Opinion
Sponsored
AI
Grid edge

Utilities can’t risk ignoring artificial intelligence

The most perilous path forward is in failing to test AI’s ability to improve grid operations, particularly at the grid edge.

Listen to the episode on:
Apple Podcast LogoSpotify Logo

Photo credit: Jon G. Fuller / VW Pics / Universal Images Group via Getty Images

Photo credit: Jon G. Fuller / VW Pics / Universal Images Group via Getty Images

The increasing complexity of the electric grid, combined with ambitious clean energy goals, demands innovation. While there is a much-needed emphasis on efforts to expand transmission and large-scale infrastructure, it is the meter that is the critical nexus between utilities and their customers.

As the first generation of “smart” meters are due for a refresh, now is the time to consider what capabilities and policies are needed to meet the climate and clean energy demands of today — and of the next two decades. 

Artificial intelligence has already transformed countless industries, from automotive to healthcare. The electric grid, though, is one of the most complicated machines ever built and has yet to harness this technology. As other industries race ahead, it's imperative that the utility industry does not stay anchored to outdated technology. The riskiest path forward is in failing to test AI’s ability to improve grid planning and operations, particularly at the grid edge. 

But traditional meter technology was largely designed to ensure reliable and accurate billing, not real-time grid operations or AI.

Even the newest smart meters lack the computational power needed to run AI models to tackle emerging use cases, like distributed energy resource integration. In the coming years, more computing power at the grid edge is necessary for DERs to provide grid services beyond capacity and energy, support local distribution constraints, and improve reliability.

The Department of Energy recently highlighted the need for AI in a report underscoring the potential for it to “significantly improve” planning, permitting, operations, reliability, and resilience. The report specifically notes that distributed AI at the grid edge can “improve grid resilience and grid cybersecurity by reducing the number of data connections required for grid operations,” as well as “process raw data and even make decisions locally.”

In combination with human oversight and rigorously validated systems to avoid introducing new security risks, all signs point to the need to test and scale AI at the grid edge. 

Distributed AI offers a solution to grid complexity 

While utilities are understandably cautious, failing to test innovative technology that can significantly and continuously improve everything from permitting to reliability is a recipe for stranded assets and compounding challenges in the years ahead.

Distributed AI uses accelerated computing at the source of data — in this case, the meter — and learns from new data to improve over time and help people make better and faster decisions. Using these tools, utilities can process and act upon large amounts of complex data in real time to manage the exponential complexity driven by electrification, electric vehicles, and mounting reliability challenges. 

With access to instant, granular data analysis, utilities can significantly improve programs targeting clean energy parity, affordability, and reliability. And they can do so without the vast majority of the data ever leaving the edge, which benefits data security.

Listen to the episode on:
Apple Podcast LogoSpotify Logo
No items found.
No items found.
No items found.
No items found.
Get in-depth coverage of the energy transition with Latitude Media newsletters

Utilities can de-risk metering and other grid edge investments

States across the country are already experimenting with how to incorporate AI into their next round of grid modernization and metering investments. Here are three things that are working: 

Competitive solicitations that include AI

In Connecticut’s grid modernization docket, regulators directed utilities to submit new final advanced metering infrastructure plans that include competitive solicitations from technology companies with “distributed intelligence” use cases.

Regulators made the point that providers must submit bids that, to the greatest degree possible, “demonstrate how the meters can be future-proofed” and “enable broad competition by allowing vendors to apply for separate components of the meter solicitation.” This type of engagement from regulators can advance utilities’ ability to seek out advanced AI technologies. 

Identifying technology that is future-proofed

Southern California Edison’s most recent rate case is a prime example of a utility identifying advanced AI capabilities and functionality to future-proof its AMI 2.0 solicitation.

SCE noted that capabilities such as high compute power, onboard sensors, edge analytics inside the meter (so-called distributed intelligence), and more advanced techniques including AI and machine learning algorithms will be needed “to support their analytics-based decision making.” Of those, high computation and processing power is particularly important to ensure that there is enough headroom to add more functionalities and applications, including from third parties, to 15-20-year metering investments. 

Prioritizing systems that are interoperable

SCE noted in the same rate case that within North America, most of the commercially available AMI solutions are proprietary and locked in with a single vendor to provide a complete end-to-end solution.

Outside of North America, though, utilities have made a concerted effort to develop interoperable and interchangeable AMI solutions that “promote multi-vendor, modular, and plug-and-play ecosystems.” SCE says that as it shifts towards a “digitized, decarbonized, and decentralized grid” that interoperability will be key. 

It’s crunch time

Utilities are facing critical investment decisions as they respond to electricity demand increases, at the very same time that AI is transforming industries across the world. Utilities need to invest in technologies that will provide value for the next two decades, particularly at the grid edge where things are most dynamic. 

Regulators and utilities can derisk these investments by testing how distributed AI can transform their operations, and then opening the door to competitive solicitations that require advanced interoperable solutions.

The status quo will not future-proof a grid that is changing rapidly. The status quo is a risk to both utilities and their customers in the diligent work to provide reliable clean energy. Holistic grid modernization is needed, and distributed AI is an investment that becomes more valuable over time — both to utilities and to the people who rely upon the grid.

Lauren Randall is the vice-president of policy and market development at Utilidata. The opinions represented in this contributed article are solely those of the author, and do not reflect the views of Latitude Media or any of its staff.

No items found.