Input Convex Neural Networks for Building MPC

Felix Bünning, Adrian Schalbetter, Ahmed Aboudonia, Mathias Hudoba de Badyn, Philipp Heer, John Lygeros
Proceedings of the 3rd Conference on Learning for Dynamics and Control, PMLR 144:251-262, 2021.

Abstract

Model Predictive Control in buildings can significantly reduce their energy consumption. The cost and effort necessary for creating and maintaining first principle models for buildings make data- driven modelling an attractive alternative in this domain. In MPC the models form the basis for an optimization problem whose solution provides the control signals to be applied to the system. The fact that this optimization problem has to be solved repeatedly in real-time implies restrictions on the learning architectures that can be used. Here, we adapt Input Convex Neural Networks that are generally only convex for one-step predictions, for use in building MPC. We introduce additional constraints to their structure and weights to achieve a convex input-output relationship for multi- step ahead predictions. We assess the consequences of the additional constraints for the model accuracy and test the models in a real-life MPC experiment in an apartment in Switzerland. In two five-day cooling experiments, MPC with Input Convex Neural Networks is able to keep room temperatures within comfort constraints while minimizing cooling energy consumption.

Cite this Paper

BibTeX
@InProceedings{pmlr-v144-bunning21a, title = {Input Convex Neural Networks for Building {MPC}}, author = {B\"unning, Felix and Schalbetter, Adrian and Aboudonia, Ahmed and de Badyn, Mathias Hudoba and Heer, Philipp and Lygeros, John}, booktitle = {Proceedings of the 3rd Conference on Learning for Dynamics and Control}, pages = {251--262}, year = {2021}, editor = {Jadbabaie, Ali and Lygeros, John and Pappas, George J. and A. Parrilo, Pablo and Recht, Benjamin and Tomlin, Claire J. and Zeilinger, Melanie N.}, volume = {144}, series = {Proceedings of Machine Learning Research}, month = {07 -- 08 June}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v144/bunning21a/bunning21a.pdf}, url = {https://proceedings.mlr.press/v144/bunning21a.html}, abstract = {Model Predictive Control in buildings can significantly reduce their energy consumption. The cost and effort necessary for creating and maintaining first principle models for buildings make data- driven modelling an attractive alternative in this domain. In MPC the models form the basis for an optimization problem whose solution provides the control signals to be applied to the system. The fact that this optimization problem has to be solved repeatedly in real-time implies restrictions on the learning architectures that can be used. Here, we adapt Input Convex Neural Networks that are generally only convex for one-step predictions, for use in building MPC. We introduce additional constraints to their structure and weights to achieve a convex input-output relationship for multi- step ahead predictions. We assess the consequences of the additional constraints for the model accuracy and test the models in a real-life MPC experiment in an apartment in Switzerland. In two five-day cooling experiments, MPC with Input Convex Neural Networks is able to keep room temperatures within comfort constraints while minimizing cooling energy consumption.} }
Endnote
%0 Conference Paper %T Input Convex Neural Networks for Building MPC %A Felix Bünning %A Adrian Schalbetter %A Ahmed Aboudonia %A Mathias Hudoba de Badyn %A Philipp Heer %A John Lygeros %B Proceedings of the 3rd Conference on Learning for Dynamics and Control %C Proceedings of Machine Learning Research %D 2021 %E Ali Jadbabaie %E John Lygeros %E George J. Pappas %E Pablo A. Parrilo %E Benjamin Recht %E Claire J. Tomlin %E Melanie N. Zeilinger %F pmlr-v144-bunning21a %I PMLR %P 251--262 %U https://proceedings.mlr.press/v144/bunning21a.html %V 144 %X Model Predictive Control in buildings can significantly reduce their energy consumption. The cost and effort necessary for creating and maintaining first principle models for buildings make data- driven modelling an attractive alternative in this domain. In MPC the models form the basis for an optimization problem whose solution provides the control signals to be applied to the system. The fact that this optimization problem has to be solved repeatedly in real-time implies restrictions on the learning architectures that can be used. Here, we adapt Input Convex Neural Networks that are generally only convex for one-step predictions, for use in building MPC. We introduce additional constraints to their structure and weights to achieve a convex input-output relationship for multi- step ahead predictions. We assess the consequences of the additional constraints for the model accuracy and test the models in a real-life MPC experiment in an apartment in Switzerland. In two five-day cooling experiments, MPC with Input Convex Neural Networks is able to keep room temperatures within comfort constraints while minimizing cooling energy consumption.
APA
Bünning, F., Schalbetter, A., Aboudonia, A., de Badyn, M.H., Heer, P. & Lygeros, J.. (2021). Input Convex Neural Networks for Building MPC. Proceedings of the 3rd Conference on Learning for Dynamics and Control, in Proceedings of Machine Learning Research 144:251-262 Available from https://proceedings.mlr.press/v144/bunning21a.html.

Related Material