Advanced Technologies are Redefining the Power Grid

Last November T&D World celebrated its 75th anniversary by looking at where the power grid has been and anticipating where it’s going. “Charging Ahead” was focused on how smart grid technology is reshaping the power grid and its operations. It has become an interconnected ecosystem of intelligent devices that is synonymous with modernization and optimization. Unfortunately, this modernization has increased the power grid’s complexity, but that’s not unexpected. Newer smart grid adaptations are addressing that problem with user-friendly interfaces.

Making smart grid technologies more user-friendly keeps the its usage increasing. A recent report from MarketsandMarkets said the global smart grid’s market is expected to expand to an estimated USD 161.1 billion by 2029. This is up from a projected USD 73.8 billion spending in 2024. Those figures represent how valuable digital technologies have become to utilities and operators, but there are adaptations that combine smart grid with other advanced technologies that are wanting more discussion.

The “Charging Ahead” article touched on several of these advanced technologies, but wasn’t able to go far enough into their cutting-edge applications or how those schemes are advancing the power grid. One of those mentioned was artificial intelligent (AI) and several of the applications associated with AI. Integrating AI into the smart grid requires care to details. Processing big-data into real-time awareness of the power grid can be tricky, as can AI’s pattern recognition capability. Combining these two for forecasting and anticipating fluctuations is only one aspect of the challenge, but as data big-grows in size it gets harder.

Size Doesn't Matter

Smart grid data analytics are helping sort through this explosion of big-data, but how much is really coming from the power grid? It’s hard to nail down the specific amounts of big-data attributed directly to the smart grid itself, but the numbers are significant. It’s been said the daily big-data production rates range from terabytes to petabytes, but with numbers of this magnitude, size doesn’t matter. It’s sufficient to say the amount of big-data is beyond the ability of a human to manage, so big-data analysis is needed.

There have been many processes developed for handling big-data, but our interest lies within an application reported on by Dimension Market Research. They published a study that said, “the smart grid data analytics market was expanding.” The study projected the market would reach USD 8.2 billion by the end of 2024. The publication went on to say that the smart grid data analytics market was anticipated to reach USD 24.2 billion by 2033, which brings up a question. What is smart grid analytics?

Essentially, smart grid data analytics is a shorter way of saying, turning the big-data into actionable information using AI. It’s a complicated subject, but there are many websites that go into great detail about AI if you are interested. For our discussion, keep it simple and look at how AI is being utilized by utilities and grid operators. Still, it’s important to know a few basics concerning AI.

It's Really Machine Learning

It should be recognized that what is being called AI is really machine learning designed to automate manual tasks. Sentient AI systems seen in works of science fiction does not exist yet. Machine learning analyzes big-data using algorithms and powerful cloud-based computing. These algorithms vary in sophistication based on the performance expected from the AI application, with the more complex reserved for the most complicated tasks.

The accepted labeling practice places all the various forms of machine learning together under the common name AI. Keeping with convention, AI is ideal for classifying, assembling, and managing big-data. It uses algorithms to organize data into well-defined categories. In addition, AI uses sophisticated sets of conditional algorithms to make predictions based on the categories and conditional probabilities.

Last year, Siemens Energy launched its Gridscale X platform, which is part of the Xcelerator portfolio. This AI-augmented platform is paving the way toward the vision of autonomous grid management, according to Siemens Energy. They went on saying it will accelerate the digital transformation in the planning, operations, and maintenance areas. Moreover it will support utilities tackling their most pressing challenges related to the energy transition and keep their grid stable and reliable.

It's About Performance

Breaking with the above labeling practice, there is another form of AI that has been placed in its own grouping called generative AI or GenAI for short. GenAI has established itself as a separate application due to its exceptional ability. GenAI uses algorithms that are somewhat different from typical AI. They take advantage of probabilistic techniques to generate what-if scenarios, which gives it the ability for extrapolation of its big-data.

To do this, GenAI analyzes the big-data to understand its underlying patterns and it creates entirely new content from the big-data it studied. This makes GenAI a bit controversial, but a very powerful tool. GenAI shows great promise for forecasting energy consumption, anticipating how loads might vary throughout the day, and envisioning congestion spots on the power grid to name a few undertakings.

In November 2024, PG&E initiated the first commercial deployment of a GenAI application for a nuclear power plant at their Diablo Canyon nuclear generating plant. The application, Neutron Enterprise Program, is a search engine developed by Atomic Canyon and the Department of Energy’s (DOE) Oak Ridge National Laboratory (ORNL). Neutron runs on the NVIDIA AI platform. It uses Atomic Canyon’s FERMI family of AI models developed in collaboration with ORNL.

PG&E said that “federal and state regulations require utilities operating nuclear power plants to manage billions of pages of technical documentation spread across multiple systems.” Finding a specific record can be a slow process consuming time and resources. Neutron uses advanced optical character recognition, and the GenAI search engine has been taught to understand the terminology used in nuclear projects, which speeds up searches from hours to seconds. Being open source software, it will be available to the industry.

Changing the Power Grid

Speaking of DOE, last April DOE announced a big step for AI grid applications with their VoltAIc Initiative. They said it’s designed to use AI to help streamline siting and permitting at the Federal, state, and local levels. DOE is investing US$13 million to develop AI-powered tools for improving the process for new grid sites. It will also aid in permitting new clean energy infrastructure. While that was taking place, DOE published its first major report on AI titled “AI for Energy,” which was followed by several others.

The publications make interesting reading and are available online at DOE’s website. They emphasize the growing interest in the advantages of advancements in AI capabilities across the energy industry. They also point out that various forms of AI are already in use worldwide and technological advances are driving the rapid expansion of these AI tools.

A quick search on the web for utilities using AI applications on their electric grid brings back a slew of hits. One described using powerful high-resolution cameras teamed with AI applications as a formidable tool for fighting wildfires. The idea is to identify the fire while it is small enough to extinguish. Utilities like PG&E, Portland General Electric, Xcel Energy, Southern California Edison, Hawaiian Electric are using the AI technology effectively and reducing wildfires in their territories.

Another one talked about labor intensive time consuming power line inspections. By using autonomous drones equipped with high-resolution cameras and integrated with AI and GIS for vegetation assessments, maintenance patrols, etc. utilities are seeing savings while efficiently meeting regulatory requirements. Utilities such as New York Power Authority. Dominion Energy, ComEd, American Electric Power, Southern Company, and others are utilizing these advanced technologies successfully.

Who Said It's Easy

These are only a small percentage of what’s happening on the power delivery system as a result of AI and GenAI being integrated into the power grid. So far AI has been doing the heavy lifting, but GenAI is beginning to pick up the slack when it comes to things like load balancing with real-time grid awareness.

Others see AI’s predictive analytics making the power grid more resilient by foretelling asset health and points of failures. If the DOE initiative for using AI to streamlining siting and permitting, delays for both will be lessen. It’s all about modernization when you remove the distractions of rapidly emerging technologies and consumer demands.

Granted, there are going to be bumps in the road as this all plays out. But it’s a positive step in dealing with the challenges the 21st century power grid is placing on the utilities and operators of the grid. There’s no option of doing nothing despite people ignoring events. The industry is just starting to begin the adoption of AI and its subsets. We can expect these advancing technologies to redefine the industry in ways we haven’t anticipated. No matter how bizarre an AI application seems someone will turn the idea into reality!