Civil Engineering Reference
In-Depth Information
•
The minimum and maximum variations of the air velocity have been set to:
−
0
.
11
/
min
≤
V
Fan
i
≤
0
.
11
/
min
•
The control action weighting factor for the water flow,
ˉ
q
i
, is four times higher
than the control action weighting factor for the air velocity,
ˉ
V
i
:
4
ˉ
V
i
=
ˉ
q
i
It is important to highlight that, as with the PNMPC algorithm for one room
presented in Sect.
5.4.1
, the last condition has been included to take into account
the differences, in terms of energy consumption and economic cost, between the
pump which flows the water to all the rooms and the fancoils, which regulates the
air velocity in each room. Then, from these last statements, it can be inferred that
the PNMPC algorithm should use the air velocity as a control signal instead of the
water flow, of course, whenever there exists any possibility.
For the remaining parameters of the model, except for the number of people in the
room (
N
p
), real data have been considered. One simulation test which was performed
against the nonlinear model explained in Sect.
4.2.3
is presented as confirmation of
the goodness of the proposed control strategy. The simulation has a duration of two
days, that is,
t
2 days). The initial boundary conditions for
all the rooms are the typical ones of a summer day. Those boundary conditions
are similar for the three rooms apart from the number of people in each room, in
this way, the corresponding energy demand for each room is different too. Part of
the boundary conditions are shown in Fig.
5.41
, where in the top graph the radiant
temperatures (north, south, east, west, ceiling and ground) for room 3 are depicted.
The corresponding radiant temperatures for rooms 1 and 2 are similar, although their
values have been augmented by
=
2880 min (48 h
=
1
◦
C respectively,
to obtain slightly boundary conditions in the three rooms. The middle graph shows
the indoor air humidity (the corresponding units appear at the left axis) and the indoor
air velocity (the corresponding units are at the right axis). Lastly, the bottom graph
shows the direct solar irradiation.
The simulation results are shown in Fig.
5.42
. Such a figure is composed of four
graphs, which represent (from top to bottom) the PMV index, the air velocity of the
fancoil, the water flow and the number of people inside each room. All the graphs
have the same colour code: blue solid line represents room 1, green dashed line refers
to room 2 and red dashed-dotted line identifies room 3. Notice that in this simulation
the lower control layer, which is in charge of the fancoil actuator, is not taken into
account unlike real results presented in previous sections. Thus, the control signal is
the water flow demanded by the cascade controller instead of the water flow valve
position and, moreover, the sawtooth shape which can be appreciated in real results
does not appear here.
As can be observed, at the beginning of the simulation, the PMV index value is
far from the optimal thermal comfort, i.e. PMV
1
◦
C and decremented by
+
0
.
−
0
.
0. Then, to be able to reach the
ideal PMV index, the PNMPC must use the two control variables in all the rooms,
=
