This master thesis deals with designing a control scheme for heating of a low energy single-family house in connection with the OPSYS
2.0. The house is heated using hydronic underfloor heating circuits supplied with hot water
from a partially solar powered heat pump. A
two-level hierarchal control scheme has been
devised. The upper level calculates a heating budget based on average house dynamics
and forecasts of energy prices, power production of the house’s photovoltaic panels, and
the weather. The lower level determines how
to distribute this heating budget to the different rooms using a lumped parameter multi-zone model of the house and forecasts of the
weather. To facilitate the design, discrete
time, non-linear, grey-box models of the system has been developed for each of the two
levels. Using these models, state observers
have been designed in the form of Kalman filters. Both levels contain a model predictive
controller, that have been developed by reformulating the models into the mixed logical dynamical framework, which has allowed the two
control problems to be formulated as mixed
integer quadratic programming problems that
are more efficiently solved. The control scheme
has been shown to achieve disturbance rejection and stabilization of the system, by testing
it using a high-fidelity model of the house supplied by the OPSYS 2.0 participants.