Distributed model predictive control using a trigger time search for optimal heating of multizone buildings

Fatima Zohra Benali (1) , Patrick Schalbart (1) , Bruno Peuportier (1)

1 CES - Centre Efficacité Énergétique des Systèmes

Optimising operation and control of buildings' electric loads is essential to mitigate peak demand and adapt to intermittent production. Predictive control has been applied to building energy management for more than twenty years. Solving an optimal control problem (OCP) allows energy saving and flexibility while maintaining thermal comfort. Resolution methods for OCP require either expert knowledge or intensive computational efforts, which complicates commercial implementation. A simple OCP resolution approach implemented into a distributed model predictive control is presented in this communication. The original contribution of this work stems from the combination of Bellman's and Pontryagin's principles. Results showed a reduction of computation time compared to a reference method with similar energy saving (28 %) when applied to a two-zone building using Time of Use tariff assumptions.

Domaines: Sciences de l'ingénieur [physics]

Fatima Zohra Benali, Patrick Schalbart, Bruno Peuportier. Distributed model predictive control using a trigger time search for optimal heating of multizone buildings. Building simulation 2023, IBPSA, Sep 2023, Shangaï, China. ⟨hal-04140658⟩