Powertrain Innovations for Future Vehicles an ARPA-E ... · ARPA-E Advanced Research Projects...

Powertrain Innovations for Future

Vehicles – an ARPA-E Perspective

Chris Atkinson, Sc.D.

Program Director

Advanced Research Projects Agency-Energy

US Department of Energy

June 3, 2015

ERC – 2015 Symposium

ARPA-E Advanced Research Projects Agency-Energy

Recommendations

Base on DARPA to be lean and

agile organization

Support creative “out-of-the-box”

transformational energy research

History

2006 Conceived by National Academies

ARPA-E Mission

Catalyze the development of transformational,

high-impact energy technologies

Ensure the U.S. maintains a lead in the development

and deployment of advanced energy technologies

Electrofuels BEEST

PETRO MOVE AMPED

REACT

SBIR/STTR

HEATS BEETIT GRIDS

ADEPT GENI

Solar

ADEPT

TransportationStationary

IMPACCT

Focused Programs 2010-2015: $1,100M, 25 programs

Technology Acceleration Model

time

cost

performance

What does ARPA-E look for?

Steam-powered Cugnot

(1769)

Benz Motorwagen

(1885)

Ford Model T

(1914)

Technologies that lead to new adoption pathways

What Makes an ARPA-E Project?

BRIDGE‣ Translates science into breakthrough technology

‣ Not researched or funded elsewhere

‣ Catalyzes new interest and investment

IMPACT‣ High impact on ARPA-E mission areas

‣ Credible path to market

‣ Large commercial application

TRANSFORM‣ Challenges what is possible

‣ Disrupts existing learning curves

‣ Leaps beyond today’s technologies

TEAM‣ Comprised of best-in-class people

‣ Cross-disciplinary skill sets

‣ Translation oriented

9

A personal note……

Model-Based Control,

Calibration & Optimization of

High Efficiency EnginesChris Atkinson, Sc.D.ATKINSON LLC

SAE 2014 High Efficiency IC Engine Symposium

TRANSIENT CYCLE REQUIREMENTS

BASELINE CALIBRATION

TRANSIENT CONTROL INPUTS

TARGET WEIGHTING & OPTIMIZATION

EMISSIONS, PERFORMANCE AND FUEL CONSUMPTION

TARGETS

SYSTEM AND OPERATING

CONSTRAINTS

INVERSE DYNAMIC DATA-DRIVEN CONTROL MODEL

FORWARD PREDICTIVE DATA-DRIVEN ENGINE

MODEL

ENGINE

MODEL-BASED CONTROL SYSTEM

ENGINE AND SYSTEM

DEVELOPED OFF-LINE

REAL-TIME, ON-LINE

IMPLEMENTATION

ATKINSON LLC 11

12

US Department of Energy SuperTruck Program 2010-2015

Daimler Trucks NA/Freightliner/Detroit Diesel

13

For the Freightliner SuperTruck,

Detroit Diesel and Atkinson LLC developed and implemented

• a fully model-based engine control system,

• using data-driven methods,

• that includes real-time, on-board optimization of fuel

efficiency,

• subject to emissions and performance constraints.

• Results include 12.2 MPG with +115% improvement

in freight efficiency

• With 50.2% peak brake thermal efficiency.

ARPA-E’s energy mandate

‣ Reduce energy imports.

‣ Improve the efficiency of energy generation, storage,

transmission, distribution and usage.

‣ Reduce energy-related emissions including CO2.

‣ Promote US innovation and hence competitiveness in the

energy arena.

– A direct match for automotive engine and powertrain

efficiency.

Total Oil Use

Imported Oil

Total Gasoline Use

Domestic Production

Total Middle Distillate Use

Zero imports with 40% increase in

LD and HD powertrain efficiency

3 overarching energy trends in the automotive

area

– Reduce imported oil and GHGs;

• 90+% of future vehicles will be either conventional or

hybrid vehicles (DOE, EIA, EPA, NRC);

– Vehicle connectivity (V2V, V2I, V2X) advances;

– Vehicle autonomy (of varying degrees).

The Automotive Industry

‣ Is a very mature, conservative industry dominated by

– Regulation (safety),

– Regulation (emissions),

– Customer preferences,

– While meeting strict cost and price constraints.

And hence has been considered ripe for disruption – cf.

Tesla, Apple, Google…

‣ The “precautionary principle” has been replaced by the Wild

West – “permission-less innovation”.

The motivation for new ARPA-E program

area

‣ Target automotive engine and powertrain technologies

beyond those proposed for 2025.

‣ Look for substantial fuel consumption and CO2 emissions

reductions beyond those currently expected.

‣ Take advantage of proposed peripheral or tangential

technologies – connectivity and autonomous operation.

– ARPA-E looks for transformational technologies and

does not fund incremental improvements.

Opportunities for maximizing vehicle efficiency

via connectivity and autonomous operation

‣Powertrain

architecture and

hardware

‣Powertrain control

technologies

0

50

100

150

200

250

300

350

400

0

10

20

30

40

50

1970 1980 1990 2000 2010 2020 2030

FuelEconomy[M

PG]

Year

Technology and fuel efficiency trends

Fuel Economy

Horsepower

Computational

power

ECU introduction

0

50

100

150

200

250

300

350

400

0

10

20

30

40

50

1970 1980 1990 2000 2010 2020 2030

FuelEconomy[M

PG]

Year

Opportunity to use connectivity &

autonomy to increase individual vehicle

fuel economy?

?

Fuel Economy

Horsepower

Computational

power

ECU introduction Connectivity

State of the Art Peak Engine Efficiency

‣ SI NA ~38.5% (Toyota Prius)

‣ CI LD ~42% (VW TDI)

‣ CI HD ~47% (Cummins and Daimler – without WHR)

Why aren’t we there yet?

CITY

HIGHWAY

Advancements in peak engine efficiency

20

30

40

50

60

70

80

90

100

50 100 150 200 250 300 350 400 450

FUELECONOMY(m

pge)(FROMCONVEN

TIONAL

VEH

ICLEDATA

)

WHEELENERGY(Wh/mile)

Advancements in peak engine efficiency

20

30

40

50

60

70

80

90

100

50 100 150 200 250 300 350 400 450

FUELECONOMY(m

pge)(FROMCONVEN

TIONAL

VEH

ICLEDATA

)

WHEELENERGY(Wh/mile)

20%

25%

30%

35%

40%

45%

50% 100%

VW Golf Diesel ~42% BTE

Real life ~24% η

Toyota Prius~38% BTE

Real life ~34% η

Advancements in peak engine efficiency

10

20

30

40

50

60

70

80

90

100

25 75 125 175 225 275 325 375 425

FUELECONOMY(m

pge)(FROMCONVEN

TIONAL

VEH

ICLEDATA

)

WHEELENERGY(Wh/mile)

2012 Chrysler 300

3.6L V-6

2013 VW Jetta

TDI 2.0L I-4

2010 Toyota Prius

~38% BTE

~42% BTE

Advancements in peak engine efficiency

10

20

30

40

50

60

70

80

90

100

25 75 125 175 225 275 325 375 425

FUELECONOMY(m

pge)(FROMCONVEN

TIONAL

VEH

ICLEDATA

)

WHEELENERGY(Wh/mile)

20%25%

30%35% 40%

45%50%

100%

2012 Chrysler 300

3.6L V-6

2013 VW Jetta

TDI 2.0L I-4

2010 Toyota Prius

~38% BTE

~42% BTE

Topic Areas of Interest

‣A – Reducing fuel consumption by improving

powertrain control through connectivity.

‣B – Improving future autonomous vehicle fuel

efficiency via new engine and powertrain

technologies.

Reducing Fuel Consumption through

Connectivity

Advanced Information and Communication Technologies (ICT)

Vehicle Connectivity + Autonomous Driving = EVs?

Vehicle Connectivity + Autonomous Driving = EVs?

Not if 90+% of vehicles will continue to be

conventionally powered or hybrids, even after 2030.

(cf. DOE, NRC, EPA, EIA)

Vehicle Connectivity is a secular trend.

‣ Vehicles will become increasingly connected, to each other

(V2V), to the infrastructure (V2I), and to the “cloud” (V2X).

‣ Vehicles will display increasing degrees of autonomy in their

operation –

– Advanced driver assistance systems

– Semi-autonomous

– Fully autonomous

Connected vehicles – V2V, V2I, V2X

DENSO, 2015

OTA

Fuel properties

Look ahead data

OTA

Fuel properties

Predictive control

Platooning

ITS

Eco-routing

GPS

Collision avoidance

Eco-Routing

V2V

V2I

V2X

What are the enabling opportunities for

increasing individual vehicle fuel efficiency?

Actuators &

sensors

Self-

awareness

Driver input

Autonomous

calibration

New powertrain systems

Self driving

Self parking

Steering

Cruise

Brakes

L1, L2

L3, L4

System level Individual vehicle

Future fuel efficiency potential with V2X?

On an individual vehicle basis

‣ Real-time optimization of powertrain control – on-board or

off-board.

– What can be done with 1000x real-time computational

capability?

– What can be done with more knowledge and

information?

– Supercomputer in the cloud?

– OTA updates of engine control and calibration?

– Bespoke, fully customized individual vehicle calibration

optimization?

What are the fuel efficiency margins due

to…

‣ Variations in fuel composition.

‣ Engineering and emissions margins.

‣ Variation in actuators and sensors, and ageing.

‣ Environmental conditions.

‣ Off-cycle operation.

‣ Driver behavior.

Future potential with Vehicle Autonomy?

‣ Advanced driver assistance systems vs. semi-autonomous

vs. fully autonomous vehicles.

‣ Autonomous EVs offer little potential for efficiency

improvement over non-autonomous EVs – aerodynamics,

weight and cabin loads dominate.

‣ Autonomous engine-powered vehicles offer significant

potential for efficiency improvements.

– Reduction in driver’s expectation of performance.

– Removing the driver from the loop.

Future potential with Vehicle Autonomy?

Autonomous Vehicle Technology – A Guide for Policymakers – Anderson et al., RAND Corporation, 2014

Improving Future Vehicle Fuel Efficiency

via New Engine and Powertrain

Technologies

Possible technology solutions

‣ Progression of existing technologies –

– downsizing and boosting, (or upsizing and NA)

– ultra-lean operation facilitated by new boost, fueling and ignition systems,

– dilution with air or EGR

– variable (high) compression ratio engines,

– Otto and Diesel cycle convergence,

– very high compression engines with knock mitigation,

– LTC, dual fuel or reactivity controlled combustion,

– friction, pumping loss and auxiliary load reduction,

– thermal loss reductions (reduced surface area available for heat transfer),

– stop/start,

– waste heat recovery, hybridization and direct energy recovery systems.

‣ New engine architectures – such as

– free piston engines,

– split-cycle engines,

– entirely new thermodynamic and combustion cycles.

‣ New engines and powertrain operating systems –

– variable displacement engines, or

– intermittent operation engines,

– Integration with electric machines or other hybrid systems.

Any new powertrain technology has to be

comparable to or better than the baseline in:

Criterion Explanation

Power Power density (or energy density including the fuel/energy

storage capacity) Customer acceptance

Efficiency Fuel economy (over real-world dynamic driving)

Regulation

Emissions Regulated criteria pollutants (and now CO2) Regulation

Cost Total cost of ownership (including capex and fuel cost)

Reliability Mean time between failures, maintainability

Utility Acceleration, driveability, NVH, cold start, off-cycle operation,

ease of use, transparency to the user, and acceptable range

Fuel acceptability Use a readily available fuel or energy source.

An Automotive Industry Comparison

Mid-size

Conventional Mid-size Hybrid Tesla Model S

Tesla:Mid-

size

Power (hp) 240 200+ 500 2.0X

Emissions

(Tailpipe) ULEV ULEV ZEV 0.0X

Cost ($) $32,000 $35,000 $95,000 3.0X

Range (mi) 540 540 265 0.5X

Refueling Rate

(MW) 20 20/0.01 0.02 0.001X

Sales Volume

(per yr) 2,500,000 250,000 20,000 0.01X

IAV 2+2 Engine (2014)

Displacement on demand:

• Electro-mechanical

decoupling of 2 cylinders

to reduce frictional losses

and pumping loss, and to

match engine output to

demand.

• Increase BMEP of each

cylinder

Claim: +14% fuel economy

over NEDC drive cycle

Toyota R&D (2014)

10 kW free piston

engine generator

42% BTE claim

Advantages: very high

CR, low friction

Achates Power opposed piston 2 stroke compression ignition engine

(2012)

Advantages: reduced heat rejection, high CR, Diesel cycle

EcoMotors 2 stroke opposed piston, opposed cylinder

compression ignition engine (2010)

Advantages: reduced surface area for heat loss, combustion

phasing, reduced pumping losses (unthrottled 2 stroke cycle,

lean)

Scuderi Split-Cycle engine (2010)

Advantages: over-expansion, reduced thermal losses,

regeneration, combustion phasing

LiquidPiston (2012)

40 hp multifuel

capable rotary

design. Unique

thermodynamic cycle.

Efficiency claims:

58% peak BTE

50% part load

Other new (or old) thermodynamic cycles?

51

Entirely New Engine Concepts?

‣ Develop an entirely new class of modular “cylinder on

demand” engines with integrated hybrid architectures, that

allow the engine to operate at peak efficiencies even under

part load conditions, suitable for conventional and

autonomous operation (at various levels).

A Notional Concept for a Highly Efficient New

Class of Engines

53

TRANSMISSIONCYL1 CYL3CYL2

EM1

C1EM2

C2

EM3

C3

Conclusions

‣ The automotive industry is considered ripe for disruption –

perhaps it is time for engine design to be disrupted.

‣ Peak engine efficiency should be de-emphasized for highly

transient operation, and part-load efficiency improved.

‣ Connectivity can offer engine control, calibration and

optimization advantages.

‣ Autonomous operation takes the driver out of the loop, and

full autonomy removes the driver’s expectations for

performance – and can enable new high efficiency engine

and powertrain options.

Chris Atkinson, Sc.D.

chris.atkinson@hq.doe.gov

Program Director

Advanced Research Projects Agency-Energy

US Department of Energy