Civil Engineering Reference
In-Depth Information
Table 5.2
Lambda (
)
choices as a function of ISE
criterion
ˉ
Strategy
Lambda (
ˉ
)
ISE criterion
(A):
J
k
1
0.1
20.01
0.3
33.46
(B):
J
k
2
0.0065
27.29
0.008
33.37
(C):
J
k
3
0.001
19.77
0.00355
33.53
(D):
J
k
4
0.012
27.83
0.2
58.53
More specifically, the results of real test figures can be divided into two parts:
the left side contains the results for the day in which the lower value of
of each
strategy has been tested, while the right side shows the results for the day in which
the higher value of
ˉ
has been tested. Besides, each column shows the PMV index
evolution, the indoor, outdoor and mean radiant temperatures, the fancoil velocity
(control signal) and the indoor relative humidity.
Table
5.3
shows a comparison among the different strategies. This comparison is
based on three performance indices: (i) first index is useful to determine if the actuator
is over-actuated, i.e. it expresses the mean number of changes in its state (on/off) per
hour, (ii) the second index determines the percentage of time the actuator is used and
(iii) the third index shows the mean energy consumption per hour associated to the
use of the actuator.
First of all, in Fig.
5.15
the results obtained for strategy (A) are shown. These
results have been obtained under similar conditions of indoor air temperature and
relative humidity. In this figure, that the PMV index is almost equal to zero with
ˉ
=
ˉ
0
.
1 can be observed. However, using a value of
ˉ
=
0
.
3, the PMV index is
around 0
15, due to the fact that the reference governor weighs more abrupt changes
in the temperature setpoint, see Eq.
5.23
. In addition, if both control signals provided
by the controller in each situation are compared, it can be noticed that in the first
case (
.
1) the control signal is approximately equal to 30% and in the other, it
is equal to 20%, with a slowly increasing trend trying to slowly reduce the error. In
both cases, five occupants were in the room. It is important to remember, although
this fact was just pointed out at the beginning of this section, that in these results as
in the results of the following strategies, there is a narrow offset in the PMV index,
that is, the PMV index is not equal to zero but is close to this value, due to the neutral
zone of 0
ˉ
=
0
.
25
◦
Cwhich has been added to avoid oscillations around the PMV setpoint.
The results registered for strategy (B) are shown in Fig.
5.16
. As in the previous
case, it can be observed that the PMV index is almost equal to zero with both weight-
ing coefficients. However, with
.
008 the time used to reach a PMV index value
equal to zero is greater than with a weighting coefficient of
ˉ
=
0
.
0065. Moreover,
if both control signals provided by the controller in each situation are compared, it
can be noticed that in the first case (
ˉ
=
0
.
0065), the control signal is approximately
equal to 50% and in the other case, the effort necessary to reach an optimal thermal
comfort situation changes continuously because of disturbances. More specifically,
ˉ
=
0
.
