Setting New Standards in Utility Resilience: Predictive Maintenance Leads the Way

Predictive maintenance is increasingly becoming the cornerstone of the utility industry’s strategy to combat the escalating threats posed by climate change and natural disasters. The Grid Resilience and Innovation Partnerships (GRIP) Program, backed by the US Department of Energy, marks a pivotal move toward reinforcing wildfire prevention and bolstering energy resilience across the United States. This strategic pivot from a reactive stance to preventative measures underscores the industry’s commitment to safeguarding the electric grid.

In this landscape, various innovative companies are deploying data-driven tools and methodologies, empowering utilities to assess and manage the health of their assets more accurately. This approach significantly curtails the risk of catastrophic wildfires, showcasing a decisive step towards protecting essential services amid growing climate volatility. Predictive maintenance plays a crucial role in enhancing grid monitoring and security efforts by leveraging technological advancements.

At the heart of these initiatives is a shift in mindset and strategy within the utility sector, exemplified by San Diego Gas & Electric (SDG&E), which has embraced predictive modeling and advanced technologies to not only enhance the reliability of electric inspections and maintenance but also to ensure safety and readiness against the challenges posed by climate change and electrification demands. This transition embodies a critical evolution from a ‘run to failure’ model to one that anticipates and mitigates risks before they escalate into emergencies.

Capturing the essence of this transformation on the show floor of DISTRIBUTECH 2024, MarketScale had the opportunity to engage with Jennifer Kaminsky, the Manager of Electric Assets and Compliance at SDG&E. Her insights shed light on the strategic and technological drivers propelling this shift towards predictive maintenance in the utility industry, offering invaluable perspectives on safeguarding communities and infrastructure against the unpredictable challenges of the future.

Jennifer’s Thoughts

“In my role at SDG&E, I’m responsible for our distribution electric inspections, as well as our corrective maintenance work. I think the utility industry traditionally had a run-to-failure model, where we just responded to events that happened. And with the space that we’re in, with wildfires, climate adaptation, and a lot of other challenges facing the industry, we’re really needing to be more predictive and proactive about how we identify damages, about making sure we have good asset inventory in order to be more predictive.

And so the presentation we did is about what at SDG&E we’ve been trying to accomplish in trying to get to that space, right? Trying to go from this very responsive, corrective approach to being more predictive and proactive about our inspections. Instead of just responding and repairing in-kind, and polka dotting around the map with, this is an old pole, this one broke, so we’re going to fix that one and replace that one.

We really need to be more thoughtful about how we spend those responsive dollars and put it more into getting ready for electrification, right? How we’re going to get ready for customers that are going to be using total electric to change these responsive dollars into that electrification, upgrading transformers, changing conductor size.

You’re not going to be ready for that when it happens and everybody’s got an electric car, batteries, you know, and there’s more things to plug in than ever before. And then just going to just safety, right? We have to be ready for the next big catastrophic wildfire and make sure that it isn’t caused by a utility. We also need to make sure that we can get power back on very quickly for our customers.

Intelligent image processing, what it does, you know, imagine yourself, you’re a lineman, you’re an inspector, you’re going out to a pole and instead of giving you a hundred questions, I mean, what type of cross-arm is on here? What type of transformer? Are there used fuses? You’re answering all these questions and you have to do that 50 times a day. It can get very repetitive.

Intelligent image processing, I like to call it spell check for linemen, it kind of gives that quality check and helps to build up our asset inventory. If you don’t know what’s out there, ou can’t necessarily be predictive about what’s going to fail and where the biggest risk is, right? So one is just asset identification, including those minor units of property.

Having qualified inspectors, having qualified linemen, do we want our linemen to be spending their time looking at a pole, finding something wrong, or do we want them spending time fixing things? So, I’m trying to shift that workforce into the work that they train for and get them back into the fixing space, rather than looking at something and saying, yep, that’s good, yep, that’s good. 

Machine learning models, if it can help surface those potential impacts and find qualified inspectors or linemen, we only have to look at those photos, it saves us time and resources and allows us to get from I found something to I’m going to fix it that much quicker.”

Article written by Sonia Gossai