Energy Dependence and COEnergy Dependence and CO22: : Nearest-Term OpportunitiesNearest-Term Opportunities

for a Dramatic Reductionfor a Dramatic Reduction

Drs. Paul J. Werbos and James A. MomohNSF, IEEE, ACUNU/MP, Howard U.

-- presenting personal, not official, viewsPJW: 80’s: EIA/DOE lead analyst for

long-term energy futures

pwerbos@nsf.gov

National Science Foundation

EngineeringDirectorate

Computer & Info.Science Directorate

ECS IIS

Control,Networks andComputational

Intelligence

EPDT:Chips,Optics,

Etc.

Robotics AI

Information Technology Research (ITR)

MegaChallenges for the 21MegaChallenges for the 21stst Century & ECS RoleCentury & ECS Role

Key Challenges To Basic Scientific Understanding:– How to build/understand real Intelligent Systems?

(computational intelligence)– How does the Universe work? (Quantum...)– What is Life? (QSB quantitative biotechnology)

Key Broader Challenges to Humanity:– Sustainable growth on earth– Cost-effective sustainable space settlement– Human potential

Some budget numbers for Some budget numbers for Fiscal Years 2003, 2004Fiscal Years 2003, 2004

Total Engineering Directorate:$540m, $565mTotal Electrical and Communication Systems

(ECS) : $73m, $75mAvailable for new awards in ECS core

programs: $36m, $35mSuccess rates for regular unsolicited

proposals in ECS core programs: 24%, 11%

www.stateofthefuture.orgCD-ROMEnergyPaper

When Oil and Gas Get Too Costly, When Oil and Gas Get Too Costly, Economically or Politically…Economically or Politically…

How do we keep our cars running?Where does the new fuel or electricity

come from? Sources? Distribution?– Rapid growth in imports of LNG

Serious hope of avoiding a crisis of dependency in time but no guarantee

HOW MUCH TOHOW MUCH TO REDUCE COREDUCE CO2,2,

WHEN?WHEN?

2004 2030 2100

Kyoto Approach:Too little

The Middle Way:6-fold goal 30 years

The Purist Way: True H2 , Too late

Carbon Dioxide Reduction

TECHNICAL PILLARS OF TECHNICAL PILLARS OF THE MIDDLE WAYTHE MIDDLE WAY

INDEPENDENTSUSTAINABLE

CAR FUELS

TRULY INTELLIGENTPOWER GRID

INDEPENDENTSUSTAINABLEELECTRICITY

•Flexible fuel (e.g. Stirling)•Plug-in hybrid – step to electric• New type fuel cell cars•Methanol production

• Autonomous• Adaptive• Cyber Control• MEMS, GPS

• Cleanest coal IGCC/sequest.• “Solar farms” • Space solar• Also Team B?

The chicken and egg problem:The chicken and egg problem:which comes first? Hwhich comes first? H22 fuel , H fuel , H22 car? car?

•Would you buy a car that only runs on H2 before your local gas station carries H2? Are PR stations enough?•Would you invest $trillion in gas stations and pipelines before people have H2 cars?• Technically: nonconvexity problem; also high costs

Long-Term Clean Alternatives to Long-Term Clean Alternatives to Carrying HCarrying H22 in Your Car Tank in Your Car Tank

Hydrogen Carriers – proven tested fuels that easily release hydrogen for use on-board a car– Methanol, our best hope (next slide)– Ammonia & other carbon-free fuels (but chicken&egg

problem again)

Electric Cars – Cleanest, most efficient, but needs R&D; can’t yet beat C; new batteries in lab exciting, but not yet… PLUG-IN HYBRIDS COULD GET US THERE.

Thermal Batteries (Long-term option if Stirling grows)

What IS Methanol?What IS Methanol?

C H

H

H

H C H

H

H

O

HC H

H

H

O

H

C HH

MethaneNatural GasScarce as OilNeeds Special Tank Ethanol

e.g From CornDrinkable

MethanolGood H CarrierCan Be BioliquidOr From Coal, Gas

Best Hope for Fuel Cell CarsBest Hope for Fuel Cell Cars Carry methanol in the gas tank Small “steam reformers” (proven known technology) to

convert methanol to hydrogen in car Carbon-tolerant alkaline fuel cell

– Unlike PEM fuel cell, no peroxide loss of energy and erosion; can be made cheaper

– Realistic hope of twice the miles/Btu of hybrids– Recent proof of air-electrode carbon tolerance (ECS$)– Plan for fuel side; can use carbon black electrodes and “Jiffy

Lube” electrolyte refreshing Breakthru battery R&D needed for all hopes!

GEM Flexibly Fuel Vehicles (FFV)GEM Flexibly Fuel Vehicles (FFV)One Tank To Hold Them AllOne Tank To Hold Them All

G: Gasoline

E: Ethanol

M: Methanol

With an FFV, you choose each day which to buyAt $100-200/car, a more open competition, level playing field, better unleash the power of the free market40% of new cars in Brazil GE flexible already

GEM Flexibility Is Well-GEM Flexibility Is Well-EstablishedEstablished

ALCOHOL FUELS

"Detroit is ready now to -- make cars that would run on any combination of gasoline and alcohol -- either ethanol, made from corn or methanol, made from natural gas or coal or even wood. Cars produce less pollution on alcohol fuels, and they perform better, too. Let us turn away from our dependence on imported oil to domestic products -- corn, natural gas, and coal -- and look for energy not just from the Middle East but from the Middle West."

Source: George Bush 1988 Campaign Brochures www.4president.org

AREAS FOR NEEDED LEGAL REFORMS

Ethanol+Methanol Fuel flexibility should be MANDATORY in new gasoline-using cars from 2006/7/8. Hybrids or advanced Stirling can also use GEM fuel tanks easily.

Incentives and research opportunities for bio-methanol should be the same as for bioethanol, biohydrogen or better

Zoning rules discouraging Distributed Generation should be modified to simplify renewable or alcohol fuel use

Grid regulation needs to be made to fit "intelligence“Leak proof tanks in gas stations for ALL fuels

Short-Term Benefit of FFVsShort-Term Benefit of FFVsto US Oil/Gas Industryto US Oil/Gas Industry

Outside US, cost of new methanol is $98/ton from remote gas (Google “Canaccord methanol”), much of which is now wasted (vented and flared).

For oil companies, new remote methanol plant will be equivalent to new proved reserves! Free market will supply methanol if FFVS.

Don’t burn CH4 to electricity! There are better new electricity sources!

Near-Term R&D: CH4-to-Methanol: small $ to follow up Catalytica, could cut cost in half

MAIN Sustainable Paths as Oil/Gas MAIN Sustainable Paths as Oil/Gas Run Out (i.e. Cost More)Run Out (i.e. Cost More)

Nuclear

Space Solar

Earth Solar

Coal

Electricity

Methanol(CH4OH)

MotorVehicles

Industry

Buildings

?

Methane

Sources: Where Does the Electricity or Sources: Where Does the Electricity or Methanol Come From If Not Oil/Gas?Methanol Come From If Not Oil/Gas?

Two scenarios: Base-Case-Present-Trends Versus Real-Hope-If-We-Act-More

Base Case:– Iran, China, eventually everyone builds fission as fast as

they can. Bin Laden Construction Co. and its less savory competitors grow very rich, very fast. 4-8¢/kwh

– Little guys (wind, rooftop solar, Anwar, ethanol) make big $ but don’t plug half the supply-demand gap

– Supply-demand gap still widens. Old coal fills the gap, filling half the world with barely survivable air (worse than China’s cities today). Not so much methanol.

– “Santa Claus drowns” Arctic Ice Cap Double or Nothing

Real Hope If We Work/Think HardReal Hope If We Work/Think Hard THREE TEAM A TECHNOLOGIES

– We know that all three CAN WORK and CAN provide all the world’s energy needs cleanly

– “IGCC” (Cool Water/Texaco/Eastmann/GE) Clean Coal Technology, Good for carbon sequestration, efficiency, wants to produce electricity and methanol together

– “solar farms” on earth with mirror or lenses: but breakthru needed on cost, new workshops?

– Space solar power – new designs from NASA-NSF-EPRI Need better (agile, international?) funding vehicle for high

risk breakthrough TEAM B hopes, like advanced large-scale biomethanol biotechnology.

Some Issues Re Earth SolarSome Issues Re Earth Solar DOE 10-year targets: 14¢/kwh PV, intermittent power, can’t compete with

coal 4¢ baseload. Cost of “balance of system” is stubborn with solar farms. Even worse for low efficiency (now 3%) “nano-based” PVs this decade.

Sandia has projected 6¢/kwh using mirrors and advanced Stirling to convert heat to electricity. But we don’t yet have the mirrors and improvements seem slow.

Cost breakthrough in “mirrors” may be possible, e.g. as spinoff of NASA work on flexible self-assembling lenses to focus light for space use. But options have yet to be scoped out and explored systematically. Workshops?

All DG (solar, biomass, wind…): Hayes says hookups are the real barrier. Solution: intelligent grid and zoning changes and improved advanced Stirling (industry)

NSF-NASA Workshop on Learning/RoboticsFor Cheaper (Competitive) Solar Power

See NSF 02-098 at www.nsf.gov &URLs

Some OutcomesSome Outcomes

98 proposals, $21 million recommended after tough merit review, $3 million funded

Previous NASA SERT program: first well-validated designs but 17¢/kwh even assuming $200/lb earth-to-LEO (Low Earth Orbit)

Now 4 designs may achieve cost breakthrus, merit follow-up. One – hybrid light-to-light laser with D-D inertial fusion and microwave beaming might get <1¢/kwh for kwh to receiver (rectenna).

Little of Texas A&M claims he can demo ability to avoid communications interference.

“Near-term vehicle” design 1st wi real hope <$200/lb

Key Needs for Space SolarKey Needs for Space SolarPartnership with NASANew Big Laser (2/4 cheap ways)Affordable launch (follow-on to ECS-

funded plasma hypersonics)Improved Robotics – REQUIRES MORE

USE OF COMPUTATIONAL INTELLIGENCE!!! (NSF/DARPA??)

Cheaper rectennas – PES/MTT partnership

Human mentors robot and then Human mentors robot and then robot improves skillrobot improves skill

Learning allowed robot to quickly learn to imitatehuman, and then improveagile movements (tennisstrokes). Learning many agile movements quickly will be crucial to enabling >80% robotic assembly in space.

Schaal, AtkesonNSF ITR project

Distribution and Grids: Upgrading the Middle Distribution and Grids: Upgrading the Middle is As Important As Cars And Sources!is As Important As Cars And Sources!

Nuclear

Space Solar

Earth Solar

Coal

Electricity

Methanol(CH4OH)

MotorVehicles

Industry

Buildings

Successful Transition Will Require An Intelligent Electric Power Grid

?

Dynamic Stochastic Optimal Power Flow Dynamic Stochastic Optimal Power Flow (DSOPF): How to Integrate the “Nervous (DSOPF): How to Integrate the “Nervous

System” of ElectricitySystem” of Electricity DSOPF02 started from EPRI

question: can we optimally manage&plan the whole grid as one system, with foresight, etc.?

Closest past precedent: Momoh’s OPF integrates &optimizes many grid functions – but deterministic and without foresight. UPGRADE!

ADP math required to add foresight and stochastics, critical to more complete integration.

ANN to I/O From Idealized Power GridANN to I/O From Idealized Power Grid

4 General Object Types (4 General Object Types (busbarbusbar, wire, G, L), wire, G, L) Net should allow Net should allow arbitrary numberarbitrary number of the 4 objectsof the 4 objects How design ANN to input and output FIELDS How design ANN to input and output FIELDS ---- variables like the SET of variables like the SET of

values for current ACROSS all objects? values for current ACROSS all objects?

Beyond Bellman: Learning & Beyond Bellman: Learning & Approximation for Optimal Management Approximation for Optimal Management

of Larger Complex Systemsof Larger Complex Systems Basic thrust is scientific. Bellman gives exact optima for

1 or 2 continuous state vars. New work allows 50-100 (thousands sometimes). Goal is to scale up in space and time -- the math we need to know to know how brains do it. And unify the recent progress.

Low lying fruit -- missile interception, vehicle/engine control, strategic games

New book from ADP02 workshop in Mexico www.eas.asu.edu/~nsfadp (IEEE Press, 2004, Si et al eds)

Emerging Ways to Get Closer to Emerging Ways to Get Closer to Brain-Like SystemsBrain-Like Systems

IEEE Computational Intelligence (CI) Society, new to 2004, about 2000 people in meetings.

Central goal: “end-to-end learning” from sensors to actuators to maximize performance of plant over future, with general-purpose learning ability.

This is DARPA’s “new cogno” in the new nano-info-bio-cogno convergence

This is end-to-end cyberinfrastructure– See hot link at bottom of www.eng.nsf.gov/ecs

What’s new is a path to make it real

Intelligent Grid Requires Intelligent Grid Requires Intelligence But Also HardwareIntelligence But Also Hardware

Brain-like intelligence is embodied intelligence; sensors, actuators and feedback on performance are essential parts of the new designs.

Reduce world CH4 to kwh: sell and upgrade Brazil’s superior transmission technology (Pilotto, Watanabe: could save California billions quickly, allow cheap electricity from underused Utah coal plants)

EPRI plan to add more communications, sensors, intelligent appliances (e.g. car chargers to turn on at quiet times at night, to exploit times of strong wind)

Interface of intelligent grid with human users, markets and regulations. (www.pserc.cornell.edu; “EPNES” at www.nsf.gov.)