5 Artificial Intelligence Grand Challenges for the Electric Power Industry

Posted to EPRI in the Digital Utility Group
image credit: EPRI
Jeremy Renshaw's picture
Senior Program Manager, Electric Power Research Institute (EPRI)

Dr. Jeremy Renshaw is the Sr. Program Manager for Artificial Intelligence at the Electric Power Research Institute (EPRI) and has been with EPRI since 2012. Dr. Renshaw manages the AI.EPRI...

  • Member since 2021
  • 13 items added with 12,117 views
  • Sep 22, 2021


The Artificial Intelligence (AI) and electric power industries are two large, impactful industry sectors that can both benefit greatly from interactions and collaborations but have not traditionally interacted significantly – both due to a lack of sufficient industry data as well as the electric power industry’s need for reliability, stability, and safety to deploy AI on physical assets.  Both industries have progressed rapidly in recent years with AI techniques rapidly being benchmarked and deployed in other industries as well as digital transformation efforts in the electric power industry underway that are collecting vast tomes of data.  EPRI is helping to accelerate these technologies by issuing 5 Grand Challenges to the AI and Electric Power industries to facilitate collaborations to benchmark, enhance, and accelerate adoption of these technologies across the electric power industry.

For many years, individual organizations have developed and implemented artificial intelligence (AI) and machine learning (ML) technologies for specific use case applications.  These models typically require significant amounts of data for training and evaluation purposes as well as significant computing resources and solve individual problems.  Computing costs have continued to drop while digital transformation activities at utilities have drastically increased the amount of available data, creating an optimal point in time to accelerate efforts related to AI adoption across the globe.

EPRI and Stanford University co-hosted a series of meetings in 2021, bringing together >100 different utilities, universities, and AI organizations to identify opportunities to apply AI/ML techniques to address industry issues, improve business operations, enhance overall safety and efficiency, etc.  These events involved an Executive Panel discussion, a training session and expert panel on AI, and culminated in a Reverse Pitch event where utilities and AI organizations met to share challenges and potential solutions.  Through these events, common themes were identified and collected into a set of grand challenges and key enablers for the AI and electric power industries.

AI Grand Challenges for the Electric Power Industry

  1. Grid-Interactive Smart Communities – The energy system of the future will connect owners and operators of buildings, homes, and power grids sharing the benefits from the advances in AI that could improve building-to-grid-operator communication, optimize cost, and improve energy utilization and energy equity for producers and consumers.  This grand challenge seeks to develop, benchmark, and scale the adoption of AI technologies that support networks of homes and buildings that interact with the power grid to optimize energy efficiency, load shifting, and usage of low or zero-carbon generation sources for economy-wide decarbonization in equitable ways for the entire community. 

  2. Energy System Resilience – Aging infrastructure, combined with severe weather events and climate change are expected to affect every aspect of the electricity sector - from generation, transmission, and distribution, to demand for electricity. Catastrophic events such as the 2021 Texas winter storm event severely disrupt the normal functioning of critical electrical grid infrastructure for significant durations.  This grand challenge seeks to develop, benchmark, and scale the adoption of AI technologies that can help to predict weather, demands, generation plant and grid conditions, and continuously optimize the system to minimize unplanned outages and intelligently control energy flow to minimize or eliminate the impact of such events in the future and reduce unplanned outage durations.

  3. Environmental Impacts – AI technologies can help power grid operators to better predict loads, optimize usage of low-and zero-carbon generation sources, improve efficiencies, and lower emissions.  AI/ML can also help identify and lessen wildfire risks, improve vegetation management, and reduce impacts on wetlands and endangered species.  This grand challenge seeks to utilize AI technologies to further reduce carbon emissions and minimize environmental impacts.

  4. Intelligent and Autonomous Plants – With the growing need for flexibility in large-scale power plants and optimizing their interaction with resources on the grid, controls and automation have become increasingly important in the utility industry. Automating tasks using AI helps reduce costs, improve efficiency, and preserve energy system assets through optimized maintenance and utilization. AI applications such as digital twins, ML/Reinforcement learning, machine vision, and automatic diagnostics, among others will enable energy system operators to focus on the most valuable maintenance, asset management, and integration tasks.  This grand challenge seeks to develop, benchmark, and scale AI applications for automation.

  5. AI-Enhanced Cybersecurity – Cybersecurity is foundational to energy systems and operations, protecting critical utility data, such as personally identifiable information, critical operational data, operational technology systems, and data for AI models.  The current and future energy generation and distribution systems rely on an increasingly digital, interconnected landscape.  AI shows promise to enhance cybersecurity improving capabilities like network monitoring, identifying suspect activity, and automatically detecting vulnerabilities in software codes.  This grand challenge seeks to advance the state-of-the-art in cybersecurity practices through effective implementation of AI solutions, benchmark those identified solutions, and scale them across the electric power industry.

Key Enabler #1:  Industrywide Data Sharing and Governance – Larger, more robust datasets will be needed to scale AI applications, frequently larger than any one utility has at its disposal.  For this reason, sharing data across industry to identify, collect, and curate key sets of data will be critical to these efforts. Data must be collected, labeled, anonymized, and stored in a secure fashion with the proper data management infrastructure. Continued efforts must be made among electric power utilities to address the data sharing challenge, as industrywide data sharing and governance will enable critical AI technology adoption.

Key Enabler #2:  Data Science Expertise and Training – There is a talent shortage for data science and AI expertise in the industry.  EPRI is working to augment the existing data science skills of staff by accelerating training efforts, including hosting training events, providing links to publicly available materials, and via a co-funded effort with the U.S. Department of Energy. All these efforts, and many more, will be needed to develop the necessary skills to match AI talent with electric power experts to scale up adoption of AI throughout the electric power sector.


Five grand challenges have been identified and are being issued to the AI and electric power industries to accelerate technology development and deployment for near term and long-term benefits to both industries. Successful implementation of these grand challenges and key enablers will provide substantial benefit to both industries both now and in the future in terms of increased safety, reliability, resiliency, reduced environmental impact, and improved economics.  These benefits will improve the quality of living for people around the world through reduced emissions, improved reliability of electricity, and lower energy costs.

How to Get Involved

Learn more about these Grand Challenges and how to get involved by attending and participating in the grand challenge breakout sessions at EPRI’s AI and Electric Power Summit on September 28-29. Register here.

Founded in 1972, EPRI is the world's preeminent independent, non-profit energy research and development organization, with offices around the world.
Matt Chester's picture
Matt Chester on Sep 22, 2021

 Catastrophic events such as the 2021 Texas winter storm event severely disrupt the normal functioning of critical electrical grid infrastructure for significant durations.  This grand challenge seeks to develop, benchmark, and scale the adoption of AI technologies that can help to predict weather, demands, generation plant and grid conditions, and continuously optimize the system to minimize unplanned outages and intelligently control energy flow to minimize or eliminate the impact of such events in the future and reduce unplanned outage durations.

I'm curious if you could talk to what hypothetically AI could have done to dampen the impact of the Texas winter storm grid issues-- you mention predicting weather but I don't think the issue was that the storm wasn't seen coming in the days ahead, just that the grid and markets were unprepared. Would AI have been able to isolate some of the problem areas so as to reduce the spread of the impact? Jump in earlier to more intelligently distribute the available energy that wasn't taken offline? 

Jeremy Renshaw's picture
Jeremy Renshaw on Sep 22, 2021

Good question, Matt.  I would say that the goal of the Resiliency grand challenge isn't to avoid any and all issues, but rather help to prepare and respond for them in an improved fashion.  A few things that AI can help to do would be to help:


1. Pre-emptively identify at risk or degraded components and dead/dying trees in the area of T&D assets, improve weather and grid load forecasting, etc. to help prepare for events.

2. Improve grid/load management issues during an event to minimize the impacted area and avoid potential cascading blackouts.

3. Aid in recovery by helping to isolate failures, downed trees, etc. to help accelerate utility efforts to get the lights back on.


Not saying that the Texas event (and other similar events) would not have happened, but the impact and duration could have been notably reduced.

Larry Kauf's picture
Larry Kauf on Sep 22, 2021

In Texas, there is a basic underlying flaw in the whole ERCOT system.  This flaw cannot simply be fixed with AI.  Unlike any other Utility in the country (and ERCOT is not a Utility but a Grid operator) Generators are not incentivized to to do anything more, than generate the maximum amount of power their plant will produce.  If they can cut costs in doing that...even better.  Build a new power plant?  Why? Who is going to pay for it? Weatherize the Pipelines supplying the Gas to the Plants?  Who's gonna pay for it?  Something as basic and simple as putting in the right Lubrication for cold weather conditions, in a Wind Nacelle, is scoffed at because a cold weather event like last winter,  only happens once every 10 years. 

There is no Capacity market behind ERCOT. You get what you get and that's backup.  There are few interconnections to other States or Grids to "borrow" power when excess is needed.  Just look at August 2020.  Certainly not a Cold Weather event,  but sustained 100 degree temps and the Power Generators could not keep up with the demand.  People wanted to be cool, they wanted to to go home cook dinner, plug in their EV's and sit by the pool.  The ERCOT solution to that was to raise the price per Kwh till the demand would naturally correct itself since people wouldn't want to be paying huge amounts for electric, during peak demand.  Only problem with that is the system is NOT transparent, so how would anyone know that prices are spiking?  Thus we saw $900/Kwh power prices and only because it was capped by ERCOT.  The system didn't work (such that it is) and we saw rolling blackouts from Dallas to Corpus.

When the average Joe, got their bills there was a revolt and no one wanted to pay those exorbitant prices.

Artificial Intelligence cant cure stupid.

Jeremy Renshaw's picture
Jeremy Renshaw on Sep 22, 2021

Larry, thanks for the comment and I agree that each event has its own unique causal factors.  I think my response to Matt's comment covers most of what I wanted to say.  The only thing I would add is the cost impact.  If low-cost tools and systems with a broad range of capabilities can be developed, then it makes sense for a business to adopt them.  That is part of the ultimate goal here.  Certainly, economics drives a big piece of adoption, so the cost vs. benefits issue needs to be addressed.  Thanks for the response!

Jeremy Renshaw's picture
Thank Jeremy for the Post!
Energy Central contributors share their experience and insights for the benefit of other Members (like you). Please show them your appreciation by leaving a comment, 'liking' this post, or following this Member.
More posts from this member

Get Published - Build a Following

The Energy Central Power Industry Network is based on one core idea - power industry professionals helping each other and advancing the industry by sharing and learning from each other.

If you have an experience or insight to share or have learned something from a conference or seminar, your peers and colleagues on Energy Central want to hear about it. It's also easy to share a link to an article you've liked or an industry resource that you think would be helpful.

                 Learn more about posting on Energy Central »