ECE PhD student Anna Stuhlmacher is working to improve the sustainability, flexibility, and resiliency of two critical infrastructure systems, the power distribution network (PDN) and the drinking water distribution network (WDN), by optimizing their interaction.
“We can use the water distribution network to provide the services that fast-ramping power plants (such as natural gas units) and voltage control devices would normally provide to the power network,” Stuhlmacher said. “The water distribution network can effectively be treated like a big battery, but a battery that has all these additional constraints and responsibilities.”
Drinking water is pushed through the WDN by way of water pumps, which are loads in the PDN. By storing water in elevated water storage tanks or changing which pumps are being used, it’s possible to shift the power consumption of the pumps. This makes the WDN inherently flexible, and when leveraged, this flexibility operates as a controllable asset for the PDN, ensuring that voltage in the PDN remains within a safe operational range.
“One of the biggest challenges is that there’s uncertainty in both of these networks,” Stuhlmacher said.
This uncertainty can increase as more renewable energy sources, which are more variable, are added to the grid. To address this uncertainty, Stuhlmacher uses robust optimization and monotonicity properties. This means that if the water pump power increases, then the hydraulic head at each node in the water distribution network and the tank levels – the two main safety parameters of concern – will also increase or decrease, respectively, across the entire network.
“We incorporate this uncertainty into the problem to make sure that our water pumping solution can handle both the water and power network constraints for any type of uncertainty that crops up,” Stuhlmacher said. “By looking at the most extreme pump power increase and the most extreme pump power decrease, we can evaluate the feasibility of the system for the entire range.”
In “Tractable Robust Drinking Water Pumping to Provide Power Network Voltage Support,” published by the IEEE Conference on Decision and Control, Stuhlmacher and her collaborators proved that their method produces robust solutions and is computationally tractable. Stuhlmacher earned her bachelor’s degree in Electrical Engineering from Boston University and her master’s degree in Electrical Engineering from U-M. She is advised by Prof. Johanna Mathieu.