Engineers model power grid to solve solar problem

After receiving federal funding in February, a team of Pitt engineers aims to create a model of an all-inclusive electrical grid that incorporates solar power.

Right now, the electricity from solar power overpowers and damages some traditional electric grids, the systems that distribute electricity to homes and offices. With solar-produced electricity on the rise, undergraduate and graduate Pitt electrical and computer engineers from the Swanson School of Engineering will study how solar power and coal power can coexist on the same grid.

“We are going to start with just one circuit ­— that might be a residential street with a few dozens houses on it — and a handful of solar panels, and try to model that better,” Tom McDermott, an assistant professor at the University of Pittsburgh’s Swanson School of Engineering leading part of the study, said.

Researchers expect the project, which started Jan. 1, to last 33 months. Pitt announced its involvement in February. According to McDermott, the entire operation, which is based in Dallas, Texas, has received $4 million in federal funds. Pitt, as a part of the operation, received $250,000.

Other partners in the three-year federal grant are the Sandia National Laboratories, the National Renewable Energy Laboratory, the Georgia Institute of Technology, the Electric Power Research Institute and CYME, a power software engineering company.

McDermott said the project aims to make a model of a hypothetical solar grid that could help integrate solar power on the electric power grid.

According to McDermott, there are also not enough sensors installed on grids to track all the abnormal voltages.

Pitt’s challenge with its model is to reduce the large-scale theoretical model to a smaller model applicable to real electric grids while still maintaining the large, theoretical model’s accuracy.

McDermott said the team is currently working on model order reduction, which is a technique for reducing the complexity of mathematical models when running simulations of hypothetical power grids.

Andrew Reiman, a third-year graduate student and research assistant at Pitt who is also working on the project, said the model order reduction allows the team to condense the information on the grid — for example, consolidating 12 homes into one “equivalent block.”

“It would reduce the clutter in the system to free up computation power to study the parts that we do care about,” Reiman said.

Laura Wieserman, an electrical engineering graduate student and another core researcher, is working on developing a nonlinear mathematical model that closer resembles how electrical current and voltage interact.

Wieserman said the researchers will convert the model into distribution software platforms. such as MATLAB — a high-level language and interactive environment used by engineers and scientists. Using the program will help planning engineers create a better design to integrate solar power into our electric power grid.

To test the model, Santino Graziani, a summer research assistant, will run sag tests, which drops the voltage and disconnects a battery to protect it.

Graziani, a senior electrical and computer engineering student, said his summer work will make future predictions about how different inverters will react with one another.

According to McDermott, the researchers hope their simulation model, which takes data from a number of different places, will be universally applicable.

“We’ll try to expand [the model] if it works to be a bigger and bigger system like Oakland, then Pittsburgh and then the whole region,” McDermott said.