Meet the Expert: Daniel Molzahn

The NSF grant winner and associate professor is working to improve the resilience of power grids while also indulging in his other love: history. 

Daniel Molzahn will readily admit he’s a Cheesehead.  

Born and brought up in Wisconsin, the associate professor at the School of Electrical and Computer Engineering attended the University of Wisconsin, Madison, for undergraduate and graduate studies. It was also at Madison that he decided to go into the family business: power engineering. 

Molzahn’s grandfather was a Navy electrician in World War II and later completed a bachelor’s in electrical engineering. He eventually was plant director at a big coal plant in Green Bay. Molzahn’s dad was also a power engineer and worked at a utility company, focusing on nuclear power.  

It was not uncommon for family vacations to include a visit to a coal mine or a nuclear power plant. Being steeped in everything power engineering eventually seeped into Molzahn’s bones. “I remember seeing all the infrastructure that goes into producing energy and it was endlessly fascinating for me,” he says.  

That endless fascination has worked its way into Molzahn’s research today—at the intersection of computation and power systems. 

An early grounding in power engineering at Madison 

The move to the University of Wisconsin–Madison for undergraduate studies was not by chance. At the time, the university was one of few that hosted a robust power engineering group, a field Molzahn knew he wanted to pursue. He remembers attending weekly seminars where power companies would present the latest in the field. “It was just a very good environment to be in,” Molzahn says.  

Madison was such a good environment that Molzahn stayed on for graduate school. He already knew he wanted to pursue advanced studies and the power engineering department had some of the best minds in the field.  

Before the advent of computers, electrical engineers would build and test big replicas of the grid to simulate failure events. “If you wanted to see what would happen if a certain transmission line or a generator failed you would just pull a resistor out of the analog circuit,” Molzahn points out. The tedious process was ripe for computing and indeed, the electric power industry was one of the early adopters of computing, Molzahn says. At Madison, his research focused on “the intersection of algorithms and computations for power grids.” 

Molzahn pursued his doctoral studies as an NSF Fellow, a grant that gave him the ability to pivot on research projects as needed. It was by chance that he landed on his dissertation topic. As it turns out, his advisor Bernie Lesieutre, had come across an academic paper that presented a simplified way of solving notoriously difficult optimization algorithms. Lesieutre was not easily convinced. He assigned his students homework to solve an optimization problem using the method outlined in the paper. They couldn’t. At first Molzahn suspected he was doing something wrong but dug deeper and realized that the method proposed in the paper could fail when applied to practical power system models. The project snowballed with Molzahn even talking to optimization experts. “Figuring out why the problem failed, how it failed, how to fix these methods when they do fail, how to use them in ways that still give you rigorous answers, that turned into my PhD dissertation,” Molzahn says.  

After his PhD and a master’s in public affairs, Molzahn followed Ian Hiskens, a professor of power engineering at Madison to the University of Michigan where Hiskens had been invited to set up a related program. After a couple of years at Michigan, Argonne National Laboratory was the next stop, where Molzahn found a lot of funding was going into managing the uncertainty in power grids. It was here that Molzahn and many of his colleagues in the field were part of the early iterations of the Grid Modernization Laboratory Consortium, a partnership between the Department of Energy and the National Labs.  

Continuing computation research at Georgia Tech 

Knowing that he eventually wanted to complement research with teaching, Molzahn found a home at Georgia Tech. A recipient of an NSF CAREER Award, Molzahn is leaning on his computation background to develop an algorithmic foundation for the resilience of power grids.  

While traditionally we have focused on the reliability of power grids, which addresses day-to-day small events, accelerating climate change needs to address resilience as well, Molzahn says. “We’re trying to manage the high-impact events such as wildfires and hurricanes and try to minimize how many people lose power and to minimize restoration times,” he says. 

To do so, the research will have to address questions like: How do I route repair crews? How do I harden the infrastructure?  

From a computational perspective, it’s a resource-constrained optimization issue. The challenge, says Molzahn, is that it’s not just a vanilla problem. The variables are inherently discrete and nonlinear, both of which are factors that muddy the landscape. Complicating the picture is uncertainty from different failure possibilities. “There are many different scenarios that have discrete variables where you have to consider nonlinearities and extreme uncertainties and that’s the space we’re working in…to come up with algorithms that address these challenges,” Molzahn says. 

Grid resilience through another avenue—improving cybersecurity resilience—is another project that Molzahn is working on with Saman Zonouz, Associate Professor at the College of Computing. The long-term outcome will be a secure AI-enabled power grid solution. 

A love of outreach and history 

While heavy-duty computation is one way that Molzahn is addressing climate change, he is also researching public perceptions of measures to combat the crisis through a fun avenue: video games.  

As part of the outreach aspect of his NSF grant, Molzahn developed a Vertically Integrated Project (VIP) that brings together diverse students with diverse expertise to build a video game about power grid resilience. Players can adopt a variety of roles like power engineer or utility manager and try activities like building new transmission lines or batteries and solar panels. They also learn to understand the cost tradeoffs of operations like placing power lines underground and balancing them against the gained resilience advantage. The project had a soft launch recently and Molzahn is excited about being able to gather insights from player perspectives, which might feed machine learning algorithms. 

An avid history buff, Molzahn makes time for a project that preserves institutional knowledge and memories down generations. At a virtual power engineering conference, Molzahn first learned of Julie Cohn, a historian who presented an hour-long session on the history of power grid control algorithms. Julie Cohn’s father, Nathan Cohn, was a reputable engineer in the field and Julie had carefully preserved all his notes and correspondence.  

Molzahn was intrigued.  

He reached out to Cohn and today they’re working together on a Sloan Foundation-funded project that interviews power engineers and tracks seminal papers and historical developments in the field. Recorded and transcribed interviews will be archived at the IEEE History Center and the University of Houston. “Historians are really good at piecing together the broader context of how people think,” Molzahn points out.  

In preserving the past, Molzahn is not fueling his passion for power systems, he’s ensuring future generations of engineers learn from history. “We’re using these interviews as a lens to understand how we got to where we are today, what assumptions have been made for modeling techniques, and what will be appropriate for challenges going forward,” Molzahn says.