Texas A&M researchers develop AI tools for a changing power grid
As electricity demand surges, Texas A&M researchers are developing AI-powered technologies to help grid operators manage growing complexity and maintain a reliable power system.
As power demand surges across Texas, researchers at Texas A&M University are developing AI tools to help make the electrical grid more efficient, reliable and resilient.
Drs. Xin Chen and Prasad Enjeti in the Department of Electrical and Computer Engineering and their research teams have created AI tools to help manage and improve power grid operations as demand increases. Chen is developing Grid Agent, which helps operators analyze and manage the electric grid, while Enjeti created Circuit AI, which helps engineers design and optimize the hardware that keeps the grid running.
“Today’s power grid needs to accommodate the rapid growth of electric vehicles, solar and wind generation, energy storage and large-scale data center loads, while also remaining resilient during extreme weather events,” said Chen. “As the system becomes increasingly complex, we are developing advanced AI tools to support grid operators with timely decision-making and professional-grade analysis.”
Leveraging AI to improve the power grid
With AI systems like Grid Agent and Circuit AI, engineers and grid operators can interact with complex engineering software using simple language. Instead of manually performing complex analyses or endlessly searching through manuals and parts lists, an operator could ask a simple question like, “Can you perform a grid analysis for me?” or “Can you locate this part for me?” Then, receive a clear, reliable answer.
Chen and Enjeti’s AI tools both address similar challenges at different levels of the power grid. Grid Agent works at an overall system level, helping operators analyze how electricity moves through the grid, identify potential problems and assess how the system would respond to disruptions. Circuit AI works at the hardware level, helping engineers select components, optimize converter designs and evaluate the reliability of the equipment that makes up today’s electric grid. Together, Grid Agent helps operators manage the grid while Circuit AI helps engineers design reliable hardware to keep it running.
Both agentic AI systems combine the reasoning capabilities of large language models (LLMs) with trusted engineering software, validated calculations and specialized data. Rather than relying on AI-generated responses alone, they ground every answer in established engineering methods to produce accurate, reliable results.
“We have several checks in place, so we’re not relying on the LLM alone or simply accepting everything it generates,” said Chen. “The AI works with trusted power system programs, combining the flexibility of an LLM with tools that engineers already know and rely on.”
With the growth of AI tools, researchers will also need ways to measure how well they perform. Chen’s team is developing a benchmark platform with standardized questions and tasks researchers can use to evaluate and compare different AI systems. The goal is to provide a consistent way to determine how well these systems understand and solve complex power system problems.
The next wave of power engineering
As AI grid assistants continue to develop, the field of power engineering could see a major shift in how engineers design, operate and maintain power systems. With AI systems working together across different levels of the power grid, engineers and operators could identify problems sooner, make more informed decisions and respond to changing conditions more quickly.
“I see Grid Agent and Circuit AI as two layers of a future ‘AI-native’ grid,” said Enjeti. “Grid Agent could detect problems at the system level, while the Circuit AI agent could assess the health of individual components and recommend solutions. This connection between the equipment and the system operator is where I believe the field is heading.”
For Chen and Enjeti, the goal is not to replace engineers, but to give them tools that can handle time-consuming analyses and help them make better decisions. As the power grid grows more complex, these AI assistants could become an increasingly important part of how engineers keep it reliable, efficient and resilient.
