Coldwater Board of Public Utilities, a Michigan public power utility, has deployed artificial intelligence to predict and get ahead of grid outages.

In an interview with APPA, CBPU Utility Director Paul Jakubczak outlined how AI is allowing the utility prevent service outages, advance grid management, and support its lineworkers.

Similar to other public power utilities that are exploring new analytics platforms, CBPU has taken a measured approach to AI usage that seeks to integrate new capacities within its existing services in ways that improve resilience and reliability. 

Jakubczak began evaluating AI with his team due to its practical usage within structural monitoring. The utility opted to adopt eSmart Systems’ platform after testing its capacity with measuring the integrity of grid components.

“It was originally presented to us through eSmart. One of the things that they talked about was the ability to use AI technology to predict whether there is a pole top pin that's gone bad, if there is an insulator that’s not wired, or if the top of a pole is rotting, and similar issues,” Jakubczak said.

CBPU first began exploring AI for advanced grid monitoring in 2023 before launching a pilot in 2024. Rather than replacing the utility’s in-house expertise, CBPU’s artificial intelligence grid monitoring allows its staff to more easily detect structural issues in advance and address them in ways that mitigate repair costs and avert outages.

Much of this has been accomplished by CBPU’s deployment of a grid monitoring drone that captures aerial photographs used to evaluate potential failure points, which the utility then delegates for repairs or replacement.

Image of damaged pole captured by CPBU's aerial drone.
Drone-captured image of pole needing repair.
Photo courtesy Coldwater Board of Public Utilities.

“If you drive by a pole, you're not going be able to see what's on top. By using this technology and our drone, we're able to gather all these different images and then analyze them back at the office to decide what work orders need to be made,” Jakubczak said.

This has been especially conducive for customer-facing reliability, as the AI drone monitoring has allowed CBPU to notice potential component failures in advance of Coldwater residents needing to alert the utility.

“Something they did with drone footage is find that we had an open wire secondary that had a broken neutral wire, and you're not going to find that until you start using multi-phased equipment. We then sent our crew out there that afternoon and they repaired it. Without that, we wouldn't have known until the customer told us that there was a problem,” Jakubczak said.

These capacities have already shown considerable promise for streamlining repairs, especially when used within an integrated platform that notices issues in advance and prioritizes their resolution. This has led CBPU to sign an agreement with eSmart Systems where its platform is now integrated within the utility’s work order management. 

“We worked with them on generating a 5-year agreement where we'll continue to upload these images to them. And then their software will take pictures and start doing predictive maintenance by prioritizing what they’ve found,” Jakubczak said.

As Jakubczak noted, certain potential grid issues might previously “have sat there for 30 to 40 years” before causing service disruptions, with CPBU’s AI monitoring allowing a level of advanced detection that prevents structural wear from accumulating.

In addition to outage prevention, CBPU’s AI monitoring is allowing the utility to advance customer safety and mitigate potential grid-related hazards.

“It's also a safety thing, because you don't know what you don't know. But if we can detect and fix things that are a public hazard or a safety issue, then it’s useful from that angle too,” he said.

Beyond the safety and reliability applications, CBPU’s AI usage is amounting to a cost savings measure for the city of Coldwater, Mich.

“I think it's going to work out from a return on investment standpoint because you can see its lower cost compared to the cost of hiring someone to constantly do line inspections. Ultimately that’s what we do - affordable, reliable and accountable,” he said.

As Jakubczak summarized, the value of AI to a public power utility like CPBU is ultimately measured in how it allows them to advance grid reliability and service to their community.

“If this helps to eliminate even one outage, it's well worth that. Especially from a customer perspective,” Jakubczak said.