Artificial Intelligence Disrupts the Energy Industry

AI chatbots such as OpenAI’s ChatGPT have garnered attention for their exponential ability to address commercial prompts, presenting a futuristic view of the technology’s potential within the energy industry. Yet AI has already been disrupting the industry for years, with established and field-proven applications in use across the electric grid. As the power system becomes increasingly interconnected, innovative use cases for advanced prediction, control, and optimization—driven by AI—are expanding in the wake of regulatory and technological market shifts to accelerate the energy transition.

Regulatory Drivers

A new regulatory regime, supported by AI, is emerging to manage an influx of grid resources. Order 2222 from the Federal Energy Regulatory Commission, or FERC, signals that new methods of aggregating distributed energy resources (DER), such as virtual power plants, will soon take hold across the grid. Incentivized by the Inflation Reduction Act, management algorithms powered by predictive AI are leveraging emerging DER to analyze historical power trends and flatten midday peaks previously filled by diesel and gas generators, all without costly infrastructure upgrades. As customers increasingly adopt DER, complex AI is becoming a necessity for optimizing an already vast array of connected resources.

Utility-Scale Flexibility

Uniting residential-scale DER enhances grid resilience and reliability. Coupled with AI, the growing availability of solar, wind, and storage technologies is allowing resources to be combined into utility-scale flexibility schemes. For example, the Brooklyn Microgrid, customized by LO3 Energy, leverages blockchain to operate a peer-to-peer platform that enables algorithmic routing of power based on local community demand. Intelligent trading and predictive management let customers procure cheaper and local-certified renewably generated resources on demand through the distributed system operator. Outage time and frequency are reduced by ensuring continuous adaptation to market conditions, enhancing demand curve efficiencies using machine learning.

Grid-Edge AI

Leveraging data amplification from a proliferation of DER, AI is providing split-second grid-edge decision-making across a spectrum of use cases to maximize energy resource efficiency, optimize customer experience, and reduce data siloes.

• NVIDIA and Utilidata’s collaborative AI-powered smart grid chip incorporates real-time, inside-the-meter power optimization using machine learning to better understand customer patterns, detect anomalies, and address inconsistencies across deployments.

• Combining edge computing and waveform data processing technologies, Landis+Gyr’s next-generation Revelo meters enable real-time AI pattern recognition in energy delivery. The ability to capture not just sub-second data streams but high resolution waveform data is a key differentiating feature, contributing to Landis+Gyr and its analytics providers’ recent success in New York State.

• In October 2014, Itron released Riva, its branded open Distributed Intelligence (DI) platform, supporting advanced applications for meter bypass, high impedance, residential neutral fault detection, location and solar awareness, transformer load management, active premise load shedding, and more. Over 4 million Itron Riva DI meters have been deployed to date, with nearly as many licensed applications in use.

Opportunities and Risks

Already, AI is accelerating the energy transition through deep learning and management to drive network efficiencies. As large datasets with sector-specific parameters emerge, energy sector work beyond grid management may be powered by AI—for instance, a chatbot system may address utility staffing shortfalls by helping lineworkers troubleshoot unfamiliar and potentially dangerous field scenarios. However, automated systems could face resource allocation dilemmas when, for example, selecting which areas of the grid to prioritize for power restoration. To gain acceptance, especially in critical industries, AI datasets will need to be truly unbiased and explainable by accountable parties or else risk reputational damage.

To learn more about AI’s impact on transmission and distribution (T&D) networks and DER, check out the Guidehouse Insights reports AI for Predictive T&D Network Management and Guidehouse Insights Leaderboard: AI Vendors for DER Integration.