Machine learning models will allow UI to forecast grid performance ahead of severe storms, troubleshoot problems areas and pinpoint investments to increase reliability
ORANGE, Conn. — August 16, 2023 — When describing the day-to-day work of an electric utility, ‘pushing the boundaries of technology’ might not be the first descriptor that comes to mind. However, Avangrid, Inc. (NYSE: AGR), a leading sustainable energy company and member of the Iberdrola Group, is looking to change that. Avangrid has launched a Data Science and Analytics team aimed at developing unique and proprietary artificial intelligence (AI) systems that are focused on improving reliability for customers. Once complete, these machine learning models will change how the company approaches grid investments, equipment upgrades, storm forecasting and more.
“We’re reimagining what’s possible for a utility when it comes to data science and analytics,” said Pedro Azagra, Avangrid CEO. “Traditionally, we’ve partnered with third parties to integrate this type of cutting-edge technology into our business. Now, we have the talent in-house to create machine learning models that Avangrid will own, and as a member of the Iberdrola Group we’re also collaborating with our global colleagues to lean on their experience and expertise. Owning these AI systems will allow us to continually improve them while reducing costs and increasing efficiency.”
Avangrid’s Data Science and Analytics team sits within the Operational Performance organization and is made up of seven data scientists, engineers and analysts who come from prestigious industries like healthcare, astrophysics and finance. The team is creating three unique AI systems: Predictive Health Analytics, GeoMesh and HealthAI. Each technology will take existing data from Avangrid companies’ electric grids and analyze it to forecast future performance of the grid, determine the condition of grid equipment or target at-risk locations for inspections and investment. Ultimately, this will lead to increased reliability for the 2.31 million electric customers served by Avangrid’s subsidiaries, including United Illuminating.
“We’re using data to answer questions that we previously thought were unanswerable,” said Catherine Stempien, Avangrid Networks President and CEO. “Utilities are no longer just a poles and wires business—we’re paving the way for the future of grid management and reliability. By creating and owning these AI systems, we’re taking greater control over the care of our network, customers and communities. Our top priority will always be the day-to-day power reliability for our customers, so we’re very excited to roll out these machine learning models for their benefit.”
Predictive Health Analytics
This project takes a proactive approach to determine the condition of substation equipment such as circuit breakers, which are like fuses at your house, and uses data to prioritize planned replacements and upgrades. Traditionally, equipment is replaced primarily based on age or if it malfunctions and causes an outage. Predictive Health Analytics will take a proactive, data-driven approach to determine equipment’s overall health and life expectancy based on numerous factors, including age, frequency of use, manufacturer and maintenance notifications. This means Avangrid will save money and help reduce outages for customers by replacing at-risk equipment, such as circuit breakers, before it causes an outage.
This project maps Avangrid’s service areas to identify the strengths and weakness of its electric networks to help forecast its performance during both blue-sky and storm scenarios. The goal is to improve understanding about how the electric grid is performing under various weather conditions so that Avangrid’s companies can better plan upgrades, storm response and more.
To accomplish this, GeoMesh breaks Avangrid’s service areas into small sections to allow the company to focus on one specific region at a time. For the chosen selection, GeoMesh makes predictions by analyzing millions of data points, such as average wind speed, precipitation type and amount, outage history and reason, population and density of tree limbs and other vegetation. All of this lets Avangrid make informed, data-based decisions on things like where and what upgrades are most needed or which customers are most likely to be impacted by a storm.
“We want to understand at a local-level where our grid is most susceptible to the impacts of extreme weather like heavy, wet snow or heat waves,” said Mark Waclawiak, Senior Manager of Operational Performance at Avangrid. “Analyzing historical weather and reliability data can help give us that insight. Once we do this, we can tailor our grid investments to harden our grid and offset the impacts of ever intensifying storms due to climate change.”
This project analyzes Avangrid’s existing millions of high-resolution photos of its street-level distribution system—poles, wires and grid equipment—to identify the assets in the photos and, eventually, catalogue their health. This increases Avangrid’s awareness of the condition of its grid equipment and helps to identify areas of concern. HealthAI will save Avangrid time and money by targeting at-risk locations for inspections and maintenance. It will also improve reliability for customers by reducing outages and improve safety for lineworkers by giving them more information before they arrive on scene.
Currently, Avangrid is training the AI system to correctly identify grid equipment in photos, such as cross arms, transformers or wire. Next, the AI system will learn to analyze and determine the health of that equipment. For instance, it will identify if the cross arm is broken or if the wire is sagging. Right now, Avangrid learns of these equipment damages or failures from customer reports, manual inspections or outages. HealthAI will make it an automatic and proactive process to make adjustments before the customer is impacted. Long term, Avangrid aims to also use HealthAI to also identify threats to its distribution network such as hanging tree limbs or dead trees that may fall onto company electric lines.