Energy management of a building cooling system with thermal storage: A randomized solution with feedforward disturbance compensation

We consider a cooling system that comprises a building composed of multiple thermally conditioned zones, a chiller plant, and a thermal storage unit. The electrical energy price is time-varying, and the goal is to minimize the electrical energy cost along some look-ahead time horizon while guaranteeing an appropriate level of comfort for the occupants of the building. To this purpose, we can appropriately set the temperatures profiles in the zones of the building and the cooling energy exchange with the storage. Since the cooling system is affected by stochastic disturbances, we adopt a stochastic formulation of the control problem, where constraints are imposed in probability and measurable disturbances are possibly compensated. The resulting chance-constrained optimization problem is then solved via a randomized approach. Numerical results show a significant reduction of the cost when the feedforward disturbance compensation scheme is adopted.