Artificial intelligence in the energy industry could mean savings for Illinois residents and could help bring businesses back online faster after power outages, but there are still concerns among some researchers.
A state Senate committee focused on the use of AI, or machine-learning technologies, and how it’s used in the state’s energy grid.
Argonne National Laboratory researcher Rick Stevens said AI can perform certain human functions quicker than humans.
“It can be more accurate in predicting the failure [of a system] than an individual engineer,” Stevens said.
That’s because increased computing power enables AI to collect and process more information faster.
Siemens engineer Sacha Fontaine said continued AI advancements could help during storm preparedness and response by being able to predict what parts of the grid may be most vulnerable.
Fontaine also said smart meters already in use in many residential areas can provide usage readings every 15 minutes, instead of once a month. AI can check if a customer is optimizing use of electricity for charging electric vehicles.
“Because if they're not, they can contact that customer and let them know what savings they can have based on their historical use and forecasting the additional savings if they did switch to EV rates,” Fontaine said.
There could be much bigger savings businesses and other large electricity users.
Mohammad Shahidehpour of the Illinois Institute of Technology said the school's small-scale smart grid using AI to optimize energy use has helped that institution save $1 million a year.
“So imagine if a small operation of this size, about 12 megawatts, is able to save about a million dollars annually, how much we could save if we could convert a larger community to a microgrid,” Shahidehpour said.
Representatives from both the Illinois Commerce Commission and Illinois Consumers Utility Board said there’s great potential for future consumer savings through such technology and the Internet of Things, or the growing ability for household and personal electronics to communicate with each other on a network.
But with AI comes challenges. School of Applied Technology’s Bob Carlson said it could leave networks vulnerable.
“Everything is the Internet of Things and that means it’s all hackable,” Carlson said.
He suggested that in developing such technologies, engineers must go in knowing how to best secure networks.
There are also concerns about how data collected by AI could compromise consumer privacy.