Utility companies can make better decisions to distribute green energy with artificial intelligence (AI) software developed by Veritone, a Denver-based software company.
AI helps utility providers plan by building predictability and limits defined by government regulations into models for energy use. Green energy, or energy generated by natural resources like wind and water, tends not to be predictable. Phenomena like weather and climate change affect the amount and timing of available power.
AI can analyze data to see if there are patterns indicating regular occurrences. That way, a utility company can determine how much power it will have at any given time. AI also allows a utility to set preferences to ensure entities like hospitals retain power during extreme weather conditions.
The key to a smooth continuation of utility operations involving green energy is an intelligent distributed energy resource management system (iDERMS). This software platform controls types of distributed energy resources (DER) like solar power.
[Like what you read in The Carbonic? Help support climate journalism]
One of the essential tasks of iDERMS is to help a utility manage multiple microgrids. A microgrid is a small network of electricity customers that uses a local supply.
“The more grids you have out there, the more data exist. AI models are capable of storing and processing large amounts of data to help utility providers work more efficiently,” said Sean McEvoy, senior vice president of energy at Veritone.
Factors That Complicate Energy Use
Utility company operations are affected by political and public health concerns. Recent problems include the conflict between Russia and Ukraine.
“The conflict in Ukraine has spurred conversations about the fact that we can’t be reliant on oil. It’s increased the rush in development on green energy,” said McEvoy.
When the pandemic began, residential and commercial property owners began investing time and savings to install equipment to capture green energy.
“We saw more lithium batteries in garages, solar panels on roofs, and wind turbines at small businesses. Everyone from residential homeowners to wineries wanted to put their excess power into the grid,” said McEvoy.
The increase in available green energy has permanently made all types of customers less reliant on utility providers.
“With so many people saying, “Let me add the power to the grid,” prices for equipment have come down. There are fewer barriers toward capturing green energy. There’s now more green energy available for all consumers,” said McEvoy.
What Comes Next
The next step for utility companies is virtual simulations of microgrids. The providers use a concept called digital twin modeling to create a mockup of a neighborhood or city.
Working at the edge is also important, said Tatjana Legans, director of product marketing for Veritone.
“Edge computing is computing done at the site of where the data is collected. Veritone is currently working on pushing intelligence and compute power to the edge of the grid, right next to a solar plant, wind turbine, or source of hydroelectricity. That way number crunching and data analysis gets done at the source of the power,” said Legans.
Edge computing allows the utility provider to quickly determine the effect of a broken, stalled, or improved piece of equipment. With the information, the utility can synchronize AI controls to solve large-scale problems.
“Say we have five solar plants on one side of the city and six wind turbines on the other. We need to maximize the energy goals for that city. How can we accomplish this?” asked Legans.
The answer is to synchronize all the equipment using AI, build a digital twin model of each power source, and utilize the lowest-cost source first.
“We then move the loop in a supervisory manner, to maximize efficiency, watching the AI do its job,” said Legans.
Legans said AI can also incorporate data from outside sources to adjust the loop for abnormalities.
“If it’s going to rain on Tuesday, there won’t be as much solar power that day as on other days of the week. Knowing this ahead of time, the utility company can leverage battery storage or arrange to have a gas or diesel turbine running in the background to fill in the gaps. We can also build an AI model of a city or neighborhood to predict demand. Veritone’s software can even do price forecasting to predict prices on a wholesale market,” said Legans.
McEvoy said inputting all of the factors into the model helps utility providers produce and deliver green energy as it is needed, “not with 24-7 fossil fuel use.”
“We’ve been developing this software intensively over the last three years. Our engineers, our data science team, our energy software team, and our energy implementation teamwork cooperatively. We’ve been able to accomplish goals like helping a Florida utility company reduce fossil fuel production at this plant by 25 percent. We expect that number to grow to more than 70 percent,” said McEvoy.
McEvoy said energy cost savings benefits more than the utility company.
“Ideally, those reduced costs get passed on to the consumer. Further, the data analysis that AI allows helps the utility provider have more energy available for that spike at the end of the day when people come home and cook dinner,” said McEvoy.
Reducing stress on the grid at critical points trained the Florida utility company and others on how to distribute power better.
“That client is now offsetting the peak periods in the day with green energy. Hopefully, this will help them meet their needs during emergencies in the future,” said McEvoy.
Legans added that green energy problems are not unique to the West.
“Other countries, including India and Thailand, never had the ability to put in large transmission grids in remote areas. They’ve been using microgrids for years. Our software looks at solutions they’re employing. As countries such as the U.S. build more microgrids, we’re learning, borrowing, and exchanging ideas,” said Legans.
Drop a line to contact@thecarbonic.com for newsletter subscriptions, tips, questions or comments.
Comments