When an electrical grid is damaged from climate change events like storms and severe weather, millions are affected. Those events cause grid anomalies that lead to environmental damage, such as deadly wildfires nationwide. In the winter of 2021, power blackouts in Texas left millions without heat and killed hundreds.
From the destructive 2020 Oregon wildfires to the recent June 2023 wildfires in Canada affecting New York, wildfires have devastated the environment and local economies. According to a Nature Sustainability study, the estimated economic impacts of California wildfires in 2018 totaled $148.5 billion.
According to Kaitlyn Albertoli, CEO and co-founder of Buzz Solutions, electrical grids are up against three major challenges from increasingly significant climate change-related weather events like wildfires and hurricanes; the aging of the transmission and distribution infrastructure that was constructed over half a century ago, the amount of electricity needed to support modern technology like electric vehicles (EV).
“But utilities are not lacking in data. IoT sensors and smart meters provide real-time monitoring of the health of our power infrastructure, but analyzing and managing this data so that utilities can act proactively is the challenge,” said Albertoli in an email interview.
One solution is to transform the way power utilities inspect infrastructure for damage, encroaching vegetation and other issues that can lead to grid-sparked fires or affect grid resiliency in times of high power demand.
“Enter artificial intelligence (AI),” said Albertoli. “Until now, much of the data analysis has been done manually—creating a long backlog of necessary maintenance.”
The New York Power Authority (NYPA) is using AI to detect anomalies in large data sets, which helps utilities prioritize critical maintenance.
“The value of using AI and computer vision to assess the health of the grid is its scalability,” said Albertoli. “For utilities like NYPA, cutting down their maintenance backlog has been huge because it has given room to focus on innovation that will directly address the three major challenges I called out earlier.”
AI does the heavy lifting
Vik Chaudhry, the co-founder, CTO and COO of Buzz Solutions, says that AI takes most of the manual analysis work out of grid inspection and is responsible for 70% time savings of person-hours for data processing and anomaly analysis. “It is trained to identify anomalies like rust, missing or broken parts and damage.”
“Accuracy is improved through a human-in-the-loop process, meaning that if a technician notices that the AI has incorrectly identified or missed an anomaly, it can manually override the system and simultaneously improve the model through human feedback,” said Chaudhry.
But the differences in how AI models are trained can make the difference in accurate predictive maintenance and the ability to predict and prevent grid problems that can lead to wildfires and blackouts.
Chaudhry says the threats the grid faces vary widely and are primarily regional. For example, you are unlikely to see massive ice storms in Florida, but the state is subject to severe hurricane damage. Both natural disasters affect the T&D infrastructure differently.”
“But what we’re seeing is that climate change is causing unpredictable weather events throughout the country—like tornadoes on the coast and devastating winter weather in Texas,” added Chaudhry.
“An AI model trained on data from across the country rather than from a single regional utility can better predict grid problems as a result of a one-of-a-kind weather event,” said Chaudhry. “Utilities are best prepared to react to and prevent significant disruptions from weather events they are used to. It’s the anomalies that lead to wildfires and blackouts.”
Chaudhry says utilities can put this technology to work almost immediately, but it requires change management inside the organization.
“Utilities are looking for a solution and are willing to adopt technologies that streamline grid repair and maintenance,” said Chaudhry. “From onboarding to full adoption, utilities can expect to see changes after five to six weeks of using the tool.”
Buzz Solutions has several North American utilities leveraging their tools to save time and cost as a part of their inspection work. “Buzz delivers insights directly to the utility with prioritized information, streamlining the path to maintenance,” said Albertoli. “Ultimately, this enables a more seamless process of analyzing inspection data while saving north of 50% cost savings and six to seven months of manual analysis time.”
Applying technology during periods of environmental changes
“There is a perfect storm of factors right now that are driving forward the adoption of technologies like ours in the utility industry, including aging grid infrastructure, climate change, technology readiness/cost-effectiveness for inspections and mandates for more frequent inspections,” said Albertoli. “As a result of these driving factors, the market is being forced into a wave of a digital transformation.”
Albertoli says utilities must inspect a certain portion of their lines yearly using drones, helicopters and fixed-wing aircraft.
“This has caused utilities to capture five to ten times the data, which is 100,000s and millions of images on an annual basis,” said Albertoli. “The current means of analyzing this data is entirely manual, with linemen and field technicians.”
Albertoli says the need for an AI-powered solution has never been more critical to save time, money and reduce the lag between image capture and fully analyzed reports.
Buzz Solutions is venture-backed and has raised $3.3 million to date.
Read the full article here
