Giuseppe Tommaso Costanzo

Fabrizio Sossan, Venkatachalam Lakshmanan, Giuseppe Tommaso Costanzo et al.

This paper describes the application of stochastic grey-box modeling to identify electrical power consumption-to-temperature models of a domestic freezer using experimental measurements. The models are formulated using stochastic differential equations (SDEs), estimated by maximum likelihood estimation (MLE), validated through the model residuals analysis and cross-validated to detect model over-fitting. A nonlinear model based on the reversed Carnot cycle is also presented and included in the modeling performance analysis. As an application of the models, we apply model predictive control (MPC) to shift the electricity consumption of a freezer in demand response experiments, thereby addressing the model selection problem also from the application point of view and showing in an experimental context the ability of MPC to exploit the freezer as a demand side resource (DSR).

Giuseppe Tommaso Costanzo, Sandro Iacovella, Frederik Ruelens et al.

Driven by the opportunity to harvest the flexibility related to building climate control for demand response applications, this work presents a data-driven control approach building upon recent advancements in reinforcement learning. More specifically, model assisted batch reinforcement learning is applied to the setting of building climate control subjected to a dynamic pricing. The underlying sequential decision making problem is cast on a markov decision problem, after which the control algorithm is detailed. In this work, fitted Q-iteration is used to construct a policy from a batch of experimental tuples. In those regions of the state space where the experimental sample density is low, virtual support samples are added using an artificial neural network. Finally, the resulting policy is shaped using domain knowledge. The control approach has been evaluated quantitatively using a simulation and qualitatively in a living lab. From the quantitative analysis it has been found that the control approach converges in approximately 20 days to obtain a control policy with a performance within 90% of the mathematical optimum. The experimental analysis confirms that within 10 to 20 days sensible policies are obtained that can be used for different outside temperature regimes.