(555c) Microgrid Power Management Using Stochastic Real-Time Optimization and Control | AIChE

(555c) Microgrid Power Management Using Stochastic Real-Time Optimization and Control


Marvin, W. A. - Presenter, University of Minnesota
Trifkovic, M., University of Minnesota
Sheikhzadeh, M., Lambton College
Daoutidis, P., University of Minnesota-Twin Cities

Pressing environmental issues, security of energy supply, and energy safety concerns drive research and development towards renewable energy sources (RES) worldwide. A key attribute of RES, such as wind and solar resources, is intermittency (uncertain and not dispatchable). One strategy for coping with the intermittency is to combine different RES with energy storage. Power management of such systems is an active field of research.

This study deals with the development of an optimization-based power management tool for a 10kW energy system built in Lambton College, Sarnia, ON. The system consists of photovoltaic arrays, wind turbine, electrolyzers, hydrogen storage tanks, and fuel cells. Our approach builds off our previous work, which developed a comprehensive dynamic model and local controllers for each component.1 In this study, we investigate the effectiveness of a stochastic real-time optimization framework as a supervisory controller. A bilateral communication between the high-level and the low-level control layer enables components’ current states updates and minimizes the effect of model mismatch between two layers.  The supervisory controller framework incorporates wind, solar and demand forecasts to determine optimal set-points and operating schedules for this multi-component system. Forecasts of these stochastic time series are handled as a scenario tree generated using auto-regressive moving average models.  We compare the performance of this framework with our previously developed decision based supervisory power controller (a reactive logic-tree approach that neglects forecasts)1, as well as with a deterministic real-time optimization approach2. We demonstrate that the inclusion of forecasts can result in significant cost and energy savings.

1. M. Trifkovic, M. Sheikhzadeh, K. Nigim, P. Daoutidis. “Modeling and Control of a Renewable Hybrid Energy System with Hydrogen Storage,” IEEE Transactions on Control Systems Technology, 2013, doi: 10.1109/TCST.2013.2248156.

2. M. Trifkovic, W.A. Marvin, M. Sheikhzadeh, P. Daoutidis. “Dynamic Real-time Optimization and Control of a Hybrid Energy System,” In proceedings of the 12thEuropean Control Conference, Zurich, Switzerland, July 17-19, 2013.