Model Predictive Control for Smart Energy Systems

In this PhD project, Economic Model Predictive Control is proposed to balance the production and consumption of electricity in a Smart Grid such that the total cost is minimized. A predictive control system has the ability to embed forecasting methodologies for the stochastic variables like electricity production, consumption, and price and even weather conditions like wind speed, temperature, or solar radiation. As well as forecasts, the predictive models used by the controllers are uncertain due to unknown events, model reduction and parameter uncertainty. Algorithms based on convex optimization and decomposition will solve this stochastic optimization problem. Control methods ranging from centralized to decentralized model predictive control using various hierarchical levels and levels of information exchange between the individual controllers are the scope of the project. In particular decomposition techniques based on price signals and predictions of future price will be investigated.