利用人工智能利用电网数据：分布式系统专家的见解.pdf

1、EPRI AI/DX Electric Power Summit|January 8,2025Leveraging AI to Utilize Grid Data:Insights from Distributed Systems Experts2Thanks for inviting us!Astrid AtkinsonCEO&Co-FounderDavid BaileyPrincipal Software EngineerWe help electric utilities tackle orchestration challengesContinuous Meter-Level Load

2、&PV ForecastingGrid Impact Simulationfor Residential EVsGrid-OptimizedManaged EV Charging3The Data Foundation for AITHE PREREQUISITES OF AIAI runs on data+scalable computingTo leverage AI in ways that are transformative,organizations require:1.A great data foundation2.(Hyper)scalable computing 5Ener

3、gy supplyCarbon intensity4.45MWMeter level:15mindemandvoltageDevice:1minLocation,typeActivityAvailable flexibilityTransformer&midline:15minload at any midline pointswitching statevoltageFeeder level:1secnet+actual demandgenerationavailable flexibilityvoltageWHATS POSSIBLE TODAYWhat does a great data

4、 landscape look like?6FEEDER ADATA POINTS&FREQUENCYEnergy supplyCarbon intensityMeter level:15min-60minSome/most metersHours-days collection delaySecondaries often not in GISDevice:some,with DERMSMultiple dashboardsMany sources missingTransformer&midline:periodicPeriodic engineering studiesMDM based

5、 analysis toolsBetter with hardwareFeeder level:1-5minnet+actual demandgenerationavailable flexibilityvoltageSolar-interconnection data onlyEV-program registration(maybe)Utility battery-SCADA telemetryCustomer battery-interconnection data,(sometimes)DERMSIncomplete GIS maps&switching stateWHERE WE A

6、RETodays utility data landscape7DATA POINTS&FREQUENCY4.45MWFEEDER ABRIDGING THE DATA GAPSUtilities can bridge data gaps to enable AI through 3 actions;additional hardware(e.g.sensors)can help but isnt necessaryEnable access to both OT&IT dataEstablish a standard data modelProvide a secure foundation