From decarbonization to resiliency, utilities need to be agile, innovative and connected to meet the demands of the future.
The most-forward thinking utilities are embracing the power of AI and big data to chart the most efficient and effective path to achieve future readiness.
Bidgely has been with Avista Utilities on its AI journey from the start. Providing electricity to nearly 403,000 customers and natural gas to ~369,000 customers across 30,000 square miles in eastern Washington, northern Idaho, and parts of southern and eastern Oregon, Avista is executing on multiple big data and AI use cases across functional areas, deploying them successfully, and scaling them across the organization for maximum grid benefit.
I had the opportunity to talk with Andrew Barrington, Products and Services Manager at Avista, to get his perspective on the utility’s journey so far, and how he sees big data continuing to transform their operations.
Andrew says Bidgely’s UtilityAI™ Analytics Workbench tool has acted as a “use case generator” for the utility. As departments share information and data-driven discoveries, it inspires new questions and prompts the rethinking of legacy problems with new approaches.
“It’s not just a single individual or a single department. It has to be cross-functional. It has to be a conversation amongst your utility,” he notes. “I’m proud to say that Avista has achieved that. We have a lot of awareness around big data. We have a functional group of about 15 users, and we meet with Bidgely every other week to come up with new use cases and talk through how we can take the data we have and elevate it further. We are able to continuously ask questions like, ‘Can we track what the potential growth is?’ or ‘Could we learn a little bit more about where this trend is going?”
EV Intelligence
In the past, Avista used in-house methods to create EV charging load profiles, which were informative but had limitations. These methods were restricted to known EV customers in utility programs, sometimes lacked accuracy, and demanded substantial engineering and analysis work.
I asked Andrew to share his view as to how the advanced features in BIdgely’s UtilityAI™ Analytics Workbench tool have enabled Avista to produce robust load profiles at scale, and gain deeper insights into customer charging habits, feasible load-shifting strategies, and smarter grid infrastructure planning for future EV demands.
“In 2021 we worked with Bidgely to analyze our AMI data to detect just over 600 accounts. But now we can see our growth over time to 2022 and 2023. In just three short years, we saw a growth of close to 1,300-1,500 vehicles come into our service territory. That’s fantastic insight for us to understand our EV growth. And what’s really interesting is each year EV time-of-use shift has shifted. The time of use when we started is much different than where we’re at today.”
Beyond total EV load, Avista is able to identify feeders and transformers that could come under strain due to EV proliferation.
“We have been able to leverage our meter data a step further and pinpoint our EV driving customers to specific substations, and start to determine the related strain at the meter and substation level, and even more granularly to the feeder level. In fact, we are working on going even further and getting to the transformer level for all of our detection and for our grid team.”
As part of that analysis, Andrew says Avista is better able to anticipate how EV charging is growing regionally — community-by-community.
“Our Spokane, Washington, AMI territory is very different when you look at communities in the north, south, east or west. We looked at a feeder in our South Hill region—which is typically higher income—and we saw that the EV adoption rate was five-to-one compared to communities in north Spokane. Now with that insight, we’re able to get to the feeder and substation level and identify varied growth models across our entire service territory. So, rather than having to enhance our grid across the board, instead of ripping the Band-Aid off and having to do
everything all at once, we can strategically start moving and enhancing our grid resiliency in alignment with where the growth is actually happening now and most likely to happen next.”
Active load management is another of Avista’s critical initiatives. Andrew says he appreciates that Bidgely’s Analytics Workbench allows the team to identify which customers are charging their vehicles during peak and off-peak hours.
“We can see that out of 2,300-ish EV owners, more than 2,000 are charging on-peak right now. We don’t have time-of-use rates yet, but we will in the future. So it’s very interesting to see the opportunity that we have to shift some of that load.”
Beyond grid planning, Andrew says the utility’s EV intelligence has also revolutionized its program marketing efforts.
“We had a program in 2023 through which we were paying a significant portion of the cost to enable customers to install an Avista level-two charger within their home. It was a strategic investment designed to enable us to learn more about charging behaviors and have the potential to control usage. But we only had 700 participants in that program, and we knew we had more than 700 EV customers within our service territory. We tried customer outreach, we tried doing surveys. We didn’t feel that the motor vehicle Department of Licensing was a path we wanted to go down. So we turned to Analytics Workbench, which was able to identify 1600 additional EV customers who were charging on our system. We could then target marketing campaigns to those customers to share awareness about the EV tools and resources that we were offering. Within hours, we provided corporate communications with a hyper-targeted list of account numbers, addresses, household energy usage, and their trending behavior.”
Heating and Cooling
Managing HVAC load is another program area that has benefited from behind-the-meter intelligence.
“One of our first Analytics Workbench use cases was to identify which customers within our service territory were high or low efficiency in terms of HVAC. We found some of our customers were using 12,000 kWhs a year in heating, compared with customers who had very high efficiency appliances who were only using around 4,000 kWhs per year. That’s incredible. That empowered us to direct market to those customers with low-efficiency appliances. We could identify which feeder and what substation they were on, and then just fully replace their appliance. I’m able to knock on the door and give them information they need.”
Equity Programs
Two use cases that Andrew was particularly excited to discuss are tied to the utility’s equity efforts.
When the USDA announced new grants to help utilities lower energy costs for rural communities, Avista initially didn’t think they had any communities that would qualify.
“When the federal grant came out to support customers within your service territory that spend more than 275 percent of the national average on energy with a grant, at first we didn’t think we’d have anyone who would be eligible because our rates are too low. But we ran the analysis with Analytics Workbench and within 10 minutes were able to identify 90 customers within one zip code that were qualified. So now we’ve got grant writers working on submitting for these grants. Without this tool, we would have discounted it. We would have said no to the opportunity.”
“We also have a Clean Energy Transformation Act (CETA) in Washington State that is focused on our low-income, vulnerable customers. We gave Bidgely a list of 150,000 accounts that are within that named community and were able to identify customers within the top 30 percent of heating usage or cooling loads. Now, we can go in and completely replace those appliances for those customers.”
The Use-Case Generation Continues
We hold bi-weekly working sessions with the Avista team to brainstorm new use cases and explore how to capture more value from behind-the-meter data to address their most pressing challenges. Andrew says those conversations, together with the discussions they now regularly have internally, are key to Avista’s continuing innovation.
“A big piece of our success is the cross functional collaboration this data allows. In my role, I am heavily involved with our energy efficiency, account and regional business managers, so I get looped into a lot of different conversations where my colleagues are saying, ‘It’d be great if we could do this.’ I get to raise my hand and say, ‘Hey, I think we can.’ It’s critical to make meter data available as a use case generator across the organization. You can’t just silo it within a single department, you have to share that awareness to achieve the maximum benefit.”
To hear more from Andrew Barrington about the Avista AMI data-driven approach, watch our on-demand webinar Seize the Data: Optimizing Grid and Customer Programs in 2024. To explore how to leverage AI and big data to improve future-readiness at your utility, connect with our team at [email protected].