Hydrogen Direct Injection Technology Challenges and Opportunities NHA

Hydrogen Direct Injection Technology - Challenges and Opportunities NHA 2008 Sacramento, CA Alan Welch*, David Mumford, Sandeep Munshi, James Holbery, Brad Boyer

H 2 ICE Research Targets Emissions Target: SULEV or Better Performance Target: Transparent to Gasoline Customer H 2 ICE Research Efficiency Target: Exceed Diesel BTE Key focus remains on efficiency Implementation -Addresses range / storage directly Target: Transparent to Customer 2

H 2 Direct Injection (DI) Collaborative Research & Development Engine / Combustion Ford Motor Co. - Research Innovation Center * (Dearborn, MI, USA) * plus Sandia National Labs, Argonne National Labs Injector / Driver Westport Innovations Inc. & GVH Gmb. H (Vancouver, BC, Canada & Dortmund, Germany). Materials Pacific Northwest National Labs** (Richland, WA, USA) ** plus Oak Ridge National Labs, Argonne National Labs 3

Hydrogen Engines in Transportation– Key Enablers Relevant Research Areas In Project Scope? Combustion & emissions technology Fuel injection technology (design & materials) Fuel storage technology & cost Hydrogen availability & cost 4

Technology – Opportunities & Results 5

Single Cylinder Research Engine • 4 -valve layout like modern diesel engine • Centrally mounted direct injector / spark plug nearby • Eliminates engine tendency to backfire in the intake manifold • Minimizes pre-ignition tendency due to shorter residence time of fuel • Compression ratio varied between 9: 1 and 16: 1 6

Injector & Environmental Conditions 7

J 43 Px Injector Animation 8

9

10

Piezoelectric Actuated Injector • Needle lift mirrors voltage • Lift control capability • Multiple injection capability 11

Variable Lift Capability 12

Ford Test Cell and Engine Specifications Type Ford single cylinder design Bore 80 -92 mm, 89 typical Stroke 80 -100 mm Displacement 0. 4 to 0. 6 L Compression Ratio Variable (approx. ) 9: 1 to 16: 1 Rated speed 6000 RPM Max. speed 7000 RPM Max. cylinder pressure Number of valves Valve Sizes Valvetrain Max. valve lift Lubrication Cylinder liner 120 bar 2 intake, 2 exhaust 35 mm intake 30 mm exhaust DOHC, direct acting mechanical bucket, toothed belt, 230 deg duration event 9. 5 mm / 9. 5 mm Dry sump Wet 13

Comparison of Various Injection Strategies 14

NOx vs. Equivalence Ratio, 1500 RPM 15

Indicated Fuel Consumption vs. Equivalence Ratio, 1500 RPM 16

Projected to Multi-cylinder Engine • 3+ years of data • Different injection rates • Mostly 1500 -4500 RPM 45% Peak • Naturally aspirated & boosted. • Discrete clouds different boost condition. 17

Technical Challenges with Hydrogen 18

Fluid Density vs. Fuel Pressure at 40 ºC Relative densities at 25 MPa or 250 bar • Dodecane (C 12 H 26) = 100% • Methane = 23% • Hydrogen = 2. 2% 19

Dynamic Viscosity vs. Fuel Pressure at 40 ºC Relative viscosity at 25 MPa or 250 bar Dodecane (C 12 H 26) = 100% Methane = 1. 5% Hydrogen = 0. 7% 20

Challenges and J 43 px– Wear Series Injector Diffusion Related Challenges. Hydrogen Diffusion / Explosive Decompression in Dielectric Coating Sliding Friction & Wear Long Term Hydrogen Diffusion in Piezoelectric Actuator ? Impact Wear 21

Technical Challenges in Hydraulic Compensator Areas of Investigation • Assembly Ease • Thermal Expansion Material • Seals Properties • Hydraulic fluid (vs. T, P) Design Internal Seal - Long Term Durability Chemistry • Hydrogen diffusion • Gas solubility 22

Modeling Impact Stress with Elastic Deformation • Wear rate appears to be low but not are not fully characterized • Slight seat leakage observed from particles in the gas stream or slight surface damage after 1 to 10 million cycles. • PNNL / ORNL test set-up will be used to model sliding impact based upon empirical data. • Modeling: Peak 500 MPa compressive stress (15% of material yield stress) 23

Material Test Facilities at PNNL High Pressure / Temperature Hydrogen • Two high-pressure autoclaves: Pure H 2 4500 PSI (300 Bar & 300°C • Piezoelectric actuators (exposure): 4 actuated & 4 nonactuated • Needle/seat related sample exposure with in-situ frictionwear apparatus • Computer-based data acquisition and control. 24

Injector Piezoelectric Actuator After High Temperature/ High Pressure Exposure Delamination Outside Surface Inside Surface SEM of sectioned actuator • Delamination of dielectric coating is seen after 200 to 500 hours, typically in hotter region, extreme operation • Current epoxy coating is function outside the glass transition temperature limit (Tg =60 C) • Continued function leads to carbon tracking and short circuits under epoxy layer. 25

Summary • The Hydrogen DI engine research program has demonstrated significant benefits: – Flexible direct injection technology for research – 45% BTE has been demonstrated on a hydrogen single cylinder research engine (multi-cylinder friction used) – Advanced NOx control strategies are possible with direct injection using mode switching – High power density 26

Next Steps – Engine research: • Further efforts on peak brake thermal efficiency • Define advanced NOx control methods • Multi-cylinder research – Injection system – validate for 1000 hour durability: • Further research on piezoelectric diffusion • Improved epoxy/ dielectric coatings • Very low friction coatings / impact coatings • Accelerated injector testing methods PNNL Materials Testing / Analysis Westport Injector Rig Rapid Aging / Validation Ford Engine Aging / Validation 27

Thank You for Your Attention www. westport. com awelch@westport. com

NOx Emissions Port Injected Hydrogen Engine • Oxides of nitrogen (NOx) can be controlled to ultralow levels using very lean strategy (at or below phi = 0. 4) • EGR and NOx traps/aftertreatment at higher loads. • At vehicle, near zero CO 2, CO, HC and PM Source: SAE 2002 -01 -0242 29

Mass Flow vs. Peak Voltage for a J 43 P 3 Piezoelectric Injector 30

Hydrogen Direct Injection Engine Technology - Opportunity Technical Goals Very High Peak Efficiency Benefit Extended vehicle range with hybrid H 2 ICE Near zero CO, HC, PM Long product window & SULEV NOx societal benefits High power / torque density Engine packaging & hardware similar to gasoline/diesel Acceptance by gasoline / diesel customers Leverage existing manufacturing & reduced capital cost 31

Hydrogen Diffusion in Piezoelectric Materials – Very Long Term Possibility 32

Needle Corrosion 0 Firing Hours 5 Firing Hours 34

Needle Corrosion 82 Firing Hours 107 Firing Hours 35

36

37

• Combustion conditions: – Reactants initially at 70 atm/650 K – Combustion at constant pressure and enthalpy • Characteristic time to reach equilibrium NO for Lambda > 1. 5 is at least an order of magnitude (or more) higher than engine combustion cycle time. • For very lean (Lambda > 1. 5, AFT < 1900 K) engine out NO is very small. • H 2 can burn up to Lambda = 4 or more resulting in very low NOx. • Lambda is the relative air-to-fuel ratio. • Calculations done with STANJAN. 38