《确保电网人工智能在电力系统网络韧性方面的承诺.pdf》由会员分享,可在线阅读,更多相关《确保电网人工智能在电力系统网络韧性方面的承诺.pdf(11页珍藏版)》请在三个皮匠报告上搜索。
1、Securing the Grid:AIs Promise for Cyber Resilience in Power SystemsRichard MacwanManager,Cyber Resilience Science Group03/31/2025NREL|2NREL|2Power Grids:The Backbone of AI and SocietyUS electricity consumption to hit 4,239B kWh by 2026(AI/data centers drive 27%of demand)*Data centers could consume 8
2、%of global electricity by 2030(up from 3%)*https:/ Rising Cyber ThreatThe U.S.accounted for 47%of all global cyberattacks targeting utilities*Cyberattacks on the utilities sector grew by an alarming 234%year-over-year by Q3 2024*https:/ Credit:https:/ Resilience:Grand Challenges Cyber Resilience:Abi
3、lity to not just prevent and defend against cyber attacks,but to withstand,recover,and adapt.Rapid increase in the scale of data and devices beyond human cognitionRapid increase in the diversity of devices,systems,and users Complex cyber physical interactions complicate decision makingTrustworthy an
4、d timely decisions for safe and reliable operationsNREL|5NREL|5AI in Preparation:Building Fortified FoundationsPredictive Risk Modeling:Analyzes historical breaches to learn and explore new topologies for a more resilient networkGraph Recurrent Attention Networks learns and generates new and more re
5、silient network realizationsLarge Language Models to extract,formalize,and verify compliance and standards requirements at scale.LLMsStandardsStructured ComplianceAutomated Compliance:From subjective assessments to scalable,verifiable and automated complianceNREL|6NREL|6AI in Defense:Smarter Threat
6、Detection Detection with Cyber and Physical:Leverage combined cyber and physical system data for a more accurate anomaly detectionExplainable and Contextual Detections:Generate system interpretable and contextual explanations for actionable and trustworthy detectionNREL|7NREL|7AI in Withstanding Att