transfer by radiation. The schematic with the dimensions of the HSI injector shell can be seen in figure 3.9.  29   Figure 3.9. Cross-section View with the Dimensions of the Test Object Used for Stage Two. Dimensions shown are in millimeters.   To run the heat transfer simulations, the properties assigned to the materials in the model can be seen in table 3.4.   30  Table 3.4. Materials’ Properties Used in the Model of Stage Two. Component Material Property  Injector nozzle M2 tool steel Thermal conductivity (W/m·K) 21.3 [46] Emissivity 0.44 [47] Insulation  Zirconia Thermal conductivity (W/m·K) 1.84 Emissivity 0.7 Heater ring Kanthal APM Thermal conductivity (W/m·K) 24 Emissivity 0.7 HSI tip  Invar Thermal conductivity (W/m·K) 10.5 [48] Emissivity 0.1 [40] Washer Copper (C110) Thermal conductivity (W/m·K) 398 [44] Emissivity 0.07 [49]  Combustion vessel Grey iron Thermal conductivity (W/m·K) 30 [41] Emissivity 0.21 [49] Injector sleeve AISI 4140 Thermal conductivity (W/m·K) 42.7 [41] Emissivity 0.44 [47]   As it could be seen in figure 3.8, the mounting of the heater ring in the HSI tip was different than the mounting discussed in section 3.3.1 (figure 3.2). This time, the heater ring was not embedded within the insulation layer. Instead, the heater was positioned at the periphery of the insulation, only touching it from the outside. This was addressed while creating the heater ring model in Icepak for the simulations of this section. To account for the small portion of the heater surface making contact with the insulation layer, an additional heater object was added between the insulation and the main heater ring model. As the contact between the heater ring and insulation layer in the real HSI tip was not very strong, a low thermal contact conductance of 1500 W/(m2·K) 31  [50] was specified at the interface of the insulation layer and the supplementary heater ring object. The detail of the heater ring object can be appreciated in figure 3.10.   Figure 3.10. Detail of the Heater Ring Model Used for the Simulation of Stage Two.  The nominal power input shown in the first column of table 3.3 was used as the heat generation in the heater ring. The ambient temperature was set to 24 °C, which means that the air simulated under the injector had a temperature of 24 °C at the beginning of the simulations. Once again, all three modes of heat transfer were present in the simulation: conduction, convection and radiation.  The mesh of the model used for the simulation was composed by 1,440,963 elements, with an overall maximum size of 0.008 m at each axis of the Cartesian coordinate system. The density 32  of the mesh was higher inside, and close to, the smaller objects of the model, such as the heater ring and air gap between the injector nozzle and HSI tip. Buffer zones were specified around these regions to avoid having big jumps in terms of the mesh element size. The mesh used for the model of stage two can be appreciated in figure 3.11.  Figure 3.11. Views of the Mesh for the Model of Stage Two Focused on the HSI Tip. The figure on the right shows the surface mesh of the HSI tip, insulation, injector nozzle, washers, and heater ring objects.   Considering the uncertainty associated with the equipment used to calculate the nominal power inputs shown in table 3.3, three cases were simulated in Icepak for each condition. One, using the nominal power input as the heat generation in the heater ring, and two more, assuming the heat generation to be equal to the nominal power input plus its uncertainty (i.e. nominal value ± its uncertainty). To show the temperature distribution in the test object, the nominal 38.5 W power input case was used to generate the temperature map shown in figure 3.12. 33   Figure 3.12. Temperature Map of the Stage Two Simulation, 38.5 W Nominal Case. Regions at a temperature above the upper colour scale limit (500 °C) appear red. The ambient temperature was set to 24 °C, which is the lower colour scale limit.  As it can be seen in figure 3.12, a large temperature drop was located at the insulation layer, showing a similar behavior as the insulation simulated in the stage one of the validation tests (figure 3.5). The temperature of the body of the injector was notably lower than the average temperature of the HSI tip object, which included both the insulation and the heater ring. This 34  appears to be a consequence of the insulating properties of the air gap separating the HSI tip from the injector nozzle. The results from the rest of the cases simulated can be seen in figure 3.13, which included a comparison with the results from the experimental tests.   35    Figure 3.13. Charts with the Comparison of Results between Experiments and Simulations of Stage Two. Error bars from the “Simulation” series are based on the uncertainty associated with the heat input value used as the boundary condition. Error bars from the “Experiment” series are based on the uncertainty of the equipment used to measure the temperature. For the temperature of the heater ring, the error bars from the “Experiment” series also include the effect of 10% of uncertainty in the input emissivity value compensation.    36  To assess the effect of the emissivity uncertainty on the temperature read by the Ultimax pyrometer, a sensitivity analysis was done. This analysis demonstrated changing the emissivity by 10% will produce a change in temperature (in kelvin) of 1.0%.  For the inner temperatures measured by thermocouples, the results from the experiments matched the results from the simulations when considering the uncertainty of each series. However, the temperature of the HSI heater read by the pyrometer did not agree with the temperature simulated of the heater.  There are some reasons that could explain this disagreement. First, as it was mentioned before, the actual geometry of the heater ring differed with the modelled geometry, which could produce different heat radiation losses owed to different surface areas. And second, the brightness of the real HSI heater was uneven, which means that the temperature of the heater varied across its surface. This contrasted with the uniform temperature of the heater simulated. The non-uniform condition of the real HSI heater surface means that the temperature read by the pyrometer was widely dependent on the measurement location, as this device is only able to capture the temperature at a single spot.  The uneven behavior of the HSI heater surface suggested the need to measure its temperature on a different way than using the Ultimax pyrometer. In chapter 4, a two-colour pyrometer method will be presented and validated, which made possible the generation of the temperature map of the whole HSI heater ring surface area at steady-state.  Now that the validation stages have been completed, the HSI injector heat transfer simulation under combustion chamber conditions will be presented.   37  3.4 Heat transfer simulation of the HSI injector Having completed the validation tests, the HSI injector model can be created with a better understanding of the properties and shapes that should be assigned to its main components. As it has been mentioned before, the temperature of the injector during normal operation is one of the concerns in the HSI injector project. For this reason, in this section, the temperatures predicted across the injector under combustion engine conditions will be presented. As no prototype was able to be tested inside a real internal combustion engine, the temperatures predicted lacked validation. However, the temperatures were used to conduct a sensitivity analysis of the main injector components and conditions driving the heat transfer across the injector, which provided some insight for areas of improvement in the design of the injector.   3.4.1 Temperature prediction of the HSI injector under combustion engine conditions To run the heat transfer simulation of the HSI injector under combustion engine conditions, a new version of the HSI injector model was created. This model was based on the “shell” injector model presented in section 3.3.2 with certain components added, such as the piezoelectric actuator inside the injector and some gaps between components filled with oil. The main dimensions of the model were the same as the model used for the “combustion vessel” simulations shown in figure 3.9. 38   Figure 3.14. Cross-section View of the HSI Injector Model Used for the Combustion Engine Simulations.  The materials and properties of the new components added to the model with respect to the “shell” injector model can be seen in table 3.5. The rest of the components had the same properties assigned as in the “shell” model from section 3.3.2.   39  Table 3.5. Materials Used for the Added Components in the Injector Model of the Combustion Engine Simulations. Component Material Property  Inner/Outer  sleeve AISI 4140 Thermal conductivity (W/m·K) 42.7 [41] Emissivity 0.44 [47] Oil gap SAE 15W-40 Thermal conductivity (W/m·K) 0.132 [51] Piezoelectric actuator Lead Zirconate Titanate (PZT) Thermal conductivity (W/m·K) 1.3 [52] Emissivity 0.4 [53] Cylinder head Grey iron Thermal conductivity (W/m·K) 30 [41] Emissivity 0.21 [49]  To run the simulations, it was decided to assign boundary conditions to several surfaces in the model, avoiding the need to create large fluid objects (e.g. air). This to improve the solving computational time in Icepak to reach a solution, and to overall simplify the model. In order for the radiation view factors to be calculated in Icepak, a medium must exist between the radiating surfaces. As such, stagnant fluid objects were incorporated between surfaces with radiation heat transfer conditions. The surface conditions assigned can be seen in figure 3.15.  40    Figure 3.15. Surface Conditions Used in the Model of the Combustion Engine Simulations.  The constant temperature of 350 K was selected according to the average temperature of the coolant used in the UBC Single cylinder research engine (SCRE) [54]. A heat transfer coefficient of 310 W/K·m2 was specified at the walls below the cylinder head and injector nozzle objects (shown in cyan in figure 3.15). The Woschni correlation [55] was used to calculate this coefficient using data from a typical low-load SCRE run during the compression stroke and a temperature of 820 K. The 820 K was calculated assuming a polytropic compression process of air (CR = 19:1) with an initial temperature of 293 K. The heat transfer coefficient was also 41  referenced to 820 K, which means that the heat transfer at the walls shown with the cyan colour was proportional to the difference between the temperature of the wall and this reference temperature of 820 K.  For the interface between the insulation/heater ring and the inner surface of the cylinder head, and the interface between the injector nozzle and the injector tip, two fluid objects with the properties of dry air at 430 K and 30 bars were modeled4. This according to the median values of temperature and pressure measured in a low-load SCRE run during the compression stroke. For the volume around the piezo actuator, a fluid object with the properties of air at standard conditions was modelled. An input power of 100 W was specified as the heat generation in the heater ring.  The “hex-dominant” mesh used for the injector model was composed by 2,868,596 cells, with an overall maximum cell size of 0.02 m at the X and Y axis, and 0.01 m at the Z axis. To have smooth mesh size transitions, “buffer zones” were set between objects of different mesh density and size. The mesh used for the model can be appreciated in figure 3.16.                                                   4 A subsequent study found a large impact of the thermal conductivity of the stagnant fluid objects on the temperatures of the HSI injector. This parameter was not considered for the sensitivity analysis presented in section 3.4.2. 42   Figure 3.16. Views of the Mesh for the Model of the Combustion Engine Simulations Focused on the HSI tip. The figure on the right shows the surface mesh of the heater ring, insulation, injector nozzle, injector tip and washer sleeve – tip objects.   To analyze the temperature of the injector for the nominal power input of 100 W, seven different temperature points were selected. Three of these points were positioned along the centre line of the injector nozzle object, two above and below the heater ring, one at the heater ring itself, and one located between the injector nozzle and injector tip. The location of the temperature points and the temperature map can be seen in figure 3.17. The results can be seen in table 3.6. 43   Figure 3.17. Temperature Map of the Combustion Engine Simulation. The numbers inside the boxes represent the temperature study points.     44  Table 3.6. Results from the Combustion Engine Simulation.  Temperature point Location Temperature (°C) 1 Heater ring – centre 1326.5 2 Heater ring – above 772.6 3 Heater ring – below 792.1 4 Between injector nozzle and injector tip  193.0 5 Injector nozzle – up 114.4 6 Injector nozzle – middle 169.4 7 Injector nozzle – tip 193.7  The heat transfer simulation predicted a temperature at the heater ring above 1300 °C when supplying 100 W to the heater. At the insulation layer, close to the heater ring, the temperatures recorded were around 735 °C. Meanwhile, in the injector nozzle, the temperatures predicted were all below 200 °C. From the temperature map, it can be seen that a large portion of the injector was at a temperature close to 80 °C (≈ 350 K), which was the assigned temperature for several surfaces within the model, representing the coolant temperature inside an engine. The sensitivity of this value in the temperature points will be discussed in section 3.4. The results from this analysis suggest that the current design of the HSI injector will be able to maintain the temperature in the main section of the injector within reasonable numbers (< 200 °C) while keeping the heater at a high temperature, which has been mentioned before to be of key importance to ensure a reliable and consistent ignition event (Chapter 2). These predictions were not able to be validated using experimental tests, as it was the case with the experiments from section 3.3. However, the validation process presented in section 3.3 provided certainty on the 45  properties that should be used for several of the components from the HSI injector model, giving some confidence on the results presented in table 3.6.  To further understand the influence of the main injector components and boundary conditions in the temperature of the HSI injector, a sensitivity analysis was carried out.  3.4.2 Sensitivity analysis of the main HSI injector components Using the computational model of the HSI injector, a sensitivity analysis was conducted to identify the most important components of the injector and boundary conditions and, as a consequence, the temperatures around the injector. To do this, several variables were selected to be studied. For each of these variables a lower bound, upper bound, and nominal value, were selected. This, in order to generate a range of temperature results based on the effect of each variable adjusted. With the results generated, the sensitivity of each variable was calculated by using equation (3.6):   𝑚𝑠 =  𝑈𝐵%,𝑜 − 𝐿𝐵%,𝑜𝑈𝐵%,𝑖 − 𝐿𝐵%,𝑖 (3.6) Where mS is the sensitivity (dimensionless), UB is the upper bound, LB is the lower bound, the subscript % represents percentage, the subscript i represents the input, and the subscript o represents the output.   In order to have the sensitivity in dimensionless terms, the lower bound and upper bound values were calculated in terms of percentage with respect to the nominal value. Equations (3.7) and (3.8) were used to do this calculation.  𝑈𝐵% =𝑈𝐵 − 𝑛𝑜𝑚𝑖𝑛𝑎𝑙𝑛𝑜𝑚𝑖𝑛𝑎𝑙 (3.7)  𝐿𝐵% =𝐿𝐵 − 𝑛𝑜𝑚𝑖𝑛𝑎𝑙𝑛𝑜𝑚𝑖𝑛𝑎𝑙 (3.8) 46  The list of the variables studied, and their respective lower bound, upper bound and nominal values can be seen in table 3.7.  Table 3.7. Variables Adjusted for the Sensitivity Analysis of the HSI Injector Model Variable Lower Bound Nominal Upper Bound Gap between inj. nozzle and inj. tip (mm) 0.4 0.5 0.6 Input power of heater ring (W) 75 100 125 Constant temperature around injector (coolant temperature) (K) 250 350 450 M2 tool steel (Injector nozzle) emissivity 0.34 0.44 0.54 Zirconia (Insulation layer) emissivity 0.4 0.6 0.8 Zirconia (Insulation layer) thermal conductivity (W/m·K) 1.64 1.84 2.04 Kanthal APM (Heater ring) emissivity 0.5 0.7 0.9 Kanthal APM (Heater ring) thermal conductivity (W/m·K) 18 24 30 HT coefficient for walls below inj. nozzle and cyl. Head (W/K·m2) 210 310 410  With these range of values for each variable, a series of heat transfer simulations were carried out, using the model presented in section 3.4.1 as the base model. The results were tabulated for each of the temperature measurement points presented in figure 3.17. Equation (3.6) was then used to calculate the sensitivity of each parameter on the temperature points selected.  The results of this study can be seen in table 3.8. 47  Table 3.8. Dimensionless Sensitivity of the HSI Injector Model Variables on the Temperature Study Points. Temperature study point  Variable 1 2 3  4 5 6 7 Gap between inj. nozzle and inj. tip -0.004 -0.020 -0.019 -0.700 0.103 0.216 0.238 Input power of heater ring 0.417 0.488 0.430 0.300 0.162 0.273 0.300 Constant temperature around injector  (coolant temperature) 0.008 0.041 0.029 0.356 0.415 0.647 0.355 Zirconia  (insulation layer) emissivity -0.024 0.072 0.053 0.054 0.027 0.049 0.054 Zirconia  (insulation layer) thermal conductivity  -0.048 -0.214 -0.139 0.000 0.026 0.008 -0.001 Kanthal APM  (Heater ring)  emissivity -0.166 -0.040 -0.057 -0.033 -0.014 -0.028 -0.033 HT coefficient for walls below inj. nozzle and cyl. head -0.011 -0.041 -0.064 -0.037 -0.003 -0.029 -0.038 Bold font indicates the parameter with the highest sensitivity in each temperature point48  To analyze the results, the sensitivities tabulated were considered in terms of their absolute value. From the above table, the following statements can be made:   For the heater ring (point 1), the variable with the largest impact was the power input (m = 0.417), meaning that, for example, a change of 5% (with respect to the nominal value) in the power input will bring a change of 2.1% (with respect to the nominal value) in the temperature of the heater ring. The variable with the second highest sensitivity was the emissivity of the heater itself (m = 0.166). This means that a change of 10% in the emissivity of the heater will produce a change in the temperature of 1.6%. The variable with the lowest sensitivity was the gap distance between the injector nozzle and injector tip.   For the points 2 and 3, which refer to spots in the insulation layer closest to the heater ring, the variable with the largest sensitivity was the power input, followed by the thermal conductivity of the Zirconia.  For the point situated between the injector nozzle and the injector tip (point 4), the variable with the highest sensitivity was the distance of the gap between the injector nozzle and the injector tip. The variable with the second highest sensitivity was the coolant temperature.  For the point located farthest from the heater ring (point 5), the coolant temperature had the largest sensitivity, with a sensitivity of 0.647. A variation of 10% in the temperature of the coolant (in Celsius) will produce a change in the temperature (in Celsius) at this point of 6.5%   For points 6 and 7, the coolant temperature was also the variable with the highest sensitivity. The variable with the second highest sensitivity was the power input. 49   In general, the power input had the largest impact for the temperature points located at, or close to, the heater ring (points 1, 2 and 3). The coolant temperature was the variable with the largest, or second largest, impact for the rest of the points scattered across the injector nozzle (points 4, 5, 6 and 7).   The thermal conductivity of the zirconia had the third highest sensitivity on the temperature of the heater ring (m = 0.048). However, its sensitivity was much lower for the points located within the injector nozzle, demonstrating how less important this value was to keep a certain temperature in the injector nozzle.   In general, the temperatures of the injector nozzle were not very sensitive to the properties of the zirconia and the emissivity of the heater ring.  In summary, the power input of the heater ring and its emissivity are key factors to ensure a high temperature in the heater ring, while the temperature of the coolant has a large effect on the temperatures across the injector nozzle. It is important to note that the insulating properties of the zirconia would improve if its thickness was increased. Nonetheless, it was decided to keep the thickness constant in the sensitivity analysis and analyze the insulating properties of the zirconia only in terms of its thermal conductivity value.    3.5 Design recommendations  From the sensitivity analysis results, some recommendations for the design of the HSI injector, from the point of view of the temperature, can be made:  50   The power input of the heater ring will affect the overall temperature of the heater ring and the main body of the injector. However, with the current HSI injector design, using a power input of 100 W, the heater can reach up to 1327 °C while keeping a temperature in the injector below 200 °C.   Special attention should be given to the surface emissivity of the material of the heater ring, as this value has a large impact on its temperature. Changes in the surface of the heater as a consequence of repeated combustion (i.e. soot deposition) could further increase the emissivity of the surface, decreasing its temperature for the same power input value.   In the sensitivity analysis, it was discovered that the coolant had a big influence on the temperature of the injector. For this reason, it must be ensured that the engine coolant can easily circulate around the injector outer sleeve, allowing the extra heat to be removed from the injector.     A gap between the injector nozzle and the injector tip should exist to insulate both components from one another, protecting the nozzle from the high temperature of the injector tip. Increasing this distance will not affect significantly the temperature of the heater ring, but it has a large impact on the temperature of the injector nozzle.   An insulation layer with a low thermal conductivity must be used to reduce the heat transfer losses from the heater to the injector nozzle. However, decreasing this thermal conductivity will not affect the temperatures of the injector nozzle too much.   One of the main elements of the HSI injector is the heater ring. A better understanding of its surface temperature could also provide some ideas on how to improve the overall design of the HSI injector. However, as it was briefly discussed at the end of section 3.3, the temperature of the 51  heater ring is non-uniform. This complicates the use of spot-constrained pyrometers, such as the Ultimax pyrometer, to provide a comprehensive temperature distribution along the heater. To tackle this problem, a pyrometer method will be developed, which should allow for the generation of the full temperature map distribution. In the next chapter, the development of this method will be presented, as well as its application on the HSI injector heater ring.        52  Chapter 4: Development and Application of a Pyrometer Method to Measure the Surface Temperature of a Hot Surface A pyrometry method was developed to produce the temperature map distribution of a HS at steady and transient states. This method consisted in a two-colour pyrometer method coupled with a single-colour pyrometer method, which allowed the temperature of the HS to be calculated during rapid cooling events. This method was validated by measuring the temperature of a stainless steel object. To test this method in a HS, the heater ring of the HSI injector was used. The produced temperature maps of the heater illustrated the different cooling effects of gaseous jets impinging at different injection pressures and durations, as well as different orientations relative to the heater.   4.1 Introduction In the HS ignition-assist method, a HS is used to initiate the combustion of NG. A jet of NG mixes with air and impinges a heated element, resulting in the ignition event. The temperature of this element is of high importance, as the ignition timing of the mixture will vary depending on this (among other factors). An application of the HS is the heater ring introduced in Chapter 3. The temperature of this element is non-uniform, as illustrated by the uneven brightness of the heater surface during normal operation. In addition, during the fuel injection event, the jets cause a temperature drop at the location of the impingement. To study the temperature of the heater before, during, and after a gaseous jet impinges on it, a pyrometry method was developed. This method involves, first, the measurement of the radiation intensity at two different wavelength bands, allowing for the calculation of the surface temperature without requiring knowing its emissivity. Then, the results are used to generate a correlation of temperature with radiation intensity collected 53  at a single wavelength band, allowing for the calculation of the temperature using only one set of radiation data for subsequent tests. For this method, a CMOS camera was used, producing data with high spatial and temporal resolution, ideal for studying the localized rapid cooling event from the NG jets in the HS ignition-assist method. A similar correlation method using a digital camera was developed by Maun et al. [32] However, in their study, a correlation of brightness intensity with measurements done using thermocouples was used, as opposed to the correlation presented in this thesis of radiation intensity with the results using a two-colour pyrometer. The developed hybrid pyrometer method was used to calculate the temperature of a stainless steel prism in order to validate it. And then, the method was used to calculate the temperature distribution of a HS, in this case, of the HSI heater ring. This was done at both steady-state and transient-state as a result of a gaseous jet impinging on its surface.   4.2 Pyrometry theory Pyrometry consists in the measurement of the thermal radiation emitted by a surface to calculate its temperature. The total thermal radiation energy is proportional to the fourth power of the temperature, as mathematically described by the Stefan–Boltzmann law equation. [56] However, in certain applications, only the energy emitted at a certain wavelength range is measured. In those cases, Plank’s Law is used (4.1) to calculate the temperature of the surface.   𝐸𝑏𝜆(𝜆, 𝑇) =  𝐶1𝜆5[𝑒𝐶2𝜆𝑇⁄ − 1] (4.1) 54  E is the spectral emissive power (typically given in W/m3 or W/m2·μm), λ is the wavelength (nm or μm), T is the absolute temperature of the surface (K), and C1 and C2 are constants, with a value of 3.74177 ✕ 108 W·μm4 /m2 and 1.43878 ✕ 104 μm·K, respectively.  Note that the subscript b in the spectral emissive power (E) stands for black body. For real surfaces, the emissive power is less than for the idealized black body case. This relationship is represented by equation (4.2), where the spectral hemispherical emissivity (ϵλ) is the fraction of spectral emissive power of a real body with respect to a black body.   𝜖𝜆(𝜆, 𝑇) =  𝐸𝜆(𝜆,𝑇)𝐸𝑏𝜆(𝜆,𝑇) (4.2) As the amount of radiation that a surface emits varies with the direction, sometimes it is more useful to work in terms of radiation intensity, which represents the radiation emitted per solid angle. Thus, the spectral emissive power described in equation (4.1) can be replaced by the spectral radiation intensity, resulting in the following equation:  𝐼𝑏𝜆(𝜆, 𝑇) =  𝐶1𝜆5[𝑒𝐶2𝜆𝑇⁄ − 1]∗𝜋 (4.3) I is the spectral radiation intensity (W/m3·sr) for a black body. Likewise, the spectral emissivity equation can also be described in terms of radiation intensity:  𝜖𝜆(𝜆, 𝑇) =  𝐼𝜆(𝜆,𝑇)𝐼𝑏𝜆(𝜆,𝑇) (4.4)  In single-colour pyrometry, the radiation intensity is measured at a single wavelength. Then, assuming that the spectral emissivity of the studied surface is known, the temperature is calculated by combining equations (4.3) and (4.4). However, the spectral emissivity of a surface generally varies with wavelength, temperature, and surface condition. [57] This impedes the use 55  of a constant emissivity value, affecting the accuracy of the temperature calculation. To overcome this, two-colour pyrometry is used. In a two-colour pyrometer, the temperature of a surface is calculated assuming that the emissivity stays constant at two different wavelengths. This avoids the need to know the emissivity of the surface, with a penalty in the accuracy of the results in case the constant emissivity assumption is incorrect. [22] To use the two-colour method, the spectral emissivity described in equation 4.4 is modified by defining it in terms of radiation intensities emitted by a black body. This is done by equating the radiation intensity that a non-black body would emit at a temperature T with the radiation intensity that a black body would emit at an apparent temperature (Ta) at a given wavelength, as seen in equation 4.5  𝐼𝜆(𝜆, 𝑇) =  𝐼𝑏𝜆(𝜆, 𝑇𝑎) (4.5) Then, equation 4.4 is combined with equation 4.5:  𝜖𝜆(𝜆) =  𝐼𝑏𝜆(𝜆,𝑇𝑎)𝐼𝑏𝜆(𝜆,𝑇) (4.6) As two-colour pyrometry assumes that the emissivity is the same at different wavelengths, and using equation 4.3 to represent the spectral radiation intensity, the following equation can be derived:  𝑒𝐶2𝜆1𝑇⁄ −1𝑒𝐶2𝜆1𝑇𝑎1⁄ −1=𝑒𝐶2𝜆2𝑇⁄ −1𝑒𝐶2𝜆1𝑇𝑎2⁄ −1 (4.7) Ta1 is the apparent temperature at wavelength one λ1, and Ta2 is the apparent temperature at wavelength two λ2.  It can be seen that the temperature of a surface can be calculated by knowing the apparent temperature of the surface at two different wavelengths, as long as the surface behaves as a grey body (emissivity independent of wavelength).   56  4.3 Apparatus and procedure  In two-colour pyrometry, the radiation intensity at two different wavelengths should be measured in order to calculate the apparent temperatures needed to solve for T. This can be done indirectly by measuring the brightness intensity of a surface and assigning a radiation intensity according to a calibrated light source. For the tests presented in this thesis, a Phantom monochrome camera (model 3815M) was used. Using a digital camera allows to increase the spatial and temporal resolution of the results compared to other methods, but can lead to additional uncertainty errors in the results. [58]  The camera consisted in a CMOS digital camera with a bit depth of 12-bits. To focus the image, a Nikon 60 mm f/2.8 lens was attached to the camera. To control the camera, the phantom camera control application software (version 2.5) was used, allowing for the adjustment of the resolution of the image, the sample rate, the exposure time and the frames recorded. For the correlation of brightness intensity with radiation intensity, a 150 W lamp attached to a Labsphere 200 mm integrating sphere (figure 4.1) was used. This device was used as reference light source, where the brightness level read by the camera at a given wavelength corresponded to a radiation intensity value given by the manufacturer. As the response of the CMOS sensor is not linear, the calibration points were used to do a linear extrapolation.  57   Figure 4.1. Integrating Sphere Used for the Calibration Process.  After the calibration process was done, the test object took the place of the sphere, keeping the same distance and orientation with respect to the camera lens. The temperature of this test object was increased until the surface facing the camera reached a glowing temperature. Then, the CMOS camera was used to measure the brightness intensity of the surface, using a resolution of 352x352 (2.2 pixels per mm), and one of the bandwidth filters attached to the lens of the camera. This was done while keeping the same aperture as in the calibration process. However, the exposure time was adjusted to maximize the range of grey tone levels read across the surface. The calibration factor calculated from the calibration process was adjusted according to the ratio of exposure times between the calibration and the experiments. This as the intensity from the integrating sphere was higher than the intensity from the test objects. Finally, the brightness level of each pixel read in the experiments was used to calculate the spectral radiation intensity, which 58  was then used to calculate the apparent temperature of each pixel. This process was repeated twice in order to use both wavelength filters. For the calibration process and the experiments, the wavelengths selected were the 700 nm and the 800 nm. This to improve the signal-to-noise ratio of the CMOS sensor, and according to the temperature range expected in the test objects (around 900 °C, as inferred from their incandescence), so that a high rate of change of radiation intensity with respect to temperature could be achieved.  Two different test objects were analyzed in the experiments. The first object consisted in a stainless steel prism (SS 303) heated up by a furnace heater. The second object consisted in a HSI heater ring, which was similar to the one discussed in Chapter 3. The experimental setup rig and results from the testing of both objects will be presented next.  4.4 Development and validation of the hybrid pyrometer method The hybrid pyrometer method consists in calculating the steady-state temperature of a surface by using two-colour pyrometry and then using the results to generate a correlation with the radiation intensity collected at a single colour (wavelength). A flowchart with the process followed to produce the results using this method can be seen in figure 4.2. 59   Figure 4.2. Flowchart of the Hybrid Pyrometer Method  To validate this method, a small stainless steel (type 303) prism was used, comparing the results with the results collected using a commercial single-colour pyrometer and a pair of thermocouple probes at steady-state.  4.4.1 Two-colour thermal imaging The first part of the hybrid pyrometer method consists in using two-colour pyrometry to calculate the temperature of a surface. This allows to calculate its temperature without requiring its emissivity value. To test the validity of the two-colour pyrometry results, a stainless steel prism was used (figure 4.3). 60   Figure 4.3. Multiview Projection of the SS Prism Object. Dimensions shown are in inches. Stainless steel type 303 was used.   This prism was designed in such a way that a Robertshaw ER1408 furnace heater could be positioned at a close distance, which would cause the prism to glow due to an increase in its temperature. Two holes were bored in the prism to insert a pair of type-k thermocouple ungrounded 1/16-inch probes (accuracy of 0.4%).    61   Figure 4.4. Test Rig for the SS Prism Tests. A post was used in front of the furnace heater to block its view to the camera.  The test rig (figure 4.4) was constructed by placing the prism on top of refractory bricks to reduce heat losses, protect the bench where the tests were conducted, and position the prism at a similar height as the height of the CMOS camera. The furnace heater was held in place by using a t-slot aluminum post. The camera was located at a distance of about 1 m from the surface of the prism facing the camera. To power up the heater, a variac AC voltage regulator was used.  To reach a glowing temperature in the prism, a voltage between 110 V and 120 V was set in the variac. The thermocouples and DAS were used to register the temperature of the prism and identify the moment when this object reached steady-state. When this happened, the camera was used to read the brightness intensity of the glowing surface of the prism at a given exposure time using the 700 nm filter and the 800 nm filter (one at a time). Figure 4.5 shows the raw image of 62  one of the frames collected with the camera when using an input voltage of 115 V and the 800 nm bandpass filter.   Figure 4.5. Raw Image of the SS Prism Taken by the Camera  The brightness intensity of the prism read by the camera was averaged over 100 frames to reduce the standard error of the mean. Then, following the procedure described in section 4.3, the temperature of each pixel in the image was calculated. The steady-state temperature of the SS prism when applying a voltage of 115 V to the furnace heater can be seen in figure 4.6. 63   Figure 4.6. Temperature Map of the SS Prism when Applying a Voltage of 115 V to the Heater  To measure the temperature of the prism using the commercial pyrometer introduced in Chapter 3, a surface emissivity of 0.85 was set according to the reported emissivity of oxidized SS 303 at a temperature of 850 °C [59]. To run the experiments, the pyrometer was mounted to a moving base, so that its lens was at the same height as the location of the SS prism (figure 4.7).   Figure 4.7. Test Rig Setup of the SS Prism Tests Using a Commercial Pyrometer  64  The temperatures at five different locations (figure 4.8) were measured by adjusting the height and position of the base. Two of these locations coincided with the locations of the thermocouple probes. The probes were positioned less than 1 mm below the surface of the prism facing the camera, trying to have a measurement of the TC close to the surface temperature of the prism.  Three different temperature results were produced by adjusting the voltage input of the furnace heater to 110 V, 115 V and 120 V. The temperatures measured using the three methods in question (two-colour method, Ultimax commercial pyrometer, and thermocouple probes) were compared with one another by plotting the results (figure 4.8).    65    Figure 4.8. Results Comparison of the SS Prism Tests. The temperature map at top-right corner was generated from the 110 volts case. The circles in the map represent the five regions analyzed to create the column charts. The blue circles also represent the locations of the tip of the thermocouple probes. The error bars of the two-colour method results were generated considering the standard deviation of the temperature of the pixels within the five regions analyzed. The error bars of the Ultimax results and the thermocouples results were generated considering the accuracy of each method.   The agreement between the three methods varied across the cases analyzed. For the 110 V case, the largest disagreement between the two-colour method and the Ultimax pyrometer was observed in point 1 (8.7%). For the rest of the points in this case, the two-colour method showed results between 3.7% and 0.0% higher than the Ultimax pyrometer, and between 5.7% and 3.1% 66  higher than the reading of the thermocouples. The agreement improved in the 115 V case and the 120 V case, where the disagreement between the two-colour method and the Ultimax pyrometer remained below 3.0%, and the disagreement between the two-colour method and the thermocouples remained below 4.3%.  The results produced by the two-colour method showed a relatively good agreement with the results produced by the use of the commercial pyrometer and the thermocouples. Specially, considering the shortcomings associated with them. For example, in the commercial pyrometer, the use of a constant emissivity value, and in the thermocouples, the reading of the temperature being slightly below the surface of the SS prism. Also, as it was mentioned before, the grey body assumption used in the two-colour method can bring some errors in the results. As such, the results are considered to be adequate to prove the validity of this method.    4.4.2 Single-colour thermal imaging With the results generated by the two-colour pyrometry part of the hybrid pyrometer method, the rest of the method can be tested and validated, allowing to calculate the temperature of a surface with only one set of radiation intensity data.  To do this, the SS prism and the results generated in section 4.4.1 were used, plotting them as a function of the radiation intensity measured at the 800 nm wavelength. The scatter plot generated using the results from the 115 V case can be seen in figure 4.9. 67   Figure 4.9. Plot of Temperature vs Radiation Intensity Calculated from the Results of the 115 V SS Tests Case  To fit the data to a function similar to the Stefan–Boltzmann law equation, the temperature was raised to the fourth power and a linear equation as a function of radiation intensity was calculated. From figure 4.9, it can be seen that the majority of the scattered points appear to accumulate at two regions above and below the red line. These two regions seem to form two different lines. This could be an artifact of the CMOS sensor of the camera used for the tests, as two different brightness intensity regions were appreciated during the calibration process when using the 800 nm bandpass filter, something unexpected considering the uniformity of the brightness of the calibration source. To check this, the pixels from the temperature map from figure 4.6 were split in half, and then re-plotted noting their location with respect to the temperature map. 68  The results can be seen in figure 4.10. The legend in the chart refers to the location of the pixels according to the temperature map shown.   Figure 4.10. Plot of Temperature vs Radiation Intensity According to Pixel Location  As it can be seen, the pixels from the top half of the temperature map were located at the top region of the scatter plot, and the pixels from the bottom half of the map were located at the bottom region of the plot. This seems to suggest that the artifact from the CMOS sensor that was 69  observed in the calibration process was also present in the images from the SS object tests, creating a temperature map with two slightly different regions (marked by the green and purple boxes in the temperature map from figure 4.10). The linear function shown in figure 4.9 assumed that all the pixels were part of a single region. To check the difference in the results between using this linear function and the results using the linear functions created from the green and purple points shown in figure 4.10, a radiation intensity point was evaluated. Evaluating a radiation intensity of 1.5 ✕ 107 W/(m3·sr) in the linear function assuming a single region (figure 4.9) results in a temperature of 832 °C; using the linear function from the green points results in a temperature of 852 °C; and using the linear function from the purple points results in a temperature of 811 °C. This means that the percentage difference between using the green points function and the single region function was of only 2.4%, while the percentage difference between using the purple points function and the single region function was 2.6%. As the differences were low, it was decided to use the linear function assuming a single region for the rest of the analyses. The effect of different regions of radiation intensity artificially created by the CMOS sensor is not expected to be a problem for the pyrometer tests using the HSI heater ring, as the size of the heater is considerably lower than the size of the artificial radiation regions. This means that the heater ring should be located within a same radiation intensity region.  Using the linear function shown in figure 4.9, the temperature of the SS prism was recalculated for the cases described in section 4.4.1. A comparison of the temperature maps for the 115 V case can be seen in figure 4.11. 70   Figure 4.11. Temperature Maps Comparison when Applying a Voltage of 115 V in the SS Prism Tests.  As it can be seen from figure 4.11, the map generated by the hybrid method looks more uniform than the map generated by the two-colour method. The brightest area in both maps appear to be at the same location, this is, at the middle region of the left side of the prism. To further analyze the differences between the temperature maps generated by both methods, a series of histograms were produced (figure 4.12). These histograms compared the number of pixels within a certain temperature range, in order to assess the validity of the hybrid method to represent the temperatures that would be calculated from the two-colour method.  71    Figure 4.12. Histograms with the Comparison of the SS prism Temperature Maps. The purple bars represent the overlapping of the “two-colour” (blue) bars with the “Hybrid” (red) bars.   A similar trend in the results can be seen in the three voltages analyzed. The two-colour method produced results with a wider range of temperatures, as indicated by the extra bins at the lower temperatures of each chart. The uniformity of the temperature maps using the hybrid method can be appreciated in the shorter range of temperature bins. At the bin with the highest temperature range, there was a good match in the results for the 120 V case and the 115 V case. The match was not as good in the 110 V case, where about 160 more pixels were registered using the two-colour 72  method compared to the hybrid method in the bin with the highest temperature range. Overall, a good agreement was reached, especially considering the problem with the CMOS sensor of the camera discussed before.  The advantage of using this method with only one set of radiation intensity data makes it very attractive to use it to measure the temperature of a surface at transient-state, such as the cooling event caused by the gaseous jets when impinging a HS to initiate the combustion process. To test this, the hybrid pyrometer method was used on the HSI injector heater ring.   4.5 Temperature measurement of the HSI heater ring The temperature of the HSI heater ring was calculated by using the hybrid pyrometer method described before. The tests had two main objectives: measure the surface temperature of the heater across its length at steady-state and observe the cooling effect of a gaseous jet impinging on its surface.  To run the tests, the HSI injector tip, introduced in section 3.3, was mounted to a 1-inch aluminum 6061 post. In turn, this post was mounted to a rotary table (figure 4.13).  73   Figure 4.13. Test Rig of the HSI tip on the Rotary Table.  This setup allowed the HSI tip to be located at the same height as the height of the camera lens, and to be rotated a full 360° turn while keeping the camera at a fixed position. To power the heater ring, a Keithley 2200-20-5 DC power supply was used. The first part of the tests consisted in the generation of the temperature-radiation intensity correlation. To do this, the CMOS camera was used to measure the brightness intensity of the heater ring at a given rotation angle using the 700 nm and the 800 nm bandpass filters. To generate a wide range of brightness intensities, four different voltage inputs were tested: 17 V, 18 V, 19 V and 20 V. After the data collection, the temperature map of each case was calculated (examples of these maps can be seen in figure 4.14), and then the data was collapsed in a single chart to generate the plot of temperature to the fourth power as a function of radiation intensity at the 800 nm wavelength band (figure 4.15). 74   Figure 4.14. Two-colour Method Temperature Maps for the 17 V and 20 V HSI Test Cases. The magenta lines were used to represent the outline of the HSI tip. The reflection of the brightness of the coil on the insulation was removed to avoid considering those pixels in the correlation process.    Figure 4.15. Plot of Temperature vs Radiation Intensity Calculated from the HSI Tests Results. 75  Using the equation shown in figure 4.15, the surface temperature of the heater along its periphery was produced by making a full rotation of the HSI tip. The temperature map at various rotation angles when applying a voltage of 20 V can be seen in figure 4.16.  Figure 4.16. Temperature Map of the HSI Heater at Various Rotation Angles.76  The temperature across the heater ring reached a maximum of 1024 °C and a minimum of 896 °C. The coolest region was observed to be close to where the power leads connected to the ring (figure 4.16, 180° frame), while the hottest region was observed to be in the “straight” section of the ring (figure 4.16, 0° frame). Nonetheless, even in this “straight” section, the temperature of the heater was not constant, varying within the 982 °C – 1024 °C range. In the experimental test presented in section 3.3.2, the temperature of the ring when applying a similar power input was measured to be 1001 °C. This number falls within the temperature range calculated with the hybrid pyrometer method. It is important to note that the HSI heater ring used in Chapter 3 was not the exact same heater ring used in Chapter 4. They shared the same main design, but some minor differences were present as both heaters were custom-made. The temperature of the HSI heater at steady-state was able to be generated using the hybrid pyrometer method. However, during normal operation of the heater inside a combustion engine, it is expected that the heater will experience a cooling event, as a consequence of a gaseous jet impinging on it to start the combustion process. Therefore, the hybrid pyrometer method was used to observe the cooling effect of a nitrogen jet impinging on the heater at different injection pressures, injection durations, and jet orientations. This to determine which gaseous jet would produce the largest and lowest temperature drop in the heater. The test rig for these cooling tests was based on the test rig shown in figure 4.13, with the addition of some components to implement the nitrogen jet to the tests. These components consisted in a custom made 0.7 mm nozzle5, a Parker 2-way 1/8-inch NPT solenoid valve, and some tubing and piping sections to connect the valve to a nitrogen tank (figure 4.17).                                                   5 Machined by Riley Cahill, former Westport intern.   77   Figure 4.17. Test Rig Detail for the HSI Cooling Tests.  To control the injection timing and duration, the solenoid valve was actuated by an Agilent 20 MHz function generator using an electromagnetic relay. The function generator would send a square 5 V pulse to the relay for a given duration, causing the solenoid valve to open. To measure the injection pressure, a Setra 0-100 psig range pressure transducer was installed upstream of the solenoid valve.  The ideal gas law equation was used to calculate the mass expelled during each injection event. This was done by calculating the mass of nitrogen within a section of the tubing and piping system before and after the injection event. This section is referred to as the “pre-solenoid valve” chamber, consisting in the piping and tubing section between the solenoid valve and a two-way 78  valve. A P&ID drawing of the piping connected to the valve and system to control the valve can be seen in figure 4.18.  Figure 4.18. P&ID Drawing of Components Connected to the Solenoid Valve.   To calculate the change in mass, equation 4.9 was used:  𝑚 =𝑃1∗𝑉𝑅𝑠∗𝑇−𝑃2∗𝑉𝑅𝑠∗𝑇 (4.9) Where m is the mass of nitrogen expelled from the injection event (calculated in grams), P1 is the initial pressure (before the injection event) (psi), P2 is the final pressure (after the injection event) (psi), Rs is the nitrogen gas constant (296.8 J/kg·K), V the volume of the section between the solenoid valve and the two-way valve (mL) and T the ambient temperature (K).  To measure the volume inside the pre-solenoid valve chamber, the tubing and piping section was filled with water by submerging it in a container with this liquid. Then, the water inside the section was poured in a graduated beaker for its quantification.  To run the HSI cooling tests, the pre-solenoid valve chamber was pressurized using the pressure regulator, and then the two-way valve was closed. After this, using the function generator, 79  the injection event was executed. The pressure transducer measured the change in pressure inside the chamber before and after the injection event, so that the mass expelled during the injection event could be calculated.  The last variable considered for the HSI cooling tests was the orientation of the nozzle with respect to the heater ring. Two orientations were tested. The first one, called “direct” orientation, impinged the heater ring directly. And the second one, called “side” orientation, impinged the heater ring from the side in a tangential way. In both orientations (figure 4.19), the distance from the nozzle to the point of the heater ring facing the camera was approximately the same (17 mm).   Figure 4.19. Top-view of the Nozzle Orientations Tested in the HSI Cooling Tests.  Table 4.1 shows the variables used to conduct the HSI cooling tests, resulting in a total of 12 cases. Table 4.2 shows the calculated mass of nitrogen expelled for the cases analyzed.    80  Table 4.1. Variables Tested in the HSI Cooling Tests. Injection pressure (psi) Injection duration (ms) Nozzle orientation 40 25 Direct 90 60 Side  95   Table 4.2. Mass Expelled According to Injection Pressure and Duration. Injection pressure (psi) Injection duration (ms) Mass expelled (g) 40 25 0.019 – 0.023 60 0.027 – 0.032 95 0.035 – 0.039 90 25 0.044 – 0.046 60 0.061 – 0.062 95 0.077 – 0.079  To illustrate the different effects of the variables tested, a series of temperature maps and plots were produced. The most interesting figures are included in this chapter. Two of the temperature maps created can be seen in figures 4.20 and 4.21. The cooling event from figure 4.20 was produced with the direct nozzle orientation, an injection pressure of 40 psi, and an injection duration of 60 ms, which resulted in a mass of 0.027 grams. In figure 4.21, the side orientation was used with an injection pressure of 90 psi and an injection duration of 95 ms, which resulted in a mass of 0.077 grams. The start of the actuation of the solenoid valve in both cases happened at the instant t = 0.     81   Figure 4.20. Temperature Map of the HSI heater during a Cooling Event Using the Direct Nozzle Orientation.  82   Figure 4.21. Temperature Map of the HSI Heater during a Cooling Event Using the Side Nozzle Orientation.  83  As it can be seen from the above figures, the cooling effect of the nitrogen jet when using the direct nozzle orientation (figure 4.20) was much stronger than when using the side orientation (figure 4.21). This in despite of the side orientation case having a nitrogen jet with a mass almost three times as much as the mass from the direct orientation case. In the side orientation case, the cooling of the heater ring is almost imperceptible, whereas the cooling effect in the direct orientation case can be easily noticed by looking at the middle region of the ring as soon as 0.06 s after the start of the injection event.  A series of plots were produced to see the temperature drop across certain pixels during the cooling event. For this, a central region within the heater ring (figure 4.22) was isolated, and the maximum and minimum temperatures at each frame inside this region were registered.   Figure 4.22. Region of the HSI Heater Used for the Temperature Plots.  The first analysis consisted in the comparison of the maximum and minimum temperatures inside the region from figure 4.22 in the cases shown in figures 4.23 and 4.24. 84   Figure 4.23. Maximum and Minimum Temperatures from the "40 psi, 60 ms, Direct" case.   Figure 4.24. Maximum and Minimum Temperatures from the "90 psi, 95 ms, Side" case.  In the direct orientation case (figure 4.23), the minimum temperature goes below the lower limit almost as soon as the cooling event starts. The maximum temperature rapidly decreases to about 900 °C, and then starts to rise. However, even after 9 seconds from the start of the injection 85  event, this point hasn’t returned to its initial temperature of more than 1000 °C. In the side orientation case (figure 4.24), the decrease in temperature of both the maximum and minimum points is substantially lower compared to the direct orientation case. The minimum temperature decreased by about 36 °C at the end of the cooling event, while the maximum temperature decreased by about 28 °C. After 7 seconds from the start of the injection event, the maximum temperature looks to be close to its initial temperature. The effect of the orientations on the maximum temperature inside the region was further analyzed by plotting the cooling events for the 40 psi injection pressure cases.   86   Figure 4.25. Comparison of Direct and Side Orientation Cases Using 40 psi as Injection Pressure.  In the direct orientation cases, the maximum temperature experienced a decrease of almost 100 °C for all the durations tested. In comparison, in the side orientation cases, the cooling effect appeared to be pretty similar among the durations tested, with only a very slight decrease in temperature.  87  The next analysis consisted in the comparison of the different injection pressures. To do this, the direct orientation cases were analyzed.  Figure 4.26. Comparison of Injection Pressures Using the Direct Nozzle Orientation  In figure 4.26, it can be seen that the cooling effect on the maximum temperature was higher in the 90 psi case, as it was expected.   Finally, the different injection durations were analyzed by plotting the 90 psi injection pressure cases.  88   Figure 4.27. Comparison of Injection Durations Using 90 psi as Injection Pressure  From the figure above, it can be seen that the maximum temperature in the side orientation cases stayed pretty constant, even in the case with the highest injection duration, where it decreased 89  by approximately 28 °C. For this case, the temperature returned to its original value about 7 seconds after the start of the jet injection event. On the other hand, the cooling effect in the direct orientation cases was more noticeable. Using this orientation, for the 25 ms injection duration, the maximum temperature decreased by 125 °C, compared to 134 °C for the 60 ms injection duration case. The temperature went below the lower limit for the 95 ms injection case.  In summary, the cooling effect was stronger when using the direct nozzle orientation than the side orientation, regardless of the injection pressure and duration. This is something very important to consider for the design of the HSI injector. A fuel injection event that impinges directly into the heater ring will cool it down substantially more than an injection event that impinges the heater on its side in a tangential manner, affecting the ignition delay of the fuel. Ideally, the jet should produce a low cooling effect in the hot surface while providing an ignitable mixture of air and fuel at the vicinity of the surface. The cooling tests also demonstrated the slow response of the heater ring to return to its initial temperature after the cooling event. When the temperature drop in the heater ring was of about 30 °C, the heater required close to 7 seconds to return to its initial temperature. This time would be too long in a real engine application considering the short timing between ignition events.  With regards to the steady-state tests, the temperature of the heater ring when applying a voltage of 20 V was calculated. The temperature across the heater varied between 896 °C and 1024 °C. The temperature distribution measured using the pyrometer method could be useful to improve the computational model discussed in Chapter 3 of the HSI injector. An improved model could result in the validation of the heater ring, which could validate the temperature prediction of the HSI injector under combustion engine conditions.    90  The hybrid pyrometer method developed in this chapter could be used in other applications, such as applications were a rapid temperature change is expected or when analyzing a moving test object. The simplicity of this method allows to use a digital camera and a bandpass filter to calculate the temperature of a surface, as long as a calibration is done beforehand, such as the correlation of two-colour pyrometry results with radiation intensity.    91  Chapter 5: Main Conclusions, limitations and Future Work An application of the hot-surface ignition-assist method was analyzed by predicting its temperature using heat transfer simulations. The absolute results could not be validated due to some limitations in the experiments and the software used, but a sensitivity analysis was conducted to find the most important components affecting its temperature. To study the temperature of the hot surface, a pyrometer was developed. This method consists in correlating the results of two-colour pyrometry with radiation intensity. To validate this method, the surface temperature of a SS prism was calculated and compared with other methods. Finally, the pyrometer method was used to analyze the cooling effect of nitrogen jets on a heater. The developed method looks promising to be used in other applications where the temperature of a surface in transient-state needs to be studied.  5.1 Summary and conclusions of Chapter 3 The HSI injector is an application of the HS ignition-assist technology that looks to address some of the challenges associated with the use of other HS devices, such as GP. As the HS and fuel injector are integrated in the same package, the temperature of the HS could cause an excessive temperature in the injector, producing undesirable results, such as damage to the injector or pyrolysis of the fuel. For this reason, a heat transfer simulation was done to study the temperature of the injector under combustion engine conditions. This was done using ANSYS Icepak. To validate the simulation of the HSI injector model, a series of experiments and simulations of increasing complexity were conducted, where some of the main components from the HSI injector model could be validated. A “shell” of the HSI injector equipped with the heater ring was tested in the final stage of the validation experiments. These tests were done inside a combustion vessel. The measured internal temperatures of the HSI injector shell showed overall good agreement with 92  the temperatures simulated, but the measured temperature of the heater ring was higher than the simulated temperature. This was attributed to be a consequence of the limitations associated with the pyrometer used to measure the temperature of the heater ring, such as the measurement being constrained to a single spot, neglecting the temperature variations that could be observed in the heater ring. In addition, the model of the heater ring that was generated in Icepak had some differences from the real object due to limitations with the design tools of the software. For these reasons, and because the tests inside the combustion vessel did not match the conditions that the injector would face inside a combustion engine, it was concluded that the temperatures predicted in the HSI injector simulation could not be validated. However, the HSI injector model was used to check the sensitivity of the main parameters in the temperature of the HSI injector, including the heater ring.   The main conclusions from this analysis were the following:  The variables with the largest effect on the temperature of heater ring were the power input and the emissivity of the heater ring surface. The emissivity of the heater ring is a property that could change as a consequence of repeated combustion events. An increase in the emissivity will require an increase in the power input in order to achieve the same heater ring surface temperature.   The temperature of the coolant was the variable with the largest effect on the internal temperature of the injector, which suggests that an efficient cooling system is needed in order to keep the injector at a relative low temperature.  The zirconia insulation layer had a larger effect on the temperature of the heater ring than on the temperatures of the injector. This suggests that the main role of the insulation layer 93  is to preserve a high temperature in the heater ring. An insulation layer with good insulating properties could reduce the input power requirements to keep a high surface temperature on the heater.   Using a wattage of 100 W, the simulation of the HSI injector model under combustion engine conditions resulted in a temperature of more than 1300 °C in the heater ring and temperatures below 200 °C in the injector. These temperatures, however, could not be validated.    The numerical prediction tool used for the simulations (Icepak) proved to be an adequate tool to simulate the temperature of the HSI injector, with certain limitations considering the lack of validation of the temperature of the heater ring object. The tool could be useful for other HS applications and designs of the HSI injector to predict temperatures under combustion engine conditions and run sensitivity analyses. This considering the limitations associated with the design tools of the software that can complicate the generation of the computational model.   5.2 Summary and conclusions of Chapter 4 Having a constant high temperature in the HS is crucial to achieve quick and consistent ignition events. As the HS is subject to gas impingements during normal operation, its temperature is expected to decrease momentarily, affecting the ignition of the fuel. To study the temperature of the HS at steady-state and during the gas impingement cooling events, a remote temperature measurement method was developed. This method consisted in a hybrid pyrometer method using a high-speed digital camera, with the potential to produce results with high spatial and temporal resolution. The developed pyrometer method works by calculating the steady-state temperature of 94  a surface using two-colour pyrometry, and then correlating the results from the two-colour method with the radiation intensity at a single wavelength for the subsequent tests, such as tests at transient-state. For the two-colour pyrometry part of the method, the temperature is calculated by measuring the radiation at a given wavelength using a bandpass filter, and then repeating the process to measure the radiation at a second wavelength using the corresponding bandpass filter. Then, the temperature results are raised to the fourth power and fitted to a linear equation as a function of the radiation intensity at a given wavelength. This method was validated using a SS prism. The validation process was divided into two sections. In the first section, the two-colour pyrometer results generated as part of the hybrid method were compared with the results using a single-colour commercial pyrometer and thermocouple probes. Then, in the second section, the temperature map of the SS prism calculated by two-colour pyrometry was compared with the temperature map calculated by the hybrid pyrometer method. The map created by the hybrid method presented a more uniform temperature gradient, as opposed to the map created by two-colour pyrometry. This was attributed to a malfunction of the high-speed camera CMOS sensor. Overall, a good agreement was achieved between two-colour pyrometry results and the single-colour pyrometer and TC results, as well as between the temperature map of the two-colour pyrometry with the map of the hybrid pyrometer method. The hybrid pyrometer method was used to analyze the temperature of the HSI injector heater ring. For an input voltage of 20 V in the heater, a maximum temperature of 1024 °C and a minimum temperature of 896 °C were calculated across its length. For the transient tests, a cooling event was produced by using a nitrogen jet impinging on the surface of the heater ring. Three variables were adjusted to generate a series of distinct nitrogen jets: injection pressure (40 psi and 90 psi), injection duration (25 ms, 60 ms and 95 ms), and jet orientation relative to the heater ring 95  (direct and side). Ideal gas law was used to quantify the mass of each nitrogen jet tested, with the mass ranging from 0.019 g to 0.079 g depending on the duration and pressure.  The main conclusions from the cooling tests were the following:  The side orientation produced a weaker cooling effect on the heater ring compared to the direct orientation. This was true even when using a gas jet with a larger mass in the side orientation case than the direct orientation case.   When focusing on the area of the heater ring impinged by the nitrogen jet, for an injection pressure of 40 psi and using the direct orientation, the maximum temperature observed showed a temperature drop of at least 100 °C in all the injection duration cases. The temperature did not return to its original value after 9 seconds from the start of the injection event.    When focusing on the area of the heater ring impinged by the nitrogen jet, for an injection pressure of 90 psi, an injection duration of 95 ms, and using the side orientation, the maximum temperature observed showed a temperature drop of about 28 °C. The temperature returned to its original value about 7 seconds after the start of the injection event.   The hybrid pyrometer method developed showed to be a good alternative to measure surface temperatures in transient cases. The method combines the advantages of using a two-colour pyrometer, such as avoiding the use of a constant emissivity value, with the simplicity of a single-colour pyrometer, such as needing the radiation intensity at only a single wavelength. This method is able to produce results with high spatial and temporal resolution, increasing its attractiveness.   96   5.3 Future work The analyses and tools discussed in this thesis should be able to contribute in the development of HS applications, including Westport’s HSI injector. Based on the results and the limitations expressed before, the recommended future works are the following:   Sensitivity analysis of the thermal conductivity of stagnant fluid objects in the simulations. In the HSI injector under combustion engine conditions simulation, the heat transfer by conduction between the radiating surfaces and the stagnant fluid objects had a larger effect on the temperatures than expected. The sensitivity of the thermal conductivity of the fluid objects should be studied to address its effect on the temperature results.  Heat transfer simulation focused on the heater ring at transient-state. The heat transfer simulation presented in Chapter 3 was done to analyze the HSI injector at steady-state. It would be interesting to perform a heat transfer simulation at transient-state focused on the heater ring during the fuel injection events, so that the temperature of the heater ring during these events may be studied.   A different heater ring computational model to have a better representation of the actual object. The created model of the heater ring had some differences compared to the actual object. The actual heater had a coiled shape and showed a non-uniform temperature across its surface. These two aspects of the heater could not be replicated in Icepak. A software with stronger modelling capabilities should be used to represent the heater ring in a more realistic way.   97   Analysis of the degradation of the heater ring surface and zirconia insulation layer. As the heater ring wants to be used as the ignition source for the NG, degradation is expected to occur on the surface of the heater ring and possibly on the zirconia insulation layer. In particular, from the temperature perspective, repeated combustion events could change the radiation properties of the Kanthal APM by increasing its surface emissivity, which would increase the power requirements. Likewise, a change in the insulating properties of the zirconia layer would also affect the temperatures across the heater and injector.  Use of a different digital camera for the hybrid pyrometer method. As it was mentioned, two distinct brightness regions were identified during the calibration process that seem to have affected the measurements of the SS prism. Although this problem is not expected to have affected the HSI heater tests, using a different camera that does not show this issue could provide a better correlation of two-colour pyrometry results with radiation intensity, improving the validity of the hybrid pyrometer method results.   Use of natural luminosity brightness as the temperature correlation variable in the hybrid method. The temperature results from two-colour pyrometry were correlated with radiation intensity in the hybrid pyrometer method. The results could have been correlated to the natural luminosity brightness intensity instead, which would reduce the need to use a bandpass filter to restrict the radiation at a given wavelength. This would be similar to the correlation done by Maun et al [32], simplifying the tests. This could open the possibility to use the hybrid method in a more common digital camera, as no extra components would be needed.  Additional injection parameters tested in the cooling tests under ambient conditions. Each of the cooling tests introduced in Chapter 4 presented jets with different masses. It would 98  be interesting to see the different cooling effects (if any) that jets with the same mass would produce by varying their injection pressure and duration. For example, a cooling jet with lower pressure but longer duration could have the same mass as another cooling jet with higher pressure but shorter duration.   Analyze the temperature loss across the heater ring. The cooling effects of the gaseous jets were analyzed by isolating a region within the heater and plotting a given pixel from this region. It would be interesting to analyze the cooling effect by considering the total average temperature of the heater, so that the total energy (i.e. temperature) loss of the heater could be studied as a function of momentum of the jet.   Repeatable cooling jet parameters. The orientation and location of the jets used for the pyrometer cooling tests were hard to control with precision. This complicated the repeatability of the tests. A different test rig and nozzle design should be used to improve this, as the cooling effect is highly sensitive to the orientation and location of the gaseous jet.    Use of a gas injector able to better replicate the pilot injection jets. The solenoid valve used for the cooling tests from Chapter 4 could not be opened for a duration of less than 20 ms. This value is considerably higher than what the duration of the pilot jets injection is expected to require in a real combustion application. In addition, the 0.7 mm size of the orifice could have been high, considering the small pilot jets that are intended to be used to initiate the combustion of NG. For this reason, running the pyrometer cooling tests with a real gas injector could be interesting, analyzing the cooling effect of the actual pilot jets.    Application of the hybrid pyrometer method and CMOS camera to calculate the surface temperature of the heater ring under combustion operating conditions. The tests presented 99  in Chapter 4 were done using the heater ring under ambient conditions. Testing the actual HSI injector under conditions similar to combustion engine conditions would be an interesting application of the pyrometer method, calculating the temperature of the heater ring just before the ignition starts, as well as identifying the location of the hottest spots in the heater, which could improve the design of the heater ring and the orientations and durations of the pilot NG jets.  100  Bibliography  1.  Greenstone M, Fan CQ. Introducing the Air Quality Life Index [Internet]. Energy Policy Institute at the University of Chicago; 2018. Available from: https://aqli.epic.uchicago.edu/wp-content/uploads/2018/11/AQLI-Annual-Report-V13.pdf  2.  Lelieveld J, Klingmüller K, Pozzer A, Pӧschl U, Fnais M, Daiber A, Münzel T. Cardiovascular disease burden from ambient air pollution in Europe reassessed using novel hazard ratio functions. European Heart Journal. 2019.   3.  Karagulian F, Belis CA, Dora CFC, Prüss-Ustün AM, Bonjour S, Adair-Rohani H, Amann M. Contributions to cities’ ambient particulate matter (PM): A systematic review of local source contributions at global level. Atmospheric environment. Elsevier; 2015;120:475–83.   4.  Annual Energy Outlook 2019 [Internet]. US Energy Information Administration; [cited 2019 Mar 19]. Available from: https://www.eia.gov/outlooks/aeo/data/browser/#/?id=58-AEO2019&cases=ref2019&sourcekey=0  5.  Harrington J, Munshi S, Nedelcu C, Ouellette P, Thompson J, Whitfield S. Direct injection of natural gas in a heavy-duty diesel engine. SAE Technical Paper; 2002.   6.  Heywood JB. Internal combustion engine fundamentals. 1st ed. McGraw-Hill; 1988.   7.  Zimmerman N, Wang JM, Jeong C-H, Wallace JS, Evans GJ. Assessing the climate trade-offs of gasoline direct injection engines. Environmental science & technology. ACS Publications; 2016;50(15):8385–92.   8.  Aesoy V, Valland H. Hot surface assisted compression ignition of natural gas in a direct injection diesel engine. SAE transactions. JSTOR; 1996;1022–30.   9.  Zhao F, Lai MC, Harrington DL. Automotive spark-ignited direct-injection gasoline engines. Progress in energy and combustion science. Elsevier; 1999;25(5):437–562.   10.  Kubesh J, King SR, Liss WE. Effect of gas composition on octane number of natural gas fuels. SAE Technical Paper; 1992.   11.  Fraser RA, Siebers DL, Edwards CF. Autoignition of methane and natural gas in a simulated diesel environment. SAE transactions. JSTOR; 1991;33–45.   12.  McTaggart-Cowan G, Mann K, Huang J, Wu N, Munshi S. Particulate Matter Reduction from a Pilot-Ignited, Direct Injection of Natural Gas Engine. ASME 2012 Internal Combustion Engine Division Fall Technical Conference. 2012. p. 427–37.    101  13.  Huang J, Munshi S, McTaggart-Cowan GP, Wagner DR, inventors; Westport Power Inc., assignee. Apparatus and method for igniting a gaseous fuel in a direct injection internal combustion engine. United State patent US 9,790,868. 2017 Oct 17.   14.  Aesoy V, Valland H. The influence of natural gas composition on ignition in a direct injection gas engine using hot surface assisted compression ignition. SAE Technical Paper; 1996.   15.  Abate V. Natural gas ignition delay study under diesel engine conditions in a combustion bomb with glow plug assist [master’s thesis]. University of Toronto; 2001. 94 p.  16.  Fabbroni M, Wallace JS. Flame propagation in natural gas fueled direct injection engines. ASME 2010 Internal Combustion Engine Division Fall Technical Conference. 2010. p. 235–45.   17.  Pan K, Wallace JS. Numerical studies of the ignition characteristics of a high-pressure gas jet in compression-ignition engines with glow plug ignition assist: Part 1—Operating condition study. International Journal of Engine Research. London, England: SAGE Publications; 2017;18(10):1035–54.   18.  Chown D. Characterization of hot surface ignition in a direct injected natural gas combustion apparatus [master’s thesis]. University of Toronto; 2014. 109 p.  19.  Gogolev IM, Wallace JS. Study of Assisted Compression Ignition in a Direct Injected Natural Gas Engine. Journal of Engineering for Gas Turbines and Power. American Society of Mechanical Engineers; 2017;139(12).   20.  Okada M. Development of CNG Direct Injection Diesel-Cycle Engine. International Conference of NG vehicles. 2004.   21.  McTaggart-Cowan G, Huang J, Turcios M, Singh A, Munshi S. Evaluation of a Hot-Surface Ignition System for a Direct-Injection of Natural Gas Engine. Proceedings of ASME 2018 Internal Combustion Engine Division Fall Technical Conference. San Diego, CA, USA: American Society of Mechanical Engineers; 2018.   22.  Pyatt E. Some consideration of the errors of brightness and two-colour types of spectral radiation pyrometer. British Journal of Applied Physics. IOP Publishing; 1954;5(7):264.   23.  Ueda T, Sato M, Nakayama K. The temperature of a single crystal diamond tool in turning. CIRP Annals. Elsevier; 1998;47(1):41–4.   24.  Furumoto T, Ueda T, Alkahari MR, Hosokawa A. Investigation of laser consolidation process for metal powder by two-color pyrometer and high-speed video camera. CIRP Annals. Elsevier; 2013;62(1):223–6.   102  25.  Müller B, Renz U. Development of a fast fiber-optic two-color pyrometer for the temperature measurement of surfaces with varying emissivities. Review of scientific instruments. AIP; 2001;72(8):3366–74.   26.  Thevenet J, Siroux M, Desmet B. Measurements of brake disc surface temperature and emissivity by two-color pyrometry. Applied Thermal Engineering. Elsevier; 2010;30(6-7):753–9.   27.  Tapetado A, Díaz-Álvarez J, Miguélez MH, Vázquez C. Two-color pyrometer for process temperature measurement during machining. Journal of lightwave technology. IEEE; 2016;34(4):1380–6.   28.  Al Huda M, Yamada K, Hosokawa A, Ueda T. Investigation of temperature at tool-chip interface in turning using two-color pyrometer. Journal of manufacturing science and engineering. American Society of Mechanical Engineers; 2002;124(2):200–7.   29.  Anselmi-Tamburini U, Campari G, Spinolo G, Lupotto P. A two-color spatial-scanning pyrometer for the determination of temperature profiles in combustion synthesis reactions. Review of scientific instruments. AIP; 1995;66(10):5006–14.   30.  Densmore JM, Biss MM, McNesby KL, Homan BE. High-speed digital color imaging pyrometry. Applied Optics. Optical Society of America; 2011;50(17):2659–65.   31.  Kuhn PB, Ma B, Connelly BC, Smooke MD, Long MB. Soot and thin-filament pyrometry using a color digital camera. Proceedings of the Combustion Institute. Elsevier; 2011;33(1):743–50.   32.  Maun JD, Sunderland PB, Urban DL. Thin-filament pyrometry with a digital still camera. Applied optics. Optical Society of America; 2007;46(4):483–8.   33.  Bardin F, Morgan S, Williams S, McBride R, Moore AJ, Jones JDC, Hand DP. Process control of laser conduction welding by thermal imaging measurement with a color camera. Applied optics. Optical Society of America; 2005;44(32):6841–8.   34.  Dagel DJ, Grossetete GD, MacCallum DO, Korey SP. Four-color imaging pyrometer for mapping temperatures of laser-based metal processes. Proceedings of SPIE, Thermosense: Thermal Infrared Applications XXXVIII. Baltimore, MD, USA; 2016.   35.  Hauer W, Zauner G. High-temperature dual-band thermal imaging by means of high-speed CMOS camera system. Proceedings of SPIE, Image Processing: Machine Vision Applications VI. Burlingame, CA, USA; 2013.   36.  Cengel YA, Cimbala JM. Fluid mechanics: Fundamentals and applications. 1st ed. McGraw-Hill; 2006.   103  37.  ANSYS Icepak user’s guide. Release 17.2. ANSYS, Inc.; 2016.   38.  Kanthal APM Resistance heating wire and resistance wire. Datasheet. [Internet]. Kanthal; 2018 [cited 2019 Mar 26]. Available from: https://www.kanthal.com/en/products/material-datasheets/wire/resistance-heating-wire-and-resistance-wire/kanthal-apm/  39.  2600 Insulating Fire Brick Data [Internet]. BNZ materials Inc.; 2013 [cited 2018 Aug 27]. Available from: http://www.bnzmaterials.com/insulating-firebrick/ifb-2600/  40.  Mikaél’A B. Infrared radiation: a handbook for applications. Springer Science & Business Media New York; 1968.   41.  ASM metals reference book: A handbook of data and information. American Society for Metals; 1981.   42.  Otsuka A, Hosono K, Tanaka R, Kitagawa K, Arai N. A survey of hemispherical total emissivity of the refractory metals in practical use. Energy. Elsevier; 2005;30(2):535–43.   43.  Table of emissivity of various surfaces [Internet]. Transmetra; [cited 2019 Mar 27]. Available from: https://www.transmetra.ch/dokumentation/dokumentation-publikationen/category/123-pyrometrie-thermografie  44.  Shackelford JF, Han YH, Kim S, Kwon SH. CRC materials science and engineering handbook. 4th ed. CRC press; 2016.   45.  Tanaka H, Sawai S, Morimoto K, Hisano K. Measurement of spectral emissivity and thermal conductivity of zirconia by thermal radiation calorimetry. Journal of thermal analysis and calorimetry. Springer; 2001;64(3):867–72.   46.  Davis JR. ASM specialty handbook: tool materials. ASM international; 1995.   47.  Singham J. Tables of emissivity of surfaces. International Journal of Heat and Mass Transfer. Elsevier; 1962;5(1-2):67–76.   48.  Trémolet de Lacheisserie É, Schlenker M, Gignoux D. Magnetism: Materials and applications. Springer Science + Business Media, Inc.; 2005.   49.  ASM ready reference: thermal properties of metals. ASM International; 2002.   50.  Yovanovich MM. Thermal interface (joint) conductance and resistance [Internet]. University of Waterloo; [cited 2018 Nov 25]. Available from: http://mhtlab.uwaterloo.ca/courses_old/ece309/notes/conduction/cont.pdf   104  51.  Lockwood F, Zhang Z, Cho S, Wang J. Thermal characteristics of new and used diesel engine oils. Proceedings of 2nd world tribology congress. Vienna, Austria: Austrian tribology society; 2001. p. 1–6.   52.  Kallaev S, Gadzhiev G, Kamilov I, Omarov Z, Sadykov S, Reznichenko L. Thermal properties of PZT-based ferroelectric ceramics. Physics of the Solid State. Springer; 2006;48(6):1169–70.   53.  Kluk P, Milewski A, Kardys W, Kogut P, Michalski P. Measurement system for parameter estimation and diagnostic of ultrasonic transducers. Optical and acoustical methods in science and technology. Institute of Physics, Polish Academy of Science; 2013;124(3):468–70.   54.  Singh AP. Characterization and system level study of air addition in a pilot ignited direct injection natural gas engine [master’s thesis]. University of British Columbia; 2019. 136 p.  55.  Woschni G. A universally applicable equation for the instantaneous heat transfer coefficient in the internal combustion engine. SAE Technical paper; 1967.   56.  Cengel YA, Ghajar AJ. Heat and mass transfer: Fundamentals and applications. 4th ed. McGraw-Hill; 2014.   57.  Kobayashi M, Ono A, Otsuki M, Sakate H, Sakuma F. A database of normal spectral emissivities of metals at high temperatures. International journal of thermophysics. Springer; 1999;20(1):299–308.   58.  Usui H, Mitsui K. Accuracy of two-color pyrometry using color high-speed cameras for measurement of luminous flames. Proceedings of SPIE, 27th International Congress on High-Speed Photography and Photonics. Xian, China; 2007.   59.  Wade WR. Measurements of total hemispherical emissivity of various oxidized metals at high temperature. Technical note 4206. National advisory committee for aeronautics; 1958.   60.  Talukdar P, Das A, Alagirusamy R, others. Simultaneous estimation of thermal conductivity and specific heat of thermal protective fabrics using experimental data of high heat flux exposure. Applied Thermal Engineering. Elsevier; 2016;107:785–96.   61.  Lawson JR, Walton WD, Bryner NP, Amon FK. Estimates of thermal properties for fire fighters’ protective clothing materials. US Department of Commerce, National Institute of Standards and Technology; 2005.     105  Appendices  Appendix A   Steps followed to generate the heat transfer models in Icepak To simulate the experiments that were presented in sections 3.3 and 3.4, the following steps were carried out: 1. The dimensions of the domain and the ambient conditions were specified, such as the temperature and pressure. 2. The model was created using predefined objects, such as cylinders and prisms. 3. The properties of the materials used for each of the components built in the model were set. 4. The boundary conditions for each component were set. This included the designation of heat generation in certain components and sides of adjacent components with a certain contact resistance. 5. A priority level was assigned to each of the objects modelled. The priority levels are used by Icepak to assign mesh and properties when objects intersect with one another or share a common surface.  6. The mesh was created using Icepak’s semi-automated meshing tool. The mesh type was set to “mesher-HD”, which consisted in a mesh containing mostly hexahedral cells. The density of the mesh was higher near the boundary of the objects simulated, and decreased in the open spaces within the system domain. To control the meshing parameters in the model at a local level, regions, defined by Icepak as “assemblies”, were defined around a group of elements depending on their size.  106  7. The basic parameters were set. These parameters included the flow regime and the gravity vector direction used to calculate the natural convection heat transfer.  8. The solver was ran until the equations described before converged to the solution.    107  Appendix B  Preliminary stage one validation tests: Heat transfer using a cartridge heater For the first of the preliminary stages, a simple experiment and simulation of heat transfer by conduction was carried out. This in order to have a better understanding of the tools and settings in Icepak to carry out the more complicated simulations, such as the simulations of the validation stages.  For this exploratory stage, a 9.5 cm long 1-inch aluminum 6061 round bar with a Watlow 1/8-inch cartridge heater inserted at its base, were used. An image for the test rig can be appreciated in figure B.1.   Figure B.1. Experimental Test Rig for Stage One of the Validation Tests  Eight type-k 24 AWG thermocouples were used in the experiment, with seven measuring the inner temperature of the round bar, and one measuring the ambient temperature. To register the temperatures, an Omega thermocouple data acquisition system (model: OM-USB-TC) was used. A variac AC voltage regulator was used to control the delivered power to the cartridge heater. The 108  power input was calculated using Ohm’s law by measuring the voltage drop across the heater employing a Circuit-Test Electronics DMR-3600 multimeter, and the electric current using a Fluke 179 multimeter. To reduce the uncertainties related to the contact resistance between the heater and the round bar, a Wakefield-Vette thermal paste was used. A small piece of refractory brick was used to protect the table under the heated round bar.  To conduct the test, a nominal power of 18.8 W was supplied to the heater by adjusting the voltage regulator. This power value was selected according to the maximum operating temperature of some of the equipment and materials used (approximately 200 °C). The experiment was ran for two hours, which provided enough time to reach steady-state in the system. This means that the supplied electrical power of 18.8 W was equal to 18.8 W of heat dissipated from the system to the surroundings. To appreciate the change of temperature throughout the experiment, the DAS collected the temperatures as soon as the voltage was supplied to the heater and until the system reached steady-state.  The simulation was performed using Ansys Icepak. In the model, the main elements of the test rig were incorporated, including the aluminum rod, refractory brick and cartridge heater. For the model parameters, a perfect contact conductance (i.e. contact resistance = 0) was specified between the heater and the rod, and the nominal electrical power of 18.8 W (± 0.5 W) was assumed as the constant heat generated by the heater.  For the boundary conditions, an ambient temperature of 19 °C was set, and heat transfer by radiation and by natural convection were activated. For the radiation heat transfer, Icepak assumes the surfaces of the participating objects to be opaque, diffuse and grey, and it calculates the view factors between the surfaces to solve for the thermal radiation. For the natural convection, the Boussinesq model approximation was used. In table B.1, the most relevant properties of the materials involved in the simulation are listed. 109   Table B.1. Materials’ Properties Used in the Model of Stage One Material Property  Aluminum 6061 Thermal conductivity (W/m·K) 154 – 180 [41] Emissivity 0.06 – 0.07 [40] Refractory clay (fire brick) Thermal conductivity (W/m·K) 0.23 [39] Emissivity 0.65 – 0.75 [40]  The mesh of the model was composed by 472,000 elements, with a maximum size of 0.01 m at each axis of the Cartesian coordinate system. The mesh type was set to “hex-dominant”, a mesh composed mostly by hexahedral elements. The density of the mesh was higher at the interface of adjoined objects and inside the main objects, decreasing as the distance between the main objects and the computational domain boundaries reduced. The mesh used for the model can be seen in figure B.2. 110   Figure B.2. Views of the Mesh for the Model of Stage One Focused on the Mesh around the Aluminum Rod  To validate the results from the simulation, the results are compared with the temperatures measured using thermocouples in the experiment described before. To do this comparison, the simulated temperature at a given location within the round bar model is compared with its corresponding location in the real round bar. If both values agree within the uncertainty associated of each method, the simulation results are said to be validated. The temperature map generated in Icepak can be appreciated in figure B.3. This temperature map was generated applying the nominal power of 18.8 W as the heat generation of the heater.  111   Figure B.3. Temperature Map of Model of Stage One. Regions at a temperature above the upper colour scale limit (180.0 °C) appear red. The ambient temperature was set to 19.0 °C, which is the lower colour scale limit.  The temperature gradient that can be appreciated above the round bar illustrates the effect of natural convection in the heat transfer from the bar to its surroundings. Another temperature gradient that can be easily recognized is located at the region of the refractory brick, as it has a relatively low thermal conductivity.  112   Figure B.4. Comparison of Results between Experiment and Simulation of Stage One. On the left side, a drawing of the cross-section of the round bar with the seven temperature measurement locations is shown. On the right side, the temperature results collected from each method are shown. Error bars from the “Experiment” are based on the uncertainty of the equipment used (i.e. thermocouples, DAS). Error bars from the “Simulation” are based on the uncertainty associated with the heat input used as the boundary condition (18.8 W ± 0.5 W).  The temperature drop across the aluminum bar remained relatively low, as one would expect considering the relatively high thermal conductivity of aluminum. In the experiment, the temperatures remained in the 170 °C – 177 °C range. In the simulation, in the 173 °C – 176 °C range. Overall, the temperatures agreed within the uncertainty associated with each method. This agreement demonstrates that the heating source, and its influence on the temperature of the aluminum bar, were correctly simulated by the software.   113  Appendix C  Preliminary stage two validation tests: Main body insulation  The second preliminary stage consisted in an experiment of heat transfer by conduction using a cartridge heater. This heater was used to heat up two aluminum objects of different sizes making indirect contact with each other. The larger object represented the main body of the HSI injector, and the smaller object represented the injector tip of the HSI injector. The objective of this stage was to recreate an important feature of the design of the HSI injector: the insulation of the main injector body from the injector tip, which main function is to hold the heater ring.  Although the materials and dimensions used in this experiment are not the same as those from the actual HSI injector, the validation of this stage should provide some confidence in the simulation of insulation mechanisms, such as an enclosed air gap and a layer of thermal insulation material between objects with large temperature differences.  Two aluminum objects were used for the experiment. A drawing of these two objects can be seen in figure C.1. The first object consisted of a 9.5 cm long aluminum 6061 round bar with a changing diameter. From the bottom side of the round bar to a distance of 6.9 cm, the bar had a 1 inch diameter. From 6.9 cm to the top side, the round bar had a 1 cm diameter. This object will be referred to as “main body” from now on. The second object consisted in a 2.3 cm long 1-inch hollow aluminum 6061 round bar, with an inner diameter of 1.2 cm. This object will be referred to as “dummy tip” from now on. 114   Figure C.1. Drawings of the Aluminum Objects Used for the Stage One Tests. On the left side, the front view of the main body object is shown. On the right side, the front view of the dummy tip is shown. Dimensions shown are in millimeters. Aluminum 6061 was used.  To conduct the test, the dummy tip was mounted on top of the main body. As the objective of the test was to insulate the objects from one another, a 1.35 mm thick cotton strip (a woven cotton fabric with a density of 550 kg/m3) was used to avoid having a direct contact between them, as it can be seen in figure C.2. For the same reason (insulation), an air gap between the inner side of the dummy tip and the outer side of the main body was created by using two O-rings around the top section of the main body.  115   Figure C.2. Detail of the Aluminum Bodies and Main Components for the Stage One Tests. The left side image shows the dummy tip dismounted from the main body. The right side image shows the dummy tip mounted to the main body, which is how the experiment was conducted.  To measure the temperature of the test objects, seven type-k 24 AWG thermocouples (accuracy of 2.2 °C) were used, with two measuring the temperature of the dummy tip, and five measuring the temperature of the main body. The locations of the thermocouples can be seen in figure C.6. One additional thermocouple was used to measure the ambient temperature around the test objects. To register the temperatures, an Omega thermocouple data acquisition system (model: OM-USB-TC) (accuracy of 0.35 °C per channel) was used.  A Watlow 1/8-inch cartridge heater was inserted at the base of the main body to increase the temperature of the test objects. The power of this device was controlled using a variac AC voltage regulator. The power input was calculated using Ohm’s law by measuring the voltage drop across the heater employing a Circuit-Test Electronics DMR-3600 multimeter (accuracy of 1.5% + 4 counts), and the electric current using a Fluke 179 multimeter (accuracy of 1.5% + 3 116  counts). To reduce the uncertainties related to the contact resistance between the heater and the inner surface of the main body where it was inserted, a Wakefield-Vette thermal paste was used. A small piece of refractory brick was used to protect the table under the test objects. A P&ID diagram of the test rig can be seen in figure C.3.  Figure C.3. P&ID of the Test Rig Used for the Stage One Tests. V represents the location of the multimeter used to measure the voltage drop. A represents the location of the multimeter used to measure the current.   To run the experiment, a nominal power of 18.5 W was supplied to the cartridge heater by adjusting the output voltage of the variac. The experiment was ran for about an hour and a half, concluding when the system reached steady-state. To appreciate the change of temperature throughout the experiment, the DAS measured the temperatures as soon as the variac was turned on.    117  For the simulation parameters, a perfect contact conductance (i.e. contact resistance = 0) was specified between the cartridge heater and the main body, and the properties of the materials of the participant objects were defined according to the values shown in table C.1. A nominal heat input of 18.5 W (± 0.5 W) was specified as the heat generation of the heater. An ambient temperature of 19.0 C was set, and heat transfer by radiation and by natural convection, using the Boussinesq model approximation, were activated in the simulation.   Table C.1 Materials’ Properties Used in the Model of Stage One Material Property  Aluminum 6061 Thermal conductivity (W/m·K) 154 – 180 [41] Emissivity 0.06 – 0.07 [40] Refractory clay (fire brick) Thermal conductivity (W/m·K) 0.23 [39] Emissivity 0.65 – 0.75 [40] Cotton fabric Thermal conductivity (W/m·K) 0.15 – 0.19 [60,61]   The final generated mesh of the model was composed by 491,113 elements, with a maximum size of 0.01 m at each axis of the Cartesian coordinate system. The mesh type was set to “hex-dominant”, a mesh composed mostly by hexahedral elements. To regulate the changes of mesh density across the solution domain, assemblies where specified around certain objects. Views of the mesh used for the model can be appreciated in figure C.4.   118   Figure C.4. Views of the Mesh for the Model of Stage One Focused on the Mesh of the Main Objects   Using the model, mesh and properties described above, a temperature map at the Y plane was produced, which can be seen in figure C.5. This map was generated using the nominal power value of 18.5 W as the heat generation in the heater.  119   Figure C.5. Temperature Map of Model of Stage One. Regions at a temperature above the upper colour scale limit (200.0 °C) appear red. The ambient temperature was set to 19 °C, which is the lower colour scale limit.   Two distinct regions of uniform temperature can be appreciated: one at the main body and one at the dummy tip. This was expected, as the thermal conductivity of the aluminum is relatively high, which would prevent the formation of temperature gradients in objects of this material. Alternatively, at the refractory brick, an evident temperature gradient can be seen, due to its low thermal conductivity. Also, the dummy tip can be seen as having a lower temperature than the main body. 120    Figure C.6. Comparison of Results between Experiment and Simulation of Stage One. On the left side, a drawing of the cross-section of the round bar with the seven temperature measurement locations is shown. On the right side, the temperature results collected from each method are shown. Error bars from the “Experiment” are based on the uncertainty of the equipment used (i.e. thermocouples, DAS). Error bars from the “Simulation” are based on the uncertainty associated with the heat input used as the boundary condition (18.5 W ± 0.5W). Temperature points 2 and 3 refer to the dummy tip object.   In both the experiment and simulation methods, in average, the dummy tip had a temperature 44 °C less than the main body. This demonstrates the correct simulation of the insulation properties of both the cotton fabric and air gap between the dummy tip and main body. In terms of the temperature at each measurement point, all the points agreed within the uncertainty associated with each method, except point 3, where the difference was of only 0.5 °C. In the 121  experiment, the temperature at point 3 was about 2.4 °C less than at point 2. This could mean that the thermocouple measuring the temperature at point 3 was not making proper contact with the surface of the round bar, which would reduce the temperature measured. In general, the model generated in Icepak was able to reproduce the same temperatures as those measured in the experiment, thus validating the results from the simulation.  

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:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0378384/manifest

Comment

Related Items