通過數字化轉型構建人工智能基礎.pdf

《通過數字化轉型構建人工智能基礎.pdf》由會員分享，可在線閱讀，更多相關《通過數字化轉型構建人工智能基礎.pdf（80頁珍藏版）》請在三個皮匠報告上搜索。

1、 2025 Electric Power Research Institute,Inc.All rights Developing the Foundation for AI through Digital TransformationArtificial Intelligence(AI)and Digital Transformation(DX)Electric Power SummitJanuary 8,2025Technical Presentation Session A4 2025 Electric Power Research Institute,Inc.All rights re

2、served.2Topics&Speakers Daniel Roesler,UtilityAPI Hallie Carrao,Google Call for Carbon Data Transparency fueling Utility Digital Transformation Cassie Shaban,EPRI EPRI and the Digital Transformation Research Initiative(DXRI)Abder Elandaloussi,Southern California Edison Digital Twins Definition,Consi

3、derations,and Use Cases Gianluca Lipari,EPRI AI for digital transformation and clean energy transition in the electric power industry,opportunities and challengesCall for Carbon Data Transparency fueling Utility Digital TransformationCreating open standards for utility connectivity and data access20

4、25-01-08-EPRI AI/DX Summit-Palo AltoSpeakers:Hallie Carrao(Google)and Daniel Roesler(UtilityAPI)Project Motivation:Why Google got involvedHallie Carrao,GoogleGoogles Energy JourneyWhy 100%Annual Renewable Energy Matching is Not EnoughInterest in the CDSCAiming to accelerate pathways to grid decarbon

5、ization by providing data that is:Better data is a foundational building block to develop the solutions needed for any clean energy future.Tour of discovery and connectivity specsconnectivity.carbondataspec.orgThe problem:Getting data once is easyGetting data at scale is hardWhats hard about getting

6、 utility customer data at scale?Need way more data nowadays than just the customers bills or monthly kwhMost customers have no idea how to get their data needed for analysisIndividual customer data is private and sensitive and requires consent to getFragmented across many utilities and other central

7、 authoritiesData formats are mostly ad-hoc or proprietary with little documentationHard to find what utilities offer what methods of access and what vendors can qualifyMany dont offer digital methods of access(paper forms or manual processes only)For those that do have digital access,vendor registra

8、tion is often slow,complex,and expensiveExisting standards for digital access are proprietary,complex,and have major gapsRegulatory oversight is highly complex because current specifications require expert knowledgeSoooo what are the components of a scalable connectivity standard?What are the compon

9、ents of a scalable connectivity standard?DiscoveryFind what utilities and other central entities that offer data access.What are the components of a scalable connectivity standard?DiscoveryFind what utilities and other central entities that offer data access.Registration+ConnectivitySign up and esta

10、blish connections with utilities.What are the components of a scalable connectivity standard?DiscoveryFind what utilities and other central entities that offer data access.Registration+ConnectivitySign up and establish connections with utilities.CDSC-WG1-01(Server Metadata)What are the components of

11、 a scalable connectivity standard?DiscoveryFind what utilities and other central entities that offer data access.Registration+ConnectivitySign up and establish connections with utilities.CDSC-WG1-01(Server Metadata)CDSC-WG1-02(Client Registration)What are the components of a scalable connectivity st

12、andard?CDSC-WG1s Connectivity Specifications:CDSC-WG1-01(Server Metadata)How utilities and other centralized entities can disclose what data and capabilities they offer.CDSC-WG1-02(Client Registration)How vendors can register,onboard,and manage their connectivity with utilities and other centralized

13、 entities.CDSC-WG1s Connectivity Specifications:CDSC-WG1-01(Server Metadata)How utilities and other centralized entities can disclose what data and capabilities they offer.CDSC-WG1-02(Client Registration)How vendors can register,onboard,and manage their connectivity with utilities and other centrali

14、zed entities.From:https:/daniel-utilityapi.github.io/Connectivity/specs/cdsc-wg1-02/overview From:https:/daniel-utilityapi.github.io/Connectivity/specs/cdsc-wg1-01/overview Who can benefit from these connectivity specifications?Other specifications!Our first two specifications(CDSC-WG1-01/02)can be

15、used by other standards to streamline third party service provider registration and onboarding(DER,EE,EV,DR,etc.).For example,our Customer Data specification(CDSC-WG3-01)uses these connectivity specs as its base.By unblocking service provider registration and onboarding,we accelerate adoption of new

16、 innovative programs,grid flexibility,and deployment of clean energy technologies.Tour of customer data specscustomerdata.carbondataspec.orgWhat are the components of a scalable customer data access standard?DiscoveryFind what utilities and other central entities that offer data access.Registration+

17、ConnectivitySign up and establish connections with utilities.CDSC-WG1-01(Server Metadata)CDSC-WG1-02(Client Registration)What are the components of a scalable customer data access standard?DiscoveryFind what utilities and other central entities that offer data access.Registration+ConnectivitySign up

18、 and establish connections with utilities.AuthorizationGet individual customer consent to access their data.CDSC-WG1-01(Server Metadata)CDSC-WG1-02(Client Registration)What are the components of a scalable customer data access standard?DiscoveryFind what utilities and other central entities that off

19、er data access.Registration+ConnectivitySign up and establish connections with utilities.AuthorizationGet individual customer consent to access their data.Data AccessRetrieve authorized customer data in a standardized format.CDSC-WG1-01(Server Metadata)CDSC-WG1-02(Client Registration)What are the co

20、mponents of a scalable customer data access standard?DiscoveryFind what utilities and other central entities that offer data access.Registration+ConnectivitySign up and establish connections with utilities.AuthorizationGet individual customer consent to access their data.Data AccessRetrieve authoriz

21、ed customer data in a standardized format.CDSC-WG1-01(Server Metadata)CDSC-WG1-02(Client Registration)CDSC-WG3-01(Customer Data)CDSC-WG3s Customer Data Specifications:CDSC-WG3-01(Customer Data)How vendors can obtain customer authorization and how the authorized customer data should be formatted.Targ

22、et use cases:Carbon AccountingDecarbonization ProjectsDistributed Grid FlexibilityBuilding BenchmarkingFrom:https:/daniel-utilityapi.github.io/Customer-Data/specs/cdsc-wg1-03/overview How we can solve the grid connectivity problem.External entities(customers,3P service providers)Step 1:External enti

23、ties establish secure connectivity with utilities and central entities.Examples:AI data tools Carbon Accounting EE/Bldg Management VPPs/Microgrids DER/Load Flexibility C&E/Data Centers EV Infrastructure Vendors/Consultants Program Administrators and more!LF Energy CDSC-Architecture DiagramUtilities

24、andother central entitiesExamples:IOUs Municipal utilities Co-ops CCAs ISOs TSOs DSOs Centralized data hubs (regional/state/national)Energy markets and more!CDSC-WG1-01-Server MetadataCDSC-WG1-02-Client RegistrationSecure,standardized protocols for utilities andother central authorities to offer dig

25、ital discovery,registration,and connectivity to third party entities.CDSC-WG3-01-Customer DataDER reporting/curtailment dataStep 2:Secure data exchange and communicationsbetween parties using connectivity from Step 1.CDSC-WG2-01-Power System DataEE program participation detailsVPP/load management pr

26、otocolsCharging infrastructure protocols(other protocols can also use these specs for establishing connectivity)How you can get involved!How to get involved:1.Take our survey:Go to carbondataspec.org Take Survey2.Join the mailing lists:https:/connectivity.carbondataspec.org/https:/customerdata.carbo

27、ndataspec.org/3.No seriously,take the survey!Q&A 2025 Electric Power Research Institute,Inc.All rights Susan Maley,Technical Executive Cassie Shaban,Program LeadColton Smith,Technical Leader,Sr January 2025Digital Transformation Research Initiative 2025 Electric Power Research Institute,Inc.All righ

28、ts reserved.35AgendaTime to Make a DecisionIntentional DXIncremental ProgressEngaging in DXEnables AIDXResearch Focus AreasStrategyPeople&ProcessDataImplementationSynthesizing for Success 2025 Electric Power Research Institute,Inc.All rights reserved.36Leveraging data and information to improve equi

29、pment and personnel performance at a fleet levelEnsuring equipment reliability and asset health is critical to meeting variable energy demand and critical dispatchabilityConnecting and integrating data streams and business functions to improve human performance and asset operationsIT and OT integrat

30、ion with more IT equipment in the DCS and shared responsibilities for maintaining and securing equipmentDigital Transformation(DX):Enabling the Energy TransformationDigital Transformation Foundational to the Energy TransformationPerformanceReliability and ResiliencyOperational EfficiencySecurity 202

31、5 Electric Power Research Institute,Inc.All rights reserved.37ImpactTime to Make a DecisionIntentional DXIncremental Progress(+)Decrease costs(+)Process optimization(+)Workforce efficiencies(+)Increase safety Energy transformation Data as a strategic asset Develop pathways for success Historically h

32、igh failure rates of DX projects and low utilization rates of installed “solutions”MotivationDX 2025 Electric Power Research Institute,Inc.All rights reserved.38MissionDigital Transformation Research Initiative Why Digital Transformation?Project TeamCassie Shaban,Colton Smith,Susan Maley,SPN Number:

33、3002027598Collaborative research and development to support digital transformation(DX)in nuclear,thermal,and renewable power generation with a focus on strategy,people,process,data,and implementationLeverages expertise from EPRI Generation,Nuclear,Energy Delivery Customer Solutions,IT,and partners w

34、ith industry leading organizations around the worldValue of CollaborationResearch Focus AreasStrategyPeople&ProcessDataImplementationSynthesizing for SuccessDX R&D RoadmapDX Maturity Model&Success StoryDX Application Areas Maturity ModelsFostering a holistic approach required for implementation of D

35、X applicationsFacilitating a forum to bring together SMEs across disciplinesFocus on the how of implementationCollaborating with industry groups(e.g.,Nuclear Information Technology Strategic Leadership,Nuclear Energy Institute(NITSL),Idaho National Laboratory,National Energy Technology Laboratory,RI

36、TSL,EPRIs WIN and SOL)Unique InsightsImpactful Content 2025 Electric Power Research Institute,Inc.All rights reserved.39Power Generating Industry DX Temperature ReadStrategy&BusinessDataTechnology ProcessOrganizational Change ManagementTraining&MetricsL5=Fully Mature and Doing Great!L4=Measurable Pr

37、ogress L3=Moving in Good DirectionL2=Making some ProgressL1=StartedL0=0,zilch,nada?2025 Electric Power Research Institute,Inc.All rights reserved.40Digital Transformation Research Initiative(DXRI)Strategy,Integration&ImplementationDoes your company wide DX strategy include AI?Enables adoption and im

38、plementation Allows for flexibility for future growth Strategy,data,infrastructure,people&cultureThings to Consider Executive level supporters and champions at every level Standards,policies and procedures Communication flows between key stakeholder teams Data continuous,catalogued,labeled and readi

39、ly available for central processing Infrastructure readiness Correct set of instrumentation for the tool Mechanism for support and maintenance of third-party modelsOpportunitiesChallengesDigital processes not well establishedLow digital maturity Low system level readinessHigh technology readiness le

40、vels for individual digital technologies 2025 Electric Power Research Institute,Inc.All rights reserved.41Reasons for FailureReference:Rabin Research Company 2025 Electric Power Research Institute,Inc.All rights reserved.42Power Generating Industry DX Challenges:People,Process and Technology“Technic

41、al side”executed byproject management discipline“People side”executed bychange management discipline 2025 Electric Power Research Institute,Inc.All rights reserved.43Power Generating Industry DX Challenges:2025 Electric Power Research Institute,Inc.All rights reserved.44Power Generating Industry DX

42、Challenges:Need to Fortify FoundationCurrent data governance practices and infrastructure is not AI readyIdentify short-and long-term goals and objectives and how they connect with plant/fleetwide goalsIdentify gaps and challenges Pinpoint quick wins to get started and milestones keep momentum going

43、Tie AI strategy with company wide DX strategy 2025 Electric Power Research Institute,Inc.All rights reserved.45Evolution of DX Use Case:Continuous Online MonitoringAdvanced Pattern Recognition Adoption of Tool:Evaluate current equipment condition Early identification of equipment anomaliesTransition

44、 to Condition-Based Maintenance DX of the Process:Building out diagnostics Using results to inform maintenance and outage planningExpert Systems for Auto Diagnosis Enhance Tool+Process:Auto diagnose failure modes Advanced data analytics to model specific failure mechanisms 2025 Electric Power Resear

45、ch Institute,Inc.All rights reserved.46Digital Transformation Research Initiative(DXRI)Strategy,Integration&ImplementationVisit the DXRI site Key Insights Technology Trends:Utilities are piloting and deploying a variety of different digital transformation(DX)applications Strategy:How to motivate emp

46、loyees to embrace DX?Focus is messaging on value proposition.oImprove safety&reducing low level tasks(manual/redundant)tasksoFreeing up time to focus on more higher quality work Organizational Change Management:It is critical for success and must be prioritized from onset and enforced through entire

47、 project Business Case:Important to demonstration quick wins where a significant impact with a clear favorable business case may be promptly deliveredProject Team:Cassie Shaban(),Colton Smith()and Susan Maley()2025 Electric Power Research Institute,Inc.All rights reserved.47 2025 Electric Power Rese

48、arch Institute,Inc.All rights TOGETHERSHAPING THE FUTURE OF ENERGY 2025 Electric Power Research Institute,Inc.All rights reserved.48Reference Slides 2025 Electric Power Research Institute,Inc.All rights reserved.49Primary use of data analytics in their jobs?Anomaly detectionForecasting asset healthC

49、ondition assessmentPlant optimizationRoot cause analysisAutomationOutage PlanningAnswering questions from plants or other teams 2025 Electric Power Research Institute,Inc.All rights reserved.50Consequences of the“Swiss Cheese Future State”Instead ofLowerROIHistory of Failed ChangesUnachieved Improve

50、mentNot What We Expected/Hoped For 2025 Electric Power Research Institute,Inc.All rights reserved.51Costs:Risks:To the project if we do not manage the people side of change wellProject delaysMissed milestonesBudget overrunsRework requiredLoss of work by project teamResistance active and passiveProje

51、ct put on holdResources not made availableObstacles appear unexpectedlyProject fails to deliver resultsProject is fully abandonedTo the organization if we do not manage the people side of change wellProductivity plunges(deep and sustained)Loss of valued employeesReduced quality of workImpact on cust

52、omersImpact on suppliersMorale declinesLegacy of failed changeStress,confusion,fatigueChange saturationTo the organization if this change does not deliver the results we expectLost investment in the projectLost opportunity to have investedin other projectsExpenses not reducedEfficiencies not gainedR

53、evenue not increasedMarket share not capturedWaste not reducedRegulations not metAvoidable Costs and Mitigable RisksDigital Twins for Electric Utilities:Definitions,Considerations,andApplicationsAbderElandaloussiMSEE,MBA,PE 1/8/2025AGENDABackgroundDriversDefinitionCore ComponentsMaturityUse CasesExi

54、stingInitiativesOpportunitiesfor InnovationConclusion53BACKGROUNDObjective:White Paper covering the over-archingdefinition and vision for Digital Twins specifically for the electric utility industry that will drive the industry toinnovate new products,services,standards,processes,and even regulation

55、(Too)Many definitions exist todayOften mistaken with functionalities of ADMS,DMS,OMS,DERMSDiverse collaboration across industry partnersUtilitiesVendorsResearch LabsRegulatory54Using Tools of Yesterday toManage the Grid of TomorrowThe grid DNA is changing as we speak and expected todrastically be di

56、fferent by 2045,but our approaches toplan,design,engineer,operate,and maintain it are notchanging at the same speedThe grid is becoming unpredictable-load,generation,customers,assetsExisting models used are often outdated due to large backlogsIt is difficult to integrate context into engineering fun

57、ctions-could be the difference maker for more informed decisionsExisting simulation tools that leverage physics-basedsimulation maybe not adequate for the complexitiesthe grid will faceBig data holds tremendous value,but is under-utilized in existing decision making across engineering,planning,and o

58、perationsClimateIndustrial LoadsEVsCyber-Physical SecuritySocio-Economic55DT DefinitionElectric Utility SpecificHigh Fidelity,virtual and visual representation of systems of systems representingelectric grid,its surroundings,and any relevant contextSystem:Capacitor Bank Feeder Substation Entire Grid

59、Real Time and/orOfflineAnalyze thecurrent state of the gridand predict future stateof the grid56DT 2.0Epiphany-We have beenusing Digital Twins for a Long TimeAdvanced Intelligence:Enhanced AI/ML capabilitiesAutonomy:Self-optimizing and decision-supportfeaturesDynamic Interoperability:Seamless integr

60、ationwith diverse systemsand data sourcesScalability:Modular,cloud-based architecturesfor widespread deploymentHolistic Understanding:Evaluate interdependenciesbetween the electrical domain and other domains across G,T,D,BTM57DigitalTwin 2.0 representsthe next evolution in digital twin technologyWhy

61、 DT 2.0?Its not for everyone.YET58IncreasingGridComplexity:-Managinghighpenetration of DERs,EVs,and inverter-basedresourcesAgingInfrastructure-Traditional manufacturerdatasheetsand extrapolationtechniquesforassetmaintenancearentenoughAnomaliesbecomingtheNorm-anomaliesaredifficult totrack andlearnfro

62、mwhentheyincreaseinfrequencyDataOverload-Bigdataisbecomingtoocumbersometo analyzethroughqueries.InsightsarelimitedLowFidelitySimulation-Howaccuratearethemodelsand predictionscomparedtoreality?DT Maturity8Descriptive TwinMost Utilities Have some Level of ImplementationLive,editableversion of grid des

63、ign and construction dataInformative Twin Integration of sensor and other live data into models to gain more contextPredictive Twin Leverage data for insights and predicting futureeventsComprehensiv e Twin Simulatefuture scenariosand consider“what if”questionsAutonomou s TwinSelf-sustaining,assist w

64、ith decision-making by informing stakeholders on business,engineering,and operations decisions in real timeFundamental 2.0 ComponentsModel Maintenance:Automated,real-time updatesfordynamicgrid modelsSimulationEngine:Multi-domain simulations(electrical,structural,cyber)DataManagement:Unify and QC mul

65、ti-modal datasetsInteractiveVisualization:3D,context-rich representationsfor better decision-makingAI/MLIntegration:Predictiveinsightsand autonomous recommendationsCybersecurity:Protecting Critical Energy Infrastructure Information(CEII)60USECASESDesignPlanningGrid OperationsAsset ManagementGenerati

66、onElectrification AI/MLProtection,Control&Automation Risk&ComplianceCybersecurity61ARCHITECTUREOpen Platform-Vendor AgnosticUniform core components across all use casesScalable-Seamless application/use case integrationMulti-Domain simulationOntology based dataCybersecurityDiverse dataset integration

67、SingleDT User InterfaceVendor1:Proprietary AlgorithmVendor2:Proprietary AlgorithmVendor3:Proprietary AlgorithmSingleDT Data Ingestion/Management62Existing InitiativesDescriptionUtilityDT ConsortiumA team ofindividualsrepresentingutilitiesandorganizationsfromNorthAmerica,objectives/focusto create an

68、initialdefinitionofthedigitaltwintobeginconversationandrefineina largertask force(ITSLC)UtilityCIMThe Common InformationModel(CIM)is maintainedandupdatedbythe CIMusers group,underthe auspices of the UCA-IUGElectricPowerResearchInstitute(EPRI)Numerouseffortsaddressingprotectionrelaymodels andsimulati

69、ons,nuclearreactors,informationand communicationstechnology(ICT),variousgeneratormodels,geospatialdataandaugmentedreality,monitoring,andadvancedanalysis.SLAC National AcceleratorLaboratoryDatadrivenanalyticstomanageextremeweathereventsDynamicandflexiblepowerflowsimulationtool tosupportresilienceanal

70、ysis,tariffdesign,renewableenergy integration,etc.Methodologiesforforecastingelectrical loadonthegridOMGDT ConsortiumObjectManagementGroup(OMG)DigitalTwinConsortium:Has workinggroupsforvariousindustries(notably,utilityindustrymissing)Theconsortiumhaspublishedfourwhite papersondecarbonizationENTSO-E,

71、EU DSO DTEuropeaneffortto supportenergytransitionthrough“digitalizingtheenergysystem,”focusedonDT fortheEU powersystem.G-PSTConsortiumGlobalPowerSystemTransformationConsortium(G-PST),aconsortiumofpowersystems operatorsandpartnersaimingtoacceleratecleanenergytransitionsatscaleCIGRE WorkingGroupC4.64:

72、“ApplicationofReal-TimeDigitalSimulationinPowerSystems”CIGREWGC4.64Whitepaperdiscussingdigitaltwinsforpowersystem equipment:”DigitalTwinforPowerEquipment”IEEEPEST&D CommitteeIEEETFonDTs ofLarge-ScalePowerSystemsCommunityofresearchersandpractitionersfromelectricutilities,academia,industrialR&D,andsof

73、twarevendors toshareactivitiesontheconceptof DTinthepowersystemspace.12OPPORTUNITIES FOR INNOVATION13NetworkModelManagementCommon DT Framework Existing SolutionsIntegration ComputationModel Interoperability Use Case development Application developmentCONCLUSION14Big data will become a liabilityif we

74、dontcollect,store,and leverage it efficiently,effectively,and in a timely mannerGrid is rapidly increasing in complexity while ourtools are not changing and adapting at the samespeedCombininginsightsfromdata withtraditionalsimulation couldoptimizeour existing decision-making processes.Are our existi

75、ngassumptions for simulationaccurately reflecting reality?The DT of the future is a scalable,open,modular,systemof systemsplatformthatmaintains a high-fidelityreplica of the physicalgrid andthe context in which its assets exist.It brings together the necessary processes,technologies,and applications

76、 thatmaintain and leverage that digital replica to effectivelyoperate,maintain,plan,design,engineer,andmonitor the electric grid.Weneedinnovationin standards,technology,processes to make the DT framework a successTHANK YOUAbder Elandaloussi Abder.E15AI-EFFECT is supported by the European Unions Hori

77、zon Europe programme under agreement 101172952.Views and opinions expressed are however those of the author(s)only and do not necessarily reflect those of the European Union or European Climate,Infrastructure and Environment Executive Agency.Neither the European Union nor the granting authority can

78、be held responsible for them.Artificial Intelligence Experimentation Facility For the Energy sectorGianluca LipariTechnical Leader EPRI EuropeArtificial Intelligence and Digital Transformation Electric Power SummitPalo Alto,January 7-9,2025Current State of AI in the European Energy Sector08/01/2025A

79、I and DX Electric Power Summit 2025-AI-EFFECT Project 68Digital Revolution AI(Artificial Intelligence)/ML(Machine Learning)/LLM(Large Language Models)are making tremendous advances and proliferating in all aspects of life.Data-Rich EnvironmentUtilities have:Lots of data:structured,unstructured,time

80、series,images,documents,simulation results.Lots of use cases to be solved to automate processes and gain insights.Rapidly changing industry:regulations,markets,resources,assets,climate,behaviours.Early-Stage PrototypesSome limited successful applications of AI in utilities for narrow sets of use cas

81、es(mostly forecasting)and early stage LLM prototypes,mostly on document search and summarization.What are the Challenges to Successful AI implementation?08/01/2025AI and DX Electric Power Summit 2025-AI-EFFECT Project 69Upcoming EU AI Act RegulationsTrust Security and Barriers to DeploymentAccess to

82、 Data and Siloed DevelopmentEnergy utilities are high reliability organisations.The EU AI Act will likely classify them as system with a high-risk(Level 3 of 4)threshold for AI They will be subject to a more rigorous approach to AI safety,involving risk assessment,transparency,and importantly valida

83、tionBreakthroughs in AI are difficult to deploy for high reliability organizations.This is due to trust,security,compute resources.Software reliability requirements are onerous,favoring mature technologies and vendors for good reasons!Integrating new AI apps into business as usual processes is chall

84、engingDespite similar use cases and datasets,utilities work independently to develop solutions to use cases.New data science teams are proliferating but little outward coordination with industry partners.Researchers and vendors experiment and develop without real datasets and with limited understand

85、ing of the use casesAI-EFFECT project overview19 Partners from 10 EU countriesAI and DX Electric Power Summit 2025-AI-EFFECT Project 7008/01/2025The AI-EFFECT project aims to establish a European Testing and Experimentation Facility(TEF)for the energy sector to develop,test,and validate AI applicati

86、ons,enhancing energy infrastructure operations,resilience,and decarbonization efforts18 partners1 associate partner10 countriesOctober 2024September 2027January 2023December 20265,65 MProject total costEU contribution5.3 MAI-EFFECT motivationAI solutions cannot be deployed at scale in the energy sec

87、tor in silos.Utilities and industry must work to develop standardized frameworks and validation methodologies to enable new,disruptive technologies to be securely deployed.New Solutions to New Energy ProblemsSecurity,Transparency and ValidationAccess to Realistic Datasets for TrainingUnderstanding o

88、f Use Case,Utility NeedsWhat Utilities NeedNew issues are emerging on energy networks that require new ways of thinking and new solutionsWhat Industry NeedsTo build solutions,industry needs to understand utility needs and use casesAI and DX Electric Power Summit 2025-AI-EFFECT Project 7108/01/2025Ob

89、jectives01Develop sets of use cases of strategic importance to the EU energy sector and a testing methodology for each use case.04Develop the governance and business model for the enduring AI-EFFECT.05Engage with European and global energy sector stakeholders about AI-EFFECT to curate use cases,trai

90、ning,and test data and to match concepts and ideas to viable datasets.02Develop and implement a modular,interoperable,and scalable framework architecture03Develop and test the end-to-end AI-EFFECT solution,for interaction with the AI-EFFECT demonstration nodes and use cases.08/01/2025AI and DX Elect

91、ric Power Summit 2025-AI-EFFECT Project 72AI-EFFECT VISIONAI-EFFECT will establish a European Testing Experimentation Facility(TEF)to enable the development,testing and validation of(AI)applications,at various stages of development and maturity,for the energy sector.4 nodes initially,future nodes ca

92、n be easily added,in Europe and globally.Modular,ScalableFreely accessible to download data and upload models for testing.Benchmarks will be published if possibleOpen and FreeLinks research institutes,utilities,vendors across EuropeCollaborativeThe structure will be set up to meet the EU AI Act regu

93、latory requirements AI in Energy Regulatory CompliantNodes built for use case specific data and models.SMEs will form centers of excellence on use cases.Use Case CentricAI and DX Electric Power Summit 2025-AI-EFFECT Project 7308/01/2025AI-EFFECT federated architectureThe AI-EFFECT federated architec

94、ture will be accessible virtually via a web platform from anywhere in Europe.Users will be able to download training data and upload their models for test and validations.The testing facilities hosting the AI-EFFECT nodes will offer the needed data,models,interfaces and simulation capabilities to de

95、ploy the test cases and validate them.Decentralized and Distributed Nodes:Multiple demonstration nodes across Europe(Denmark,Netherlands,Portugal,Germany).Each node focuses on specific use cases like multi-energy systems,congestion management,energy efficiency,and DER integration.AI-EFFECT Digital P

96、latform:Ensures interoperability,scalability,secure data exchange,and trustworthy testing procedures.Virtual connection of existing European computer and lab facilitiesTwo-Layer Structure:Top Layer:Centralized platform with AI-EFFECT catalogue and functionalities.Bottom Layer:Distributed testing nod

97、es interconnected using the VILLASframework08/01/2025AI and DX Electric Power Summit 2025-AI-EFFECT Project 74AI-EFFECT Project-Danish Demo Site1.Focus Area:-Multi-Energy Systems2.Objectives:-Integrate AI to optimize the operation of multi-energy systems.-Enhance energy efficiency and flexibility.-I

98、mprove the integration of renewable energy sources.3.Key Activities:-Develop AI algorithms for real-time energy management.-Test and validate AI solutions in a real-world environment.-Collaborate with local energy operators and stakeholders.Danish Node District Heating and Sector Coupling08/01/2025A

99、I and DX Electric Power Summit 2025-AI-EFFECT Project 76AI-EFFECT Project-Dutch Demo Site1.Focus Area:-Transmission system Congestion Management2.Objectives:-Use AI to predict and manage grid congestion.-Enhance grid stability and reliability.-Support the integration of distributed energy resources(

100、DERs).3.Key Activities:-Develop predictive models for congestion management.-Implement AI-based solutions for dynamic grid control.-Engage with grid operators and regulatory bodies.Transmission System Congestion Management Dutch Node08/01/2025AI and DX Electric Power Summit 2025-AI-EFFECT Project 77

101、AI-EFFECT Project-Portuguese Demo Site1.Focus Area:-Energy Efficiency2.Objectives:-Apply AI to improve energy efficiency in buildings and industries.-Reduce energy consumption and costs.-Promote sustainable energy practices.3.Key Activities:-Develop AI tools for energy monitoring and optimization.-T

102、est AI solutions in various building types and industrial settings.-Collaborate with energy service companies and end-users.Energy Community Data Space Portuguese Node08/01/2025AI and DX Electric Power Summit 2025-AI-EFFECT Project 78AI-EFFECT Project-German Demo Site1.Focus Area:-DER Integration2.O

103、bjectives:-Integrate AI to manage and optimize distributed energy resources.-Enhance the resilience and flexibility of the energy grid.-Support the transition to a low-carbon energy system.3.Key Activities:-Develop AI algorithms for DER management.-Test and validate AI solutions in a distributed ene

104、rgy environment.-Work with local utilities and technology providers.DER Integration in Distribution System German Node08/01/2025AI and DX Electric Power Summit 2025-AI-EFFECT Project 79AI-EFFECT is supported by the European Unions Horizon Europe programme under agreement 101172952.Views and opinions

105、 expressed are however those of the author(s)only and do not necessarily reflect those of the European Union or European Climate,Infrastructure and Environment Executive Agency.Neither the European Union nor the granting authority can be held responsible for them.infoai-effect.euhttps:/ Lipari Project coordinator gianluca-lipari 2025 Electric Power Research Institute,Inc.All rights reserved.81 2025 Electric Power Research Institute,Inc.All rights TOGETHERSHAPING THE FUTURE OF ENERGY