Comfort Control Techniques for the Users of a Room - Comfort Control in Buildings - page 205

each room. Thus, the available water is a resource shared by all fancoil systems of
the bioclimatic building. In this case, the challenge arises when the energy demanded
by the occupied rooms in the building is greater than the one available. This energy
source, the solar collector field, depends on the solar irradiation available to feed the
absorption machine, that is, cloud transients sometimes do not allow the absorption
machine to operate at its nominal capacity. When the whole building is considered,
the available cool or hot water supplied by the absorption machine for the fancoil
system of each room is a resource shared by all fancoil systems of the bioclimatic
building.
However, when the energy demand is excessively high, for instance in hot sum-
mer and cold winter days, or when the solar collector cannot adequately feed the
absorption machine due to cloud transients, the control system may not maintain the
thermal comfort for all the rooms at the same time. In those cases, the absorption
machine must be fed by an auxiliary energy source, usually a gas heater or a boiler,
losing part of the benefits obtained when the energy is only produced by renewable
sources. Therefore, an optimal distribution of the available energy among the rooms
is compulsory, i.e. thermal comfort must be maximised considering the availability of
energy and variables such as the number of occupants, the room features (priorities)
and energy demand in each room in the building.
This problem is dealt with in Álvarez et al. ( 2013 ) where the problem previously
pointed out is solved through an optimal energy demand distribution among the
rooms of the CDdI-CIESOL-ARFRISOL building. Moreover, inÁlvarez et al. ( 2013 )
a tradeoff for the thermal comfort of all the rooms is wanted, considering that only
the energy generated by the solar collector is available, i.e. it is not allowed to turn
on the gas heater. The optimal energy distribution among the rooms is carried out
by means of a centralised control system for the whole building. The centralised
controller is a PNMPC one explained in Sect. 5.2.3 which includes an optimisation
problem. Noteworthy is that the optimisation problem can become hard-to-solve
when the number of rooms in the building increases, making the computational
time needed to provide the optimal solution exceed the constraints imposed by the
controller sample time. To deal with such an optimisation problem, a Lagragian
dual method has been deployed for the optimisation (Bazaraa et al. 2006 ), which
allows us to solve several optimisation problems in parallel and hence, to reduce
the computational effort proportionally to the available processing elements when
necessary. In order to validate the effectiveness of the proposed strategy, it is tested
through the nonlinear model presented in Sect. 4.2.2.1 .
5.6.2 Controller and Optimisation Method
As in this case, the system (the building) is considered to be composed of M sub-
systems, each sub-system i has its own cost function similar to the one presented in
Eq. 5.8 where both, the forced matrix response, G PNMPC i , and the free response
vector, f i , of each system are calculated through Algorithm 2. In this way, the