Civil Engineering Reference
In-Depth Information
y
1
y
2
y
3
y
4
setpoint
22
20
18
20
40
60
80
100
120
140
time [min]
3
u
1
u
2
u
3
u
4
2
1
0
20
40
60
80
100
120
140
time [min]
10
8
solar plant temperature output
6
20
40
60
80
100
120
140
time [min]
8
solar plant flow output
6
4
20
40
60
80
100
120
140
time [min]
Fig. 6.3
Subsystem outputs and controls, solar plant temperature and water flow with DMPC
controllers.
Source:
As a courtesy of the authors (Scherer et al.
2014
)
With the agents network active, each agent always took into account the next
agent. Thus, when the last agent could not reach its reference, the previous agent
reduced its own consumption of chilled water. When the penultimate agent helped
the last agent, it ended up moving away from its own reference, so the previous
agent also reduced its consumption. The process was repeated until the first agent
reduced its consumption. This behaviour indicated that all agents cooperated with
each other, resulting in operations where all agents were practically equally sharing
the available resources. These behaviours can be observed in Fig.
6.3
mainly in two
periods of time: between 40 and 80 min and between 110 and 130 min. In other
words, the DMPC strategy is able to adequately counteract the disturbances caused
by the input temperature fluctuations.
6.3 Comfort Control Using Day-Ahead Pricing
address the problem of comfort control in an efficient way. More specifically, these
strategies were based on MPC techniques and they try both to maintain thermal com-
fort, and to reduce the use of the HVAC system. Nevertheless, there is an important
factor which was not taken into account by these strategies: energy price, i.e. they
try to reduce energy consumption, by penalising the use of the HVAC system, but
they do not consider its economic cost. Therefore, as energy price has a dynamic
behaviour along the whole day and also as a function of the period of the year,
see Fig.
6.4
, energy can be consumed at these instants in which it is less expensive.
