EVALUATION OF CURRENT AND FUTURE ATKINSON ENGINE TECHNOLOGIES 2 nd CRC Advanced Fuel and Engine Efficiency Workshop 11/2/2016 Charles Schenk, U.S. EPA Developmental data: internal EPA use only 1
Background As part of the rulemaking establishing the model year (MY) 2017 2025 light duty vehicle GHG standards, EPA made a regulatory commitment to conduct a Midterm Evaluation (MTE) of longerterm standards for MY 2022 2025. https://www3.epa.gov/otaq/climate/mte.htm Work began immediately following the 2012 FRM Draft TAR released for public comment 7/16 Comment period closed in 9/16 In 2012, Mazda introduced their SkyActiv G family of engines in the U.S. Notable characteristics: First implementation of Atkinson Cycle outside of HEVs/PHEVs (as far as we knew) Very high geometric compression ratio (13:1 U.S., 14:1 E.U. and Japan) EPA engineering staff thought it warranted a closer look and added SkyActiv G to the list of engines and transmissions that would be benchmarked as part of our powertrain technology assessment activities Benchmarking and model data used in MTE Technical Assessment Report (TAR) Subsequently other Atkinson engines have been released Toyota ESTEC 2GR FKS/FXS V6, SAE 2015 01 1972; 1NR FKE 1.3L I3 and 2NR FKE 1.5L I4 cegr/atkinson Hyundai Kappa 1.6L GDI, SAE 2016 01 0667; Nu 2.0L PFI 2
Benchmarking results on the 2014 U.S. Mazda 2.0L engine 3
Benchmarking Overview Benchmarked Mazda 2.0L (13:1 CR) engine Tier 2 E0 93 AKI Tier 3 E10 86 AKI Implemented into Hardware in Loop (HIL) test bed Validated to baseline vehicle test data Tested engine in a simulated future vehicle 4
Benchmarking Engines fully instrumented CAN data for available EPIDs All ECU I/O measured and logged Cylinder pressure on all cylinders Exhaust emissions Temperatures, pressures, etc. SAE Papers 2016 01 1007 Benchmarking and Hardware in the Loop Operation of a 2014 MAZDA SkyActiv 2.0L 13:1 Compression Ratio Engine 2016 01 0565 Air Flow Optimization and Calibration in High Compression Ratio Naturally Aspirated SI Engines with Cooled EGR 5
Mazda 2.0L Engine Benchmarking Atkinson Cycle Effects of LIVC cam phasing: Allows high geometric expansion ratio (13:1) Reduced effective compression ratio Varies from 5 11 due to intake cam phasing Decreases in cylinder temperatures and knock sensitivity Reduced pumping losses (at throttle) SAE Technical Paper 2016 01 1007 6
Benchmarking Intake Cam Phasing for Atkinson Cycle Intake retard from latest IVC Intake manifold pressure (kpa) Atkinson Internal EGR 7
Benchmarking Some improvement with Octane LEV III Fuel (E10, 88 AKI) Tier 2 Certification Fuel (E0, 93 AKI) -No change in torque curve from octane 8
Benchmarking 93 AKI 88 AKI comparisons BTE (93 AKI) BTE (88 AKI) (%) +3% BTE max Spark (93 AKI) Spark (88 AKI) (BTDC) +6 spark FTP HWFET -Efficiency differences mostly along the low speed torque curve -Caused by spark advance allowed by higher octane 9
Hardware in the loop (HIL) cycle testing on the 2014 U.S. Mazda 2.0L engine 10
Engine Hardware in Loop (HIL) Testing Vehicle Configuration VSIM (EPA s vehicle HIL model) is based on EPA s full vehicle simulation ALPHA model Allows test cell to drive an engine as a virtual vehicle Can infinitely vary: Drive cycle Vehicle test weight and road loads Transmission parameters Shift logic controlled by ALPHAshift Optimizes gearing for best efficiency Simple transmission thermal model used to calculate higher losses of cold transmission during FTP 11
HIL Testing Baseline Cycle Validation Validated VSIM with 2.0L SkyActiv to 2014 Mazda3 chassis test data Compared key characteristics to actual vehicle CAN bus data Engine speed, gear, fuel flow Cycle mph Chassis Engine Engine Speed Gear Total Fuel (g) SAE Technical Paper 2016 01 1007 12
HIL Testing Baseline Fuel Economy Baseline vehicle Bag 1 Bag 2 Bag 3 FTP HWFE 35.0 36.5 42.8 37.7 58.5 35.0 36.6 42.8 37.8 58.8 35.5 36.9 42.7 38.1 58.6 Average 35.2 36.7 42.8 37.9 58.6 Cert data 35.1 37.9 43.1 38.6 56.9 % error 0% 3% 1% 2% 3% Three repetitions completed for each tested configuration Baseline and future vehicles Standard test-test variability was very small for all cases Baseline HIL data correlated well with 2014 Mazda3 certification test data SAE Technical Paper 2016 01 1007 13
HIL Testing Future Vehicle Specification Unmodified 2014 Mazda 2.0L engine (same as baseline HIL case) Approximated with 2025 midsize car Assumed footprint of current Mazda6 Maintained baseline acceleration performance (power/weight) Added features to 2025 midsize car: Future 8-speed transmission Based on current 8-speed ZF transmission (8HP50) Includes expected reductions in spin and pump losses Active trans warmup (assume thermal loop) Stop-start (calculation adjustment only) Road load reductions (two levels, L1 and L2) SAE Technical Paper 2016 01 1007 14
HIL Testing Future Vehicle Road Load Used a reference road load based on average of several high volume 2008 midsize cars to properly reflect reductions in the Federal Rulemaking (FRM) Applied road load reductions in two levels (L1, L2) 2008 Mazda6 was almost identical to average 2008 vehicle SAE Technical Paper 2016 01 1007 15
HIL Testing Future Vehicle Road Load Starting with 2008 reference road loads, applied two levels of reductions: Level Weight Reduction Rolling Resist. Reduction CdA Reduction L1 10% 20% 20% L2 15% 30% 25% Resulting in the following test coefficients for 2025 midsize car L1 and L2: 2008 Mazda6 2025 Midsize Car L1 2025 Midsize Car L2 ETW 3625 3250 3125 A (lb) 29.7 24.3 22.5 B (lb/mph) 0.3810 0.0279 0.1622 C (lb/mph 2 ) 0.01811 0.01765 0.01456 CRR 0.0089 0.0071 0.0062 CdA (m 2 ) 0.76 0.60 0.57 SAE Technical Paper 2016 01 1007 16
HIL Testing Future Vehicle Fuel Economy Cycle test results of Skyactiv 2.0L engine as 2025 midsize car (mpg): 2025 midsize car L1 Bag 1 Bag 2 Bag 3 FTP HWFE 40.6 39.3 45.7 41.2 64.9 40.7 38.6 45.6 40.8 64.5 40.5 38.3 45.5 40.6 64.2 Average 40.6 38.7 45.6 40.9 64.5 2025 midsize car L2 Bag 1 Bag 2 Bag 3 FTP HWFE 41.5 39.8 46.7 41.9 67 41.6 40.3 47.0 42.3 67.2 41.6 40.0 46.8 42.1 67.1 Average 41.6 40.0 46.8 42.1 67.1 These are raw results, prior to adjustment for assumed stopstart operation SAE Technical Paper 2016 01 1007 17
HIL Testing Future Vehicle Fuel Economy with Idle Start Stop Made adjustments assuming the 2025 midsize car would be equipped with a stop start device Enable conditions: > 120s run time AND Coolant temp > 80C 2025 midsize car L1: start stop adjustments CBE corrected Total Idle Adj total FE FE adj g/mi adj Bag 1 247.3 4.3 242.9 40.6 41.3 215.0 Bag 2 278.3 18.4 259.9 38.7 41.5 214.3 Bag 3 230.4 9.4 220.9 45.6 47.5 186.9 FTP total 257.9 12.8 245.1 40.9 43.0 206.7 HWFE 64.5 137.7 Combined 50.6 175.6 2025 midsize car L2: start stop adjustments CBE corrected Total Idle Adj total FE FE adj g/mi adj Bag 1 241.7 3.9 237.8 41.6 42.2 210.4 Bag 2 269.0 17.6 251.4 40.0 42.8 207.5 Bag 3 214.3 9.0 205.4 46.8 48.9 181.8 FTP total 247.6 12.2 235.4 42.1 44.3 200.8 HWFE 67.1 132.4 Combined 52.3 170.0 SAE Technical Paper 2016 01 1007 18
HIL Testing Future Vehicle Results FTP and HWFET cycles (combined) for the 2025 midsize cars yielded 170 176 g/mi CO2 (L2 results shown below) Total Fuel (g) Idle Fuel (g) Adjusted Fuel (g) FE (mpg) g/mi CO2 Bag 1 241.7 3.9 237.8 42.2 210.4 Bag 2 269.0 17.6 251.4 42.8 207.5 Bag 3 214.3 9.0 205.4 48.9 181.8 FTP (total) 247.6 12.2 235.4 44.3 200.8 HWFE 67.1 132.4 Combined 52.3 170.0 The 2025 GHG compliance level for a midsize car with a 48 ft 2 footprint is 154 g/mi Possible A/C credits anticipated to be up to 18.8 g/mi This suggests a target range of 154 173 g/mi The HIL test results suggest this hypothetical vehicle has the potential to obtain compliance levels with the existing 2.0L Skyactiv engine SAE Technical Paper 2016 01 1007 19
GT-POWER Atkinson engine futuring 14:1 CR, cooled EGR (cegr), cylinder deactivation (CDA) 20
GT POWER Validation 13:1 Engine Benchmarking Data Fuel: 42.9 MJ/kg, 96 RON Tier 2 certification gasoline (E0) Dynamometer test data over more than 200 speed and load points GT Power Maps generated from 0.5 bar to 13 bar BMEP Modeled BSFC was significantly higher below ~0.5 bar BMEP load and GT Power sometimes estimated unreasonably high BSFC at 0 bar BMEP. BSFC at below 0.5 bar BMEP was therefore estimated by using a low fidelity extrapolation method. SAE Technical Paper 2016 01 0565 21
GT POWER Modeling 14:1 and cegr Incremental FC effectiveness of cegr alone: ~ 2 5% Incremental FC effectiveness of cegr + 14:1 CR: ~ 4 5% SAE Technical Paper 2016 01 0565 22
GT POWER Modeling 14:1, cegr, and CDA 2 cylinder deac/4 2 cylinder deac/6 BSFC Combined cegr and CDA -BSFC reduction from reduced pumping losses at partial load SAE Technical Paper 2016 01 0565 23
Future Work Proof of concept engine development based on 14:1CR 2.0L EU version of the engine cegr 2017 SAE Congress paper CDA Combustion improvements Make further improvements to GT Power Model Further model validation as data becomes available Validate EGR and kinetic knock models Burn duration Model a larger DOE space Sweep EGR rates, spark timing & camshaft phasing within model Explore use of MathWorks model based Calibration Toolbox for rapid development of engine control and calibration Further investigate conditions and limitations for implementation of cylinder deactivation 24