Strengthening our power grid through AI

Editor's note: This expert Q&A is part of our “AI is everywhere ... now what?” special project exploring the potential (and potential pitfalls) of artificial intelligence in our lives. Explore more topics and takes on the project page.

Artificial intelligence is changing the nature of communication, transportation, finance, health care, education and many other fields. 

As these changes happen, Arizona State University researchers are devoted to maximizing the potential and minimizing the pitfalls of AI across a broad spectrum of disciplines. 

Anamitra Pal, an assistant professor of electrical engineering with ASU’s Ira A. Fulton Schools of Engineering, is developing AI to support our power grid. 

Here, he talks about his work with this technology. 

Question: What is the focus of your research?

Answer: My work seeks to maintain the reliability and resilience of the electric power grid in the presence of a diverse generation mix and the increasing frequency and intensity of extreme weather events. Essentially, my goal is to keep the power on, irrespective of whether the outside temperatures are at 120 F or 20 F. 

Q: How is AI changing research within your field?

A: The electric power system is interspersed with a variety of sensors that monitor its health on a regular basis. However, it is extremely difficult for human beings to make sense of the plethora of data produced by these sensors and make correct decisions before the system reaches the point of no return, such as a blackout or a brownout. It is in solving this problem — namely, extracting meaningful, actionable information from raw data in a timely manner — that AI is having the most impact in the field of power and energy systems.

Q: How is AI technology shaping the broader application of work from your field?

A: AI has the ability to overcome the limitations of conventional approaches by discovering intricate patterns that are embedded deeply in the data. Similarly, it can incorporate prior system knowledge into its learning mechanism — we call it physics-informed learning. These two attributes are enabling AI to support a variety of power system applications, such as enhanced situational awareness, adaptive protection and robust control.

Q: What opportunities related to AI in your field are motivating you?

A: I am particularly excited about explaining the operation of AI in the context of electric power system decision-making. For instance, one of my ongoing projects funded by the U.S. National Science Foundation is aimed at providing guarantees to the performance of black-box In the area of software testing, "black box" means the tester has no knowledge of the software's inner workings, mimicking an outsider's perspective. "Gray box" means the tester has partial knowledge of the system.AI models for a monitoring problem in power systems. Providing such guarantees are expected to make the black-box AI models “gray(-box),” which in turn will increase the trust of power system operators in using them for making important decisions.

Q: What challenges related to AI in your field need to be addressed?

A: A major challenge associated with AI in the field of power and energy systems is with regard to its black-box nature. The electric power system is a mission-critical system since its successful operation is an indicator of the prosperity of a nation. Therefore, before applying any technology to the power grid, the technology must be properly vetted. Unfortunately, the rapid progress of AI has made it extremely difficult to comprehensively examine all the consequences of using it to make decisions that will impact electric power system operation.

Q: Why is ASU important to the successful development of AI?

A: ASU has one of the best power system and data science research programs in the country, both in terms of quality as well as quantity. The diversity of the expertise that ASU faculty possess enables us to tackle multidisciplinary problems that are beyond the capability of an individual researcher or even a specialized group of researchers working in one domain or another. This is crucial for the successful application of AI in the field of power and energy systems because the nature of the problems are such that they require knowledge from all domains for finding the optimal answer.