A National Infrastructure for Artificial Intelligence on the Grid

OPEN 2018
El Segundo,
Project Term:
08/12/2019 - 08/11/2022

Critical Need:

The electric grid evolved in an environment of severe information scarcity. Its operation relies on overbuilt capacity, responsive fossil-fuel based generators with inherent energy storage (fuel and rotating mass), unidirectional distribution systems, and operator discretion in the face of uncertainty. An increasingly complex grid with diverse resources alongside fast, automated disturbance responses requires a new level of insight, without which the system can become uneconomical, unreliable, and unstable. For example, wide-area power oscillations never predicted by system models can occur, unknown amounts of generation behind customer meters can be suddenly lost, and protection systems can trip unexpectedly causing outages. In spite of the nationwide installation of phasor measurement units, sensor data use falls far short of its potential. The utility industry’s pace of innovation has been and continues to be hampered by a lack of data accessibility; ineffective data quality solutions; skill and tool mismatch; a lack of artificial intelligence (AI) and machine learning (ML) experts; and difficulty in transitioning research into production.

Project Innovation + Advantages:

PingThings will develop a national infrastructure for analytics and artificial intelligence (AI) on the power grid using a three-pronged approach. First, a scalable, cloud-based platform will store, process, analyze, and visualize grid sensor data. Second, massive open and accessible datasets will be created through (a) deploying grid sensors to capture wide-scale and localized grid behavior, (b) simulating and executing grid models to generate virtual sensor data, and (c) establishing a secure data exchange mechanism. Third, a diverse research community will be developed through focused educational content, online code sharing, and data and AI competitions. The project’s goal is to accelerate the development of data-driven use cases to improve grid operation and analysis.

Potential Impact:

This project will make research on DOE target impact areas easier to conduct by providing state-of-the-art tools and open access to necessary sensor datasets for analytic development and training of ML and Deep Learning (DL) models.


Development of a national infrastructure for AI will create a more secure power grid through situational awareness from grid data analytics, and help ensure the U.S. maintains its energy and AI technological lead.


Realizing data-driven applications using ML and DL to drive efficient use of grid assets will also yield a corresponding decrease in emissions, primarily through more effective algorithms for integrating and operating distributed energy resources.


Each component of the proposed AI infrastructure requires innovation, increases in value when integrated into the other components, and improves on current solutions. Access to data will drive AI/ML techniques within the power sector that will transform the way the grid is managed and enable a more efficient, economic, and reliable operation.


ARPA-E Program Director:
Dr. Patrick McGrath
Project Contact:
Mr. Sean Murphy
Press and General Inquiries Email:
Project Contact Email:

Related Projects

Release Date: