Big Data for Energy and Environment

SPR ING 2014 CONFERENCE

MaCkENzIE ROOMJEN-HSuN HuaNG ENGINEERING CENtERStaNFORD uNIVERSItY

MaY 15, 2014

Stanford university Energy and Environment affiliates Program

S T A N F O R D U N I V E R S I T Y

S P R I N G 2 0 1 4 C O N F E R E N C E

2

WELCOME

The advent of massive data sets and advanced analytical techniques is changing information technology in fundamental ways. In the past, data was used primarily to confirm hypotheses. Increasingly, data is used to generate

hypotheses, indicate trends, identify anomalies, and make predictions. This new paradigm is impacting research in diverse fields and business in diverse industries.

The impact is particularly great in energy and environment. This is driven by the proliferation of sensors measuring physical properties, the availability of vast networks of human data, the algorithms to extract information, and the computing power to do this in real time. This is an area where data, tools, and applications are closely linked.

This conference on big data in energy and environment showcases the latest work in four broad areas. The session on data science techniques and embedded sensors looks at the fundamental analytical tools and sources of data that enable applications. The session on oil and gas exploration and production shows how more accurate information and models are revolutionizing drilling and extraction.

The session on environment and climate change examines the role of big data in valuing natural resources and adaptation to climate change. The session on societal networks and the electric grid addresses the interdependency of social science and efficient energy management. Each of these sessions includes an innovative and exciting startup company, presented by the entrepreneurial founders.

Thanks to your support, the Energy and Environment affiliates Program is one of Stanford’s largest and fastest growing corporate affiliates programs. It spans all of Stanford University with a focus on energy, environment, materials, chemistry, and sustainability. These semi-annual meetings provide a remarkable opportunity to see cutting-edge research, meet the professors and graduate students who are doing the research, and interact with other companies.

Thank you for your participation.

aGENDatHuRSDaY, MaY 15, 2014

8:00 am to 9:00 am Registration and breakfast

9:00 am to 9:05 amWelcome Steve Eglash, Executive Director, Energy and Environment Affiliates Program

9:05 am to 9:15 amIntroduction to Big Data for Energy and EnvironmentJeff Koseff, Co-Director, Stanford Woods Institute for the Environment, Professor, Civil and Environmental Engineering

General Data Science Techniques and Sensors

Moderator: Sally Benson, Director, Global Climate and Energy Project, Director, Precourt Institute for Energy, Professor, Energy Resources Engineering

9:15 am to 9:45 amData Science and AnalyticsMargot Gerritsen, Director, Institute for Computational and Mathematical Engineering, Associate Professor, Energy Resources Engineering

9:45 am to 10:15 amSoftware for Embedded Sensing SystemsPhil Levis, Associate Professor, Computer Science and Electrical Engineering

Oil and Gas Exploration and Production

10:15 am to 10:45 amAnalyzing Seismic Data from Large Permanent Arrays to Monitor Subsurface ProcessesBiondo Biondi, Professor, Geophysics

10:45 am to 11:00 am Break

11:00 am to 11:30 amMultiple-Point Geostatistics: Stochastic Modeling with Training ImagesJef Caers, Professor, Energy Resources Engineering

11:30 am to 11:40 amOverview of Stanford Data Science Initiative (SDSI)Hector Garcia-Molina, Director, SDSI, Professor, Electrical Engineering and Computer Science

11:40 am to 12:10 pm Student and Postdoctoral Poster Preview Presentations

12:10 pm to 1:30 pm Lunch and Poster Viewing

1:30 pm to 1:35 pmEnergy and Environment Affiliates Program OverviewSteve Eglash, Executive Director, Energy and Environment Affiliates Program

Environment and Climate Change

Moderator: Jeff Koseff, Co-Director, Stanford Woods Institute for the Environment, Professor, Civil and Environmental Engineering

1:35 pm to 2:00 pm Earth Genome ProjectMary Ruckelshaus, Managing Director, Natural Capital Project

2:00 pm to 2:25 pmIncorporating Big Data into Climate Risk ManagementNoah Diffenbaugh, Associate Professor, Environmental Earth Systems Science

2:25 pm to 2:45 pmBig Data for AgricultureAnnie Hazlehurst, Founder, Faridan and Sara Menker, CEO and Founder, Gro Intelligence

2:45 pm to 3:05 pm Environment and Climate Change Panel Discussion

3:05 pm to 3:20 pm Break

Electric Grid, Energy Efficiency, and Social Networks

Moderator: Barton “Buzz” Thompson Jr., Co-Director, Stanford Woods Institute for the Environment, Professor, Natural Resources Law

3:20 pm to 3:45 pmBig Data and Networked Systems: Insights for Transportation SystemsBalaji Prabhakar, Professor, Electrical Engineering and Computer Science

3:45 pm to 4:10 pmElectric Grid AnalyticsRam Rajagopal, Assistant Professor, Civil and Environmental Engineering

4:10 pm to 4:30 pmThe Business Case for Big Data in SustainabilityAmit Narayan, CEO and Founder, AutoGrid

4:30 pm to 4:50 pm Electric Grid, Energy Efficiency, and Social Networks Panel Discussion

4:50 pm to 5:00 pm Wrap up

5:00 pm to 6:00 pm Wine and cheese reception

S P R I N G 2 0 1 4 C O N F E R E N C E

3

44

Energy and environmental research is a major theme at Stanford and is found in nearly every academic discipline.

Stanford emphasizes an interdisciplinary approach, recognizing that energy and environmental research activities are interwoven within the fabric of nearly all Stanford academics.

FINDING RESEaRCH GERMaNE tO YOuR INtEREStSIt can be difficult for those outside the Stanford community to quickly grasp the

scope of energy and environment research activities on campus or to understand where specific research is being conducted in a particular area of commercial interest. Among the benefits that the Energy and Environment Affiliates Program aims to deliver to our members is assistance in understanding what specific research is taking place at Stanford, who is conducting it, and how to engage directly with those activities.

For details, visit the Affiliates Program website at eeap.stanford.edu.

MEMBER BENEFItSAffiliates Program members are entitled to the full range of benefits, including:• SupportforPh.D.studentresearchthroughparticipationintheFellow-Mentor-

Advisor Program;• In-depthinteractionswithfacultyandgraduatestudents;• Companyvisitsbyfacultyandgraduatestudents;• Facilitatedgraduatestudentrecruitingopportunities;• Accesstoresearchpapersandcomputermodels;• UseofadirectoryofStanfordresearchactivitiesinenergyandtheenvironment;• OpportunitytoestablishavisitingscientistatStanford;and• InvitationstoAffiliatesProgramsymposiaandsemi-annualconferences.

FELLOW-MENtOR-aDVISOR PROGRaMAffiliatesProgrammembershipincludesparticipationintheFellow-Mentor-Advisor(FMA)Program.ThegoaloftheFMAProgramistoestablisharelationshipbetweenaPh.D.student(“Fellow”),anemployeeofthecompany(“Mentor”),andaprofessor(“Advisor”).Aportionoftheannualmembershipfeeisgiventoaprofessorofthecompany’schoosingtohelpsupportaPh.D.Fellow.TheAffiliatesProgramcanassistthecompanyinmakingthischoice.TheFMAProgramisaboutcreating a relationship. It is not sponsored research, but it is an opportunity for the company to get close to the professor, the student, and the student’s research.

Membership in the Energy and Environment Affiliates Program is US$150,000 per year. Affiliates Program revenue is used to support Ph.D. student research, symposia and workshops, seed projects, equipment purchases, faculty and graduate student travel, and program operations. One FMA Program allocation is included in the annual membership fee. Companies can obtain additional FMA Program allocations for US$75,000 each.

COnTACT InFOrMATIOnSteve EglashExecutive Director, Energy and Environment Affiliates ProgramEmail: seglash@stanford.eduPhone: 650-721-1637Mobile: 650-799-2267

aBOut StEVE EGLaSHSteve Eglash is responsible for developing and managing

interactions for corporations and other organizations that have an interest in Stanford’s research, faculty, and graduate students in

energy-related and environmental fields. His background is in renewable energy, business, technology, and finance. Eglash was president and CEO of Cyrium Technologies, a solar energy startup company, and a consultant and advisor to the national renewable Energy Laboratory and the U.S. Department of Energy. He received his Ph.D. and M.S. degrees from Stanford University, and a B.S. degree from U.C. Berkeley, all in electrical engineering.

JOIN TODAYCONNECT WITH STANFORD

UNIVERSITY AND OTHER

COPORATIONS AND

ORGANIZATIONS

a F F I L I a t E S M E M B E R S H I P

SPE ak ERS aND MODER atORSBiondo BiondiProfessor, GeophysicsBiondo Biondi graduated from PolitecnicodiMilanoin1984andreceivedanM.S.(1988)andaPh.D.(1990)ingeophysicsfrom Stanford. He is director of the Stanford Exploration Project (SEP).SEPisanacademicconsortium whose mission is to develop innovative seismic imaging methodologies and to educate the next generation of leaders in applied seismology. SEP’s activities are supported by about 25 companies involved inOil&Gasexplorationandproduction.Biondiisalsoco-directoroftheStanfordCenterforComputationalEarthand

Environmental Science(CEES).CEESleadsthe School of Earth Sciences’ computational-

oriented research and educational programs. Biondi has made contributions on several aspects of seismic imaging, ranging from velocity estimation to parallel algorithms for seismic migration. Since the early nineties he has been at the forefront of the development ofwave-equationimagingandinversionmethods.In2004the Society of Exploration Geophysicists(SEG)honoredBiondi with the Reginald FessendenAward.In2006Biondi published the book “3DSeismicImaging”thatwasthe first book to introduce the theory of seismic imaging from the3-Dperspective.

Sally BensonDirector, Global Climate and Energy Project, Director, Precourt Institute for Energy, Professor, Energy Resources EngineeringA groundwater hydrologist and reservoir engineer, Benson has conducted research on a range of issues related to energy and

the environment, including techniques to reduce greenhouse gas emissions

bycapturingCO2 from power plants and pumping it underground for permanent sequestration. Benson’s research interests also include technologies and energysystemsforalow-carbon future, groundwater quality and remediation, and geotechnical instrumentation for subsurface characterization and monitoring. Benson was also a coordinating lead author onthe2005IntergovernmentalPanelonClimateChange(IPCC)SpecialReportonCarbonDioxideCaptureandStorage.Benson earned a bachelor’s degree in geology from Barnard CollegeatColumbiaUniversity,andamaster’sandPhDinmaterials science and mineral engineeringfromUC-Berkeley.

Jeff CaersProfessor, Energy Resources EngineeringJefCaersisdirectoroftheStanfordCenterforReservoir

Forecasting, an industrial affiliates program in reservoir forecasting with over 25 partners

from the energy industry.

Caers’researchinterestsarein the area of geostatistics, spatial-temporalmodelingandmodeling uncertainty applied to various areas in the Earth Sciences.CaersiscurrentlyEditor-in-ChiefofComputers&Geosciences.Caershaswrittentwobest-sellingbooksentitled“PetroleumGeostatistics”(SPE)and“ModelingUncertaintyintheEarthSciences”(Wiley-Blackwell).

Noah DiffenbaughAssociate Professor, Environmental Earth Systems ScienceNoahDiffenbaughstudiesthe climate system, including the processes by which climate change could impact

agriculture, water resources, and human health. Diffenbaughis currently an Editor of the

peer-reviewjournalGeophysicalResearch Letters, and a Lead AuthorforWorkingGroupIIofthe Intergovernmental Panel onClimateChange(IPCC).DiffenbaughisarecipientoftheJames R. Holton Award from the AmericanGeophysicalUnion,aCAREERawardfromtheNational Science Foundation, andaTermanFellowshipfromStanfordUniversity.Hehas also been recognized asaKavliFellowbytheU.S.National Academy of Sciences, andasaGoogleScienceCommunicationFellow.

Hector Garcia-MolinaDirector, Stanford Data Science Initiative, Professor, Electrical Engineering and Computer ScienceHectorGarcia-MolinawasthechairmanoftheComputerScienceDepartmentfromJanuary2001toDecember2004.From1997to2001he

was a member the President’s Information TechnologyAdvisory Committee

(PITAC).FromAugust1994toDecember1997hewastheDirectoroftheComputerSystems Laboratory at Stanford. From1979to1991hewasonthefacultyoftheComputerScienceDepartmentatPrincetonUniversity,Princeton,New Jersey. His research interests include distributed computing systems, digital libraries and database systems. He received a BS in electrical engineering from the Instituto TecnologicodeMonterrey,Mexico,in1974.FromStanfordUniversity,Stanford,California,hereceivedin1975aMSinelectrical engineering and a PhDincomputersciencein1979.HeholdsanhonoraryPhDfromETHZurich(2007).

5

S P R I N G 2 0 1 4 C O N F E R E N C E

Margot GerritsenDirector, Institute for Computational and Mathematical Engineering, Associate Professor, Energy Resources EngineeringGerritsen’sworkisaboutunderstanding and simulating complicated fluid flow problems.Gerritsen’sresearchfocuses on the design of highly accurate and efficient parallel computational methods to predict the performance of enhanced oil recovery methods.Outsidepetroleumengineering,Gerritsenisactivein coastal ocean simulation with colleagues from the

DepartmentofCivilandEnvironmental Engineering, yacht research and pterosaur

flight mechanics with colleagues fromtheDepartmentofMechanicalandAeronauticalEngineering, and the design of search algorithms in collaboration with the Library of CongressandcolleaguesfromtheInstituteofComputationalandMathematicalEngineering.She teaches courses in both energy related topics (reservoir simulation, energy, andtheenvironment)inherdepartment, and mathematics for engineers through the InstituteofComputationalandMathematicalEngineering.

annie HazlehurstFounder, FaridanHazlehurst is the founder of Faridan, an innovative investment vehicle focused on high growth technology companies in Africa and the MiddleEast.Previously,she

workedatDraperFisherJurvetson where she focused oninvestmentsinIT,energy,agriculture and materials. She hasalsoworkedwithCapricornInvestmentGroup,Mubadala,theIFCandMorganStanley.After finishing college, she

conducted environmental field research inZambiaandtaught English in parts of Southeast

Asia. Hazlehurst graduated fromBrownUniversitywithaB.S. in Biotechnology where she completed a Hughes Fellowship that evaluated the development of tissue engineering and stem cell technologies. She receivedanM.S.Environmentand Resources from Stanford Universitywhereshefocusedon materials science and engineering for energy related technologies.ShehasanMBAfromtheStanfordGraduateSchool of Business.

Jeff koseffCo-Director, Stanford Woods Institute for the Environment, Professor, Civil and Environmental EngineeringKoseff received his bachelor’s degree in civil engineering fromtheUniversityoftheWitwatersrandin1976.After working for a year in

Johannesburg, he came to Stanford where he received anM.Sincivilengineering in 1978andaPh.D.

incivilengineeringin1983.Koseff joined the faculty of theDepartmentofCivilandEnvironmental Engineering in 1984,whereheservedastheDepartmentChairfrom1995to1999andthenastheSeniorAssociateDeanofEngineeringfrom1999to2003.Since2003,hehasservedasthePerryL.McCartyDirectoroftheStanfordWoodsInstitutefor the Environment. At the 2011CivilandEnvironmentalEngineering graduation ceremony, Koseff received theEugeneL.GrantAwardfor teaching in recognition of“hiscontinueddedicationand excellence in teaching as voted by the students of the Department.”EugeneGrantwasalong-servingandhighlydistinguished member of the facultyatStanford.Theawardwas established by his students to honor his contributions. KoseffalsoreceivedtheGrantAwardin1995.

Philip LevisAssociate Professor, Computer Science and Electrical EngineeringPhilip Alexander Levis is an Associate Professor of

ComputerScienceand Electrical Engineering at Stanford University,wherehe heads the

Stanford Information Networks Group(SING).Hisresearchcenters on computing systems that interact with or represent the physical world, including low-powercomputing,wirelessnetworks, sensor networks, embedded systems, and graphics systems. He has been awardedtheOkawaFellowship,

anNSFCAREERaward,andaMicrosoftNewFacultyFellowship. He’s authored over 60peer-reviewedpublications,including three best paper awards, one test of time award, and one most influential paper award at premier conferences. Levis has a Sc.B. in Biology andComputerSciencewith Honors from Brown University,aM.S.inComputerSciencefromtheUniversityofColoradoatBoulder,andaPh.D.inComputerSciencefromtheUniversityofCalifornia,Berkeley.Hehasaself-destructiveaversiontolow-hangingfruitandadeepappreciation for excellent engineering.

Sara MenkerCEO and Founder, Gro IntelligenceSaraMenkerisFounderandCEOofGroIntelligence,abusiness that uses big data to enable global food security and isco-headquarteredinNairobiandNewYork.GroIntelligence

develops products that serve as risk management tools resulting in better access to

capital and higher productivity for agriculture. Prior to founding GroIntelligence,MenkerwasaVicePresidentinMorganStanley’sCommoditiesGroup.She studied Economics and AfricanStudiesatMountHolyokeCollegeandtheLondon School of Economics, andearnedanMBAfromColumbiaUniversity.MenkerisaTrusteeoftheMandela

6

S P R I N G 2 0 1 4 C O N F E R E N C E

Balaji PrabhakarProfessor, Electrical Engineering and Computer Science Prabhakar’s

research interests are in computer networks; notably, in designing algorithms for theInternetandforDataCenters.Recently,hehasbeen interested in Societal Networks: networks vital for society’s functioning, such as transportation, electricity and recycling systems. He has been involved in developing and deploying incentive mechanisms to move commuterstooff-peaktimesso that congestion, fuel and pollution costs are reduced. HehasbeenaTermanFellowatStanfordUniversityandaFellow of the Alfred P. Sloan Foundation. He has received theCAREERawardfromtheU.S.National Science Foundation, the Erlang Prize, the Rollo DavidsonPrize,anddeliveredthe Lunteren Lectures. He is aco-recipientofseveralbestpaper awards.

Ram RajagopalAssistant Professor, Civil and Environmental EngineeringProfessor Rajagopal leads a laboratory for creating sustainable engineering

systems, with renewable energy systems as one of the main focus areas. He received his

Ph.D.inelectricalengineeringand computer sciences and

master’s degree in statistics fromtheUniversityofCalifornia,Berkeley. He has specialized in creating and deploying large sensing systems, and using the generated data together with novel statistical algorithms and stochastic control to achieve sustainable transportation, energy, and infrastructure networks.

Mary RuckelshausManaging Director, Natural Capital ProjectMaryRuckelshausoverseesallworkoftheNaturalCapitalProject, including the science, software, applications, strategy, partnerships, and fundraising. Ruckelshaus is based in Seattle, WA,whereshepreviouslyled the Ecosystem Science ProgramatNOAA’sNWFisheriesScienceCenter.Priorto that, she was an Assistant Professor of biological sciences atTheFloridaStateUniversity

(1994-1997).Themain focus of her recent work is on developing ecological models including

estimates of the flow of environmental services under different management regimes in marine and freshwater systems worldwide. Ruckelshaus servesontheScienceCouncilofTheNatureConservancyandisChairofitsWashingtonBoardofTrustees,andisapastchair of the Science Advisory BoardoftheNationalCenterfor Ecological Analysis and Synthesis(NCEAS).RuckelshauswasChiefScientistforthePugetSoundPartnership,apublic-private institution charged with achieving recovery of the Puget Sound terrestrial, freshwater

and marine ecosystems. Ruckelshaus has a bachelor’s degree in human biology from StanfordUniversity,amaster’sdegree in fisheries from the UniversityofWashington,andadoctoral degree in botany, also fromWashington.

Barton “Buzz” thompson Jr.Co-Director, Stanford Woods Institute for the Environment, Professor, Natural Resources LawThompson’sresearchfocuseson the sustainable use of natural

resources and the effective reform of regulatory institutions. Theauthorofseveral books

on water, the environment, andproperty,Thompsonhas published articles on such diverse topics as water markets, fisheries management, biodiversity protection, land conservation, the use of economics and market tools in environmental regulation, and cognitive barriers to resource management.Thompsonischairman of the board of the Resources Legacy Fund and the Resources Legacy Fund Foundation,aCaliforniatrusteeforTheNatureConservancy,and a board member of both theAmericanFarmlandTrustand the Sonoran Institute. He previously served as a member of the Science Advisory Board fortheU.S.EnvironmentalProtectionAgency.In2008,theSupremeCourtappointedThompsontoserveasthespecialmasterinMontanav.Wyoming(137Original).

InstituteofDevelopmentStudies(MINDS)andistheChairoftheBoardofTruthAid,an organization which harnesses the power of multimedia in order to effect social change. She is a fellow of the African Leadership Initiative of the Aspen Institute and was named aYoungGlobalLeaderbytheWorldEconomicForum.

amit NarayanCEO and Founder, AutoGridAmit Narayan is the Founder andCEOofAutoGrid,Inc.From2010to2012,hewasthe

DirectorofSmartGridResearchinModeling&Simulationat Stanford University,

where he continues to lead an interdisciplinary project related to modeling, optimization and control of the electricity grid and associated electricity markets. Prior to founding AutoGrid,NarayanwastheVice President of Products at the publicly traded company MagmaDesignAutomation,Inc.(Nasdaq:LAVA),whereheled the product development and product management teamsresponsibleforMagma’sflagship product in the design implementation area. NarayanreceivedhisB.Tech.in Electrical Engineering from IndianInstituteofTechnologyatKanpurandPh.D.fromUniversityofCaliforniaatBerkeley. He has published over 25 papers about design automation,holdssevenU.S.patents and is an active advisor to several startup companies in the Bay Area.

7

S P R I N G 2 0 1 4 C O N F E R E N C E

8

S P R I N G 2 0 1 4 C O N F E R E N C E

S T U D E N T A N D P O S T D O C T O R A L P O S T E R S

StuDENt/POStDOCtORaL aDVISOR tItLE

Adrian Albert, Electrical Engineering James Sweeney and Ram Rajagopal

Data-Driven Energy Demand Management for the Smart Grid

Jason Chang, GeophysicsTaylor Dahlke, GeophysicsNori Nakata, Geophysics

Biondo Biondi Resolving Subsurface Structure from Cross-Correlations of Continuously Recorded Ambient Noise at Long Beach, CA

Jungsuk Kwac, Electrical Engineering

Ram Rajagopal Segmenting Customers from Smart Meter Data

Lewis Li, Energy Resources Engineering

Jef Caers Using Analogous Data for Subsurface Characterization

Chinmoy Mandayam, Electrical EngineeringChenguang Zhu, Computer Science

Balaji Prabhakar Congestion and Parking Relief Incentives

Sid Patel, Civil and Environmental Engineering

Ram Rajagopal Aggregation for Load Servicing

Raffi Sevlian, Electrical Engineering Ram Rajagopal A Model for the Effect of Aggregation on Short Term Load Forecasting

Saahil Shenoy, Physics Dimitry Gorinevsky Risk Management in Forecasting for Electricity Markets

Pejman Tahmasebi, Energy Resources Engineering

Jef Caers High Performance Computational Methods for Reservoir Modeling

Chuan Tian, Energy Resources Engineering

Roland N. Horne Machine Learning for Well Test Interpretation

9

S P R I N G 2 0 1 4 C O N F E R E N C E

Jason ChangDepartment:GeophysicsAdvisor: Biondo BiondiPosterTitle:ResolvingSubsurface Structure from Cross-CorrelationsofContinuouslyRecordedAmbientNoiseatLongBeach,CA(jointposterwithTaylorDahlkeandNoriNakata)ResearchDescription:Byusingcross-correlationtechniques,continuous recordings of the ambient seismic noise field can be transformed into coherent seismicsignal.Whenthesemethods are applied to data recorded by a seismic ar-ray that is large, dense, and long-duration,theresultingrecovered signals can be used toaddresslocal-scaleenvi-ronmental problems. Such an array is located in Long Beach, CA,whereChangisabletoprocess large amounts of data to resolve wave types that were previously difficult to recover, includinghigh-frequency

surface waves and body waves. Theseseismicwaves have the potential to be used to resolve

near-surfacestructuresthatarevital to fields such as earth-quake hazard analysis and groundwater monitoring.

taylor DahlkeDepartment:GeophysicsAdvisor: Biondo BiondiPosterTitle:ResolvingSubsurface Structure from Cross-CorrelationsofContinuouslyRecordedAmbientNoiseatLongBeach,CA(jointposterwithJasonChangandNoriNakata)ResearchDescription:Dahlke’sresearchusescross-correlationsgenerated from massive quantities of ambient seismic noiseasvirtualshot-receiverpair signals. By transforming

these signals into thefrequency-wavenumber domain, he can find phase infor-mation which can

betransformedintotime-delaysurfaces, ultimately allowing him to extract local wavespeed information to get tomograph-icresults.Becausecross-cor-relation data volume extends over time and virtual shot for eachreceiverpoint,Dahlkehasample sampling from which he cangeneratestatistics.Thesestatistics can be used to gener-ate basin wide models for earthquake hazard modelling.

Jungsuk kwacDepartment:ElectricalEngineeringAdvisor: Ram RajagopalPosterTitle:SegmentingCustomersfromSmartMeterDataResearchDescription:InKwac’sPhDresearch,hedevelops

methodologies for segmenting customer energy consumption life-styles and for tar-geting customers

inenergyprograms.Duetothelarge number of people and the huge size of the consump-tion data set, any methodology should guarantee scalability.

Lewis LiAdvisor:JefCaersDepartment:EnergyResourcesEngineeringPosterTitle:UsingAnalogousDataforSubsurfaceCharacterizationResearchDescription:Themaindifficulty behind forecasting reservoir performance is a lack

of information re-garding the sub-surface.Wellboredata can provide accurate, but sparse informa-

tion.Traditionalgeostatisticsfocuses on parametric model-ing to interpolate the well data for characterizing the reser-voir.Thisisbasedonrandom

adrian albertDepartment:ElectricalEngineeringAdvisors: James Sweeney and Ram RajagopalPosterTitle:Data-DrivenEnergyDemandManagementfortheSmartGridResearchDescription:Demand-sidemanagement

is seen as a cost-effective,environmentally-considerate alternative to investments in

costly, polluting generation capacity to address the increas-ingdemandforelectricity.Tobetter monitor demand, utility companies have deployed millionsof“smartmeters”thatcollect energy consumption dataatfine(sub-hourly)timescales. Yet little understanding exists currently about how such information might be used to improve operational prac-tices. Albert introduces novel static and dynamic models of individual residential user con-sumption that use smart meter datatoidentifyhigh-potentialusersfortargetingdemand-side energy management pro-grams and shows that certain user characteristics (appliances, lifestyle)maybepredictedusing features computed from the data.

StuDENt aND POStDOCtORaL SCHOLaRS

S P R I N G 2 0 1 4 C O N F E R E N C E

functiontheory(RFT),andcanbedifficultfornon-expertstograsp. Furthermore, it requires a subjective mathematical decomposition of the data into trendandresidual.Conversely,wepresentanon-parametricapproach that does not rely on RFT.Instead,itreliesoncom-puting statistics on analogous data (such as from neighboring fields)anddirectlywithoutanymodeling produce interpola-tions between existing wells.

Chinmoy MandayamDepartment:ElectricalEngineeringAdvisor: Balaji PrabahakarPosterTitle:CAPRI:Congestionand Parking Relief Incentives (jointposterwithChenguangZhu)ResearchDescription:CAPRIis a program to incentivize Stanford commuters to shift automobile commute times to offpeak hours, and to take

up bicycling and walking com-mutes.CAPRIincludes a status system and magic boxes

featuring personalized offers to encourage participation. Currently,thereareover4,500registered users, including over 850walkersandbikers.SinceApril2012,morethan550,000automobile commutes and over50,000walkingandbikingtripshavebeenrecorded.Tobetter understand the impact

of incentive mechanisms on users’ commute behavior, we employ a generative user model to classify users based on performance and character-ize the temporal dynamics of behavioralshift.Theresultsshow that our system has a siz-able impact on users’ decision to shift automobile commute times to offpeak hours.

Nori Nakata Department:GeophysicsAdvisor: Biondo BiondiPosterTitle:ResolvingSubsurface Structure from Cross-CorrelationsofContinuouslyRecordedAmbientNoiseatLongBeach,CA(jointposterwithJasonChangandTaylorDahlke)ResearchDescription:Withcross-correlationtechniques,Nakata extracts useful subsur-

face information from continuous seismic data. Pre-andpost-correlation filters allow him to ob-

taincross-correlationfunctionswithhighsignal-to-noiseratio.Nakata can further use these functions to image subsurface. Inthisstudy,heappliesatwo-steppost-correlationfiltertoretrieve body waves from ambi-entnoise,andinvertP-wavevelocities using tomography techniques.

Sid PatelDepartment:CivilandEnvironmental EngineeringAdvisor: Ram RajagopalPosterTitle:AggregationforLoad ServicingResearchDescription:Theproliferation of smart meters enables a load serving entity (LSE)toaggregatecustomersaccording to their consumption patterns. Patel demonstrates a method for constructing groups of customers who will be the cheapest to service at wholesalemarketprices.Usingsmart meter data from a region inCaliforniaandaggregatingmore of these customers to-gether, their consumption can be forecasted more accurately, which allows an LSE to mitigate

financial risks in its wholesale market trans-actions. Patel observes that the consumption of

aggregates of customers with similar consumption patterns can be forecasted more ac-curately than that of random aggregatesofcustomers.Themodel he proposes enables an LSE to offer discounted rates tolow-costcustomersbecauseit can purchase electricity for them more cheaply than it can for the general population.

Raffi SevlianDepartment:ElectricalEngineeringAdvisor: Ram RajagopalPosterTitle:AModelfortheEffect of Aggregation on Short TermLoadForecastingResearchDescription:Inthiswork, Sevlian proposes a simple empirical scaling law that describes load forecast-ing accuracy at different levels of aggregation. He shows that for the short term forecasting

problem, aggre-gating more users will improve the relative forecast-ing performance up to a point.

Beyond this point, no more improvement in relative perfor-mance can be obtained.

Saahil ShenoyDepartment:PhysicsAdvisor:DimitryGorinevskyPosterTitle:RiskManagementin Forecasting for Electricity MarketsResearchDescription:Robuststatistics of load forecasting of energy demand is usually focused on mean values in related statistical models and ignores rare peak events. It

also models the distribution of deviations naively as a normal distri-bution. Extreme value theory, a

well-establishedbranchofstatistics that analyzes large deviation events that are long

10

S P R I N G 2 0 1 4 C O N F E R E N C E

ingimage.Then,smallpiecesof the patterns in the training image are stochastically an-chored to the available dataset. Thisposterwillpresentalgo-rithmic and hardware solutions that address the high compu-tational cost for very large and realreservoirs.Thealgorithmicsolutionusesamulti-scaleframework to present the train-ingimagemoreefficiently.Thehardware solution uses cluster computing,GPUs,andHadoopto increase efficiency. By com-bining these solutions we were abletosimulateamulti-millioncell grid in a matter of seconds.

Chuan tianDepartment:EnergyResourcesEngineeringAdvisor: Roland N. HornePosterTitle:MachineLearningforWellTestInterpretationResearchDescription:Tian’sresearch is focused on the

interpretation of flow rate – pres-sure – tempera-ture(q-p-T)datafrom Permanent Downhole

Gauges(PDGs)usingmachinelearningapproaches.Theideais to develop the algorithm to describe the pattern behind q-p-Tdata,whichcontainsthe

reservoir information implicitly. By inputting a simple flow rate history into the algorithm and askingforp-Tprediction,Tianis able to extract the reservoir parameters as conventional welltesting.ThismethodhelpstoutilizePDGsdatatoestimate reservoir parameters without additional operations, e.g. shut in the wells.

Chenguang zhuDepartment:ComputerScienceAdvisor: Balaji PrabhakarPosterTitle:CAPRI:Congestionand Parking Relief Incentives (jointposterwithChinmoyMandayam)ResearchDescription:CAPRIis a program to incentivize Stanford commuters to shift automobile commute times to off peak hours, and to take up bicycling and walking commutes.CAPRIincludesastatus system and magic boxes featuring personalized offers to encourage participation. Currently,thereareover4,500registered users, including over 850walkersandbikers.SinceApril2012,morethan550,000automobile commutes and over50,000walkingandbiking

trips have been recorded.Tobetter under-stand the impact of incentive mechanisms on

users’ commute behavior, we employ a generative user

model to classify users based on performance and character-ize the temporal dynamics of behavioralshift.Theresultsshow that our system has a siz-able impact on users’ decision to shift automobile commute times to off peak hours.

tailed, ignores the bulk of the data.Thisresearchdevelopsstatistical modeling and esti-mation approaches by combin-ing robust regression and long tailestimation.Theapproachallows Shenoy to estimate the regression model, distribution body, distribution tails, and boundaries between the body and the tails. As an application example, the model is esti-mated for historical power load data from an electrical utility. Practical usefulness of the model is illustrated by stochas-tic optimization of electricity orderinday-aheadmarket.

Pejman tahmasebiDepartment:EnergyResourcesEngineeringAdvisor:JefCaersPosterTitle:HighPerformanceComputationalMethodsforReservoirModelingResearchDescription:Geo-statistical methods are used widely in the earth sciences. Mosttraditionalmethodsrelyontwo-pointcovariance,whichhas been proven to be insuffi-cient to capture the complexity in reservoir modeling. Recently,

multiple-pointgeostatistical simulations using higher-orderstatistics have received more at-

tention.Theconceptistofirstpresent the prior conceptual informationthrougha3Dtrain-

11

The Energy and Environment affiliates Program is a membership program that helps corporations and other organizations to engage with Stanford faculty and graduate students in energy-related and environmental areas. The Affiliates Program is about establishing relationships and facilitating communication.

E E a P M E M B E R S

Yang & Yamazaki Environment & Energy Bldg., MC 4240473 Via Ortega

Stanford, CA 94305http://eeap.stanford.edu

Cateringby:

We’vewontheAwardofExcellence forourGreenachievements

Videotaping by:

Geballe Laboratoryfor Advanced Materials

The Geballe Laboratory for Advanced Materials (GLAM) facilitates education and research on advanced materials in science and engineering.

The Woods Institute is pioneering innovative approaches to meet the environmental challenges of the 21st century.

The Precourt Institute for Energy (PIE) engages in a broad-ranging, interdisciplinary program of research and education on energy.

Energy and EnvironmentA� iliates Program

S P R I N G 2 0 1 4 C O N F E R E N C E

Printed on recycled paper.