While we’ve seen a recent explosion in focus on artificial intelligence in the last few years, AI has been around and in development for a long time. Machine learning engines have been helping operators and technicians better perform their tasks for many years now. So what has changed to make everyone suddenly shift focus?
The answer is Generative AI, or GenAI for short, and it’s the topic of a recent interview Rick Kephart gave Public Utilities Fortnightly. In the interview, Rick helped illustrate how the rise of GenAI will reshape the power industry, and the benefits users can expect to see in the near future.
A new AI
The key difference in GenAI is what makes it truly transformative. Traditional machine learning tools were great for modeling, pattern recognition, and detecting anomalies as they could learn and differentiate between what was normal and abnormal. However, those tools couldn’t perform any tasks that resembled reasoning.
“We’ve been using aspects of AI in power for a long time, particularly things like machine learning, but they really didn’t have the ability to reason.”
The large language models (LLM) in GenAI have changed all that. Teams no longer need to explicitly program every rule and response. Instead, models are trained on all the information available using LLMs, and they use that data to generate new insights.
“Now with Generative AI, you can have the AI system reason over those results, and it can generate output—original content—and come back with a diagnosis, and you can interact with it in natural language.”
Fit for purpose tools
Most people, when they think of GenAI, think of the publicly available LLMs like ChatGPT. However, these are not the models that will drive the future of the power industry. The largest software vendors are already offering private LLMs that do not release information back into the public domain, helping keep intellectual property and operations private, while still empowering teams to leverage the power of AI.
Moreover, a generic model is typically of limited value, so private LLMs can be modified, either through training or through external datasets including retrieval-augmented generation AI engines—tools that use a specified, customized data set to help refine analysis for a company’s unique operations.
The big picture for power
There will be many benefits to GenAI for power companies. First and foremost, teams will be able to keep plants running more efficiently, effectively, reliably, and safely through better predictive maintenance and improved operations. GenAI will help teams detect failures earlier and can provide decision support to operators and technicians.
The larger grid will also be impacted, as powerful AI engines will help teams navigate the intermittency of renewable generation and the complexity of weather patterns and events. AI tools will augment the human operator to increase situational awareness and improve the ability to react to a changing environment or market factors.
In addition, LLMs will help companies lock-in institutional knowledge. Best practices and insights will no longer leave the company when the best personnel retire or move into new roles, as they can be taught to AI engines to help support and upskill new personnel across the life of the facility.
Rick shares even more insight in the full interview, including specific examples of AI in action. Head over to Public Utilities Fortnightly to get a glimpse into what might be the future of your plant!
