Tuyển tập công trình Hội nghị Cơ điện tử toàn quốc lần thứ 6 739
Mã bài: 160
Development of an adaptive temperature control for HVAC to
intelligent energy management system in buildings at DaNang city
Đề xuất bộ điều khiển thích nghi nhiệt độ cho thiết bị HVAC để quản lý
thông minh hệ thống năng lượng trong các tòa nhà tại TP Đà Nẵng
N. M. Tri
1
, N. H. Anh
2
, T. Q. Tuan
3
, H.D. Hoan
2
, N.D.H.Phuong
4

1
Vietnam Electricity-EVN

2
QuyNhon University, Vietnam
;
3
IDEA, Grenoble-INPG, France
Senior Member, IEEE
Abstract:
In order to use loads in an active and intelligent way to resolve technical problems in the networks or
contribute to ancillary services (smart grid), this paper presents a new method of air-conditioning control that
allows to reduce the peak consumption by maintaining thermal comforts. This control is based on the variable
set-point temperature of air conditioning adapted to the permissible power. This power can be fixed by outdoor

periods. Load management is dedicated to
control systems which monitor and plan the
energy demand of a building or larger zone.
They can be programmed to control lighting,
thermal comfort equipment, HVAC,
refrigeration equipment, pumps, valves and
motors (Fig. 1).
The sector of the building presents one of the
greatest potentials of energy efficiency and
reduction of the gas emissions. The use of the
loads in an active and intelligent way and
optimal load management is one of the major
L
740 N. M. Tri

, N. H. Anh, T. Q. Tuan

, H.D. Hoan

, N.D.H.Phuong
VCM2012
concerns of the managers, the providers and the
consumers of energy.
The peak consumption reduction is one of the
most effective solutions of energy management
systems. This reduction presents many interests:
 For the customers: reduce the bill for the
subscription and consumption in peak hours,
 For the DNO (Distribution Network
Operator): avoid the congestion and the

and permits to contribute to ancillary services in
distribution.

Figure 1. Energy and load management system
In Vietnam, the electric power consumption
is always superior to the electric power
production. The load sheddings in on-peak
periods in order to avoid over load or blackouts
are inevitable. The air-conditioning takes an
important part in the tertiary and residential
buildings. This is why the direct load control of
air conditioning presents one of the best
solutions to reduce peak consumption.
In this paper, an adaptive control of air-
conditioning units is proposed. Then the
proposed solution is applied for a distribution
network in order to reduce the peak load
consumption and to avoid congestion.

2. Load control
2.1 Techniques of load control
Techniques of load control can be presented
in the following: Time- Of- Use-Tariff,
Interruptible Load Tariffs, Distribution System
Loss Reduction (ex: reactive compensation).

2.2 Measurement for load management
Load control can be realised for lighting load
and HVAC. Load control strategies can be
presented as following:

associated with the evaporation of refrigerant is
transmitted to the ambient air;
- A compressor compressing the gaseous
fluids, increasing pressure and temperature;
- An exchanger condenser where gas
transfers its heat by condensing;
- A relief valve decreasing the pressure of the
liquid refrigerant before its evaporation in the
heat exchanger.
In this paper, the electrical analogue model
for an air conditioned house proposed in [7] is
used Fig. 2 shows the model of an air
conditioning developed with EMTP-RV. From
this model, we propose a new method based on
the adaptive control of air-conditioner for load
management system.

Iac S(t)
Is
Rw

Tw
Ti
IinstFigure 2. Electrical analogue model for an air
conditioned house
Where:
 Rw, Cw: the equivalent thermal conduction
resistance and thermal storage capacity of
the house (wall, base, roof)
 Rc, Ci: the equivalent thermal conduction
resistance of the average air infiltration and
thermal capacity of the air inside the house
 To, Tw, Ti: the exterior temperature, the
wall temperature and interior temperature
 Is : the current source of two components
(solar irradiation and the portion of internal
heat sources involved in this indirect heating
of air)
 Iinst: the current source of heat source
produced by lamp, computer, the body…

RcCi
To
Ci
tIacS
Ci
Iinst
dt
dTi
11
)(
 (2)
Where Tw and Ti are the unknown variables.
The Fig. 3 shows the EMTP-RV model of the
air conditioner that is built from this differential
equation system. In order to connect to
distribution network, the air conditioner is
modeled with EMTP-RV by a current injection.
f(u)1Fm1
f(u)1Fm12
f(u)1Fm13
f(u)1

+
+
+
+

sum7
!h
Int3
!h
Int4
f(u)
1
2
3Fm2 2
c
#P_AC#

C1
+
+
+
+

sum8
c
#T_Set#


scope
Q_AC
-1

Gain 2
scope
scp6
f(s)
fs1

f(u)1Fm2 5
scope
P_AC
P_Depas
f(u)=0
1
2Relay
Iinst
scope
F _ A C

Figure 3. Air conditioner modelized using
EMTP-RV
3.2. Proposed adaptive control


Classical
regulator

Air
conditioning

M
eteorologies
Conditions P_permissibleTemperature
regulator PID
and Fuzzy

-

T_room

P_total

Temperature
Set-point

Thermal model


reduction power. This value is transmitted to
each air conditioning in order to modify the set-
point temperature. In a distribution network with
different houses, the permissible power signal of
DNO is generated from sub stations.

4. Real-time control using Zigbee sensor
network for energy management system
in buildings
In the light of developments in microelectro-
mechanical systems (MEMS), along with
progress made in communication and embedded
smart sensors, the residential sector has a huge
potential for mitigating demand. The
possibilities of creating networks between home
appliances, sensors and wireless media, enable
the control of domestic equipment locally or
remotely via the Internet. The development of
WNS, with the Zigbee technology allows us to
establish more sophisticated control based on the
combination of measured information and
intelligent control in order to improve the use of
electrical equipments.
The advantages of this type of technology
ZigBee include:
 Elimination of all costs related to the
physical connection of devices.
 Possibility of establishing a single
communication interface between several
devices, using a communication protocol

Figure 5. Architecture of system
The Wireless Electrical Power Sensor
comprises two components: a power sensor and
a communication unit. The first component is
used to measure power consumption
instantaneously. The communication unit
transmits this information to the central cont rol
unit. The wireless electrical power sensor
receives the consum ption information and
detects excessive power (beyond the authorized
power limit).
air
conditioning

air
conditioning
Tuyển tập công trình Hội nghị Cơ điện tử toàn quốc lần thứ 6 743
Mã bài: 160
The central control unit : in our system, the
control unit is a computer equipped with a
wireless communication module and control
software. The central control unit analyzes the
information received. The central control unit
processes the temperature and power
measurements and makes decisions to control
the air conditionings in an intelligent manner,
maintaining comfort and keeping power
consumption below the authorized power limit.
The array of Wireless Temperature Sensors is
programmed to measure temperature within the


Figure 6. Daily variation of residential loads in
each house without air conditioners +
1

R3
+
1

R1
LF
LF1
Slack: 20.5kVRMSLL/_0
Phase:0
+
5nF

C1
p1 p2
N1 N2
ALM 70_1 30m

N1 N2
ALM 35 _157m
PI
p1 p2
N1 N2
ALM 35 _121m
PI
p1 p2
N1 N2
ALM 35_ 130m
PI
p1 p2
N1 N2
ALM 35_ 127m
PI
p1 p2
N1 N2
AL9 5_50S_470m
PI
1 2
DY_1

20/0.42
+
S_HTA

20.5kVRMSLL /_0
Slack:LF1
p
V_pu

N
Load_AirConditi oning
L_AC_6
L_AC
N
Load_AirConditioning
L_AC_7
L_AC
N
Load_AirConditio ni ng
L_AC_14
L_AC
N
Load_AirConditio ni ng
L_AC_9
L_AC
N
Load_AirConditi oning
L_AC_10
L_AC
N
Load_AirConditioning
L_AC_12
L_AC
N
Load_AirConditioning
L_AC_13
L_AC
N
Load_AirConditioning

rated power of the HV/LV transformer.
Fig. 10 shows the total power measured at the
transformer. It shows that there is a 10% overload
0 2 4 6 8 10 12 14 16 18 20 22 24
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
Times (H)
Power load (kW, kVAR )
744 N. M. Tri

, N. H. Anh, T. Q. Tuan

, H.D. Hoan

, N.D.H.Phuong
VCM2012
between 17 and 21H. The power of air-conditioner
and the interior temperature of the house at bus 4
are presented in Figs. 11 and 12. P
Q
S
Smax = 100 kVA

Figure 10. Total power (classical control) of the network
0 2 4 6 8 10 12 14 16 18 20 22 24
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
5.5
6
Times (H)
Air-conditioning power (kW)

Figure 11. Power of air conditioner of the house at bus 4
0 2 4 6 8 10 12 14 16 18 20 22 24
18.5
19
19.5

(with classical control)

It shows that the interior temperature is
maintained at 20°C (±1°C). The thermal comfort is
assured for all houses.
Fig. 13 shows the three phase voltage variation
at bus 4. In light load (0-6H) the voltage is high,
and in heavy load (10-22H) the voltage is low. The
voltage is always maintained between 0.9 and 1.1
pu.

5.2 Proposed method
In this case, the adaptive control is applied for
all air-conditioners in this network. The set-point
temperature of each house is 20°C (±1°C).
0 2 4 6 8 10 12 14 16 18 20 22 24
0
20
40
60
80
100
120
Times (H)
Total power (kW, kVAR, kVA)P
Q
S

2
3
4
5
6
Times (H)
Air conditioning power (kW)

Figure 15. Power of air conditioner of the house at bus 4
0 2 4 6 8 10 12 14 16 18 20 22 24
18.5
19
19.5
20
20.5
21
21.5
Times (H)
Interior temperature (°C )

Figure 16. Interior temperature of the house at bus 4
0 2 4 6 8 10 12 14 16 18 20 22 24
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7

) is 21°C. When the
permissible power is lower than 0.9Smax (90kVA)
the thermal comfort is broken. It means that with a
peak load reduction to avoid congestion superior to
10%, the comfort is not maintained with
T
setpoint
=21°C (±1°C). If the permissible power,
fixed by DNO, is 0.8 Smax (80 kVA), the maximal
temperature is increased to 23.5°C. This is
equivalent to 20% of load shedding.
6. Conclusion

The results of simulation show that the
proposed method permits to reduce efficiently the
peak consumption while maintaining thermal
comfort. The suggested method can be applied for
the various types of loads (ex: heating) and
adapted to the context in the future by taking into
account the economic and technical signals from
manager and DNO (ex: congestion, dynamic
tariff…). The obtained results show that this
method can be applied to contribute to ancillary
services such as voltage control in distribution
networks.
The proposed solution is applied to a group of
loads or buildings (such as a virtual consumer) in
order to reduce the peak consumption (or
congestion management) in a large distribution
network. In order to reduce peak consumption, this

Trans. on Power Systems, Vo. 23, No. 3, Aug.
2008.
746 N. M. Tri

, N. H. Anh, T. Q. Tuan

, H.D. Hoan

, N.D.H.Phuong
VCM2012
[6] D. Bargiotas and J.D. Birdwell, "Residential
air conditioner dynamic model for direct load
control," IEEE Trans. Power Delivery, vol. 3,
no.4, pp.2119-2126, October 1988.
[7] Suresh Kumar K.S. “Control Strategies for
Energy Conservation in room air conditioning
units – Matlab/Simulink simulation Study”,
Department of Electrical Engineering,
National Institute of Technology Calicut,
Calicut-673601, Kerala State, India.
[8] Qinhua, H., et al. "Two ANN-Based Models for
a Real MVAC System", International
Conference on Wireless Communications,
Networking and Mobile Computing, WiCom
2007.
[9] K. Le, T. Tran-Quoc, JC Sabonnadière, Ch.
Kieny, N. Hadjsaid, “Peak load reduction by
using heating regulators”, CIRED, Vienna,
21-24 May 2007.