Energy and Buildings 40 (2008) 20522058

Contents lists available at ScienceDirect

Energy and Buildings

journal homepage: www.elsevier.com/locate/enbuild

Cooling load reduction by using thermal mass and night ventilation

Lina Yang *, Yuguo Li
Department of Mechanical Engineering, The University of Hong Kong, Pokfulam Road, Hong Kong, China

A R T I C L E I N F O

A B S T R A C T

Article history:
Received 6 May 2008
Accepted 26 May 2008

We provide a quantitative understanding of the relationship between thermal mass and cooling load, i.e.
the effect of thermal mass on energy consumption of air-conditioning in ofce buildings. A simple ofcebuilding model with air-conditioning at daytime and free cooling at nighttime is analyzed in detail to
quantify the hourly and overall variation of cooling load of air-conditioning. As an important parameter,
an increase of time constant can effectively reduce the cooling load, by as much as more than 60% when
the time constant is more than 400 h. However, when the time constant is larger than 1000 h, a further
increase may slightly increase the cooling load, as a too large time constant may also postpone the heat
release of thermal mass until the daytime. For the most effective reduction of cooling load, the interior
and exterior convective heat transfer numbers need to be matched.
2008 Elsevier B.V. All rights reserved.

Keywords:
Cooling load
Cooling load ratio
Thermal mass
Night ventilation
Air-conditioning

1. Introduction
Buildings consume more than 30% of the primary energy
worldwide. In China, buildings account for 23% in 2003 of the total
energy use and is expected to increase to 30% by 2010 [1]. 65% of
the building energy consumption in 2003 in China was due to
heating, ventilation and air-conditioning [1]. Hence, improving
building energy efciency has become one of the critical issues for
overall national energy strategy in China.
The use of thermal mass in a building can reduce peak heating
or cooling load, and subsequently building energy consumption, in
particular when it is integrated with night ventilation. Thermal
mass is dened as the thermal materials that can absorb heat, store
it and release it later. Thermal mass includes building envelope,
furniture, internal walls, etc. Thermal storage capacity of building
mass is one of the factors describing the building thermal
performance [2]. In naturally ventilated buildings, thermal mass
is effective for reducing the air temperature uctuation [3].
Many studies investigated the relationship between thermal
mass and indoor air temperature, and the effect of thermal mass
and night ventilation on cooling load; as reviewed by Balaras [4].
16 different simplied models for estimating the cooling load of a
building, considering the buildings thermal mass, were summarized and compared in Ref. [4]. Parameters describing the effects of
thermal mass include the effective heat storage capacity [5,6],
diurnal heat capacity [7], thermal effectiveness parameter [8],

* Corresponding author. Tel.: +852 2816 2625; fax: +852 2858 5415.
E-mail address: liyg@hku.hk (L. Yang).
0378-7788/$ see front matter 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.enbuild.2008.05.014

admittance factor [9], and total thermal time constant [10]. The
effective layer thickness of external walls [11] and the surface area
of thermal storage [12,13] also signicantly affect the thermal
mass performance.
Existing studies showed that the reduction in cooling load by
using thermal mass vary between 18 and 50% [8,1416]. But these
studies were mostly based on the laboratory monitoring or eld
experiments, without systematic theoretical studies. Hence this
paper aims to provide a detailed theoretical analysis on the
relationship between use of thermal mass and reduction of cooling
load. Through a simple building model, all parameters affecting the
thermal mass performance are quantitatively evaluated and
analyzed.
2. A simple building model
Here only the warm climates are considered, while the results
and analysis can also be easily extended to the cold climates. Fig. 1
shows a simple ofce-building model with daytime air-conditioning and night ventilation. The air temperature distribution in the
building is uniform. Both internal and external thermal storage
materials are modeled as a thermal mass wall. All building
envelope except the thermal mass wall is perfectly insulated. As
shown later, the location of thermal mass relative to insulation and
effect of insulation may be analyzed by changing the interior and
exterior convective heat transfer numbers. Thermal radiation
between room surfaces is ignored. All heat gain (including solar
heat gain) and heat generation in the building is lumped into one
heat source term, i.e. E at daytime and no indoor heat gain is
considered at nighttime. The temperature distribution in the

L. Yang, Y. Li / Energy and Buildings 40 (2008) 20522058

2053

Nomenclature
Ai
Ao
cm
cp
E
g
hi
ho
M
qv
Qcl
t
TE
Ti
Tm
To
Tset
To
DTo

interior surface area of thermal mass (m2)

exterior surface area of thermal mass (m2)
heat capacity of the thermal mass (J/kg K)
heat capacity of air (J/kg K)
heat generation rate (W)
acceleration of gravity (m/s2)
interior convective heat transfer coefcient
(W/m2 K)
exterior convective heat transfer coefcient
(W/m2 K)
mass of the thermal mass (kg)
night ventilation rate (m3/s)
cooling load (W)
time (s)
temperature rise due to internal heat gain (K)
indoor air temperature (K)
thermal mass temperature (K)
outdoor air temperature (K)
indoor air setting temperature at daytime (K)
mean outdoor temperature (K)
amplitude of outdoor air temperature uctuation
(K)

Greek symbols
b
phase shift (s)
li
interior convective heat transfer number
lo
exterior convective heat transfer number
j
cooling load ratio
jt
total cooling load ratio
r
air density (kg/m3)
t
time constant (s)
v
frequency of outdoor temperature variation (1/s)

Fig. 1. A simple one-zone building model with periodic outdoor air temperature
variation when: (a) daytime, the air-conditioning (AC) system is on and the indoor
air temperature is kept constant and (b) nighttime, AC is off and the building is
ventilated at a constant ventilation rate.

where DTo and To are independent of time and DTo  0; v is the

frequency of the outdoor temperature uctuation with a value of
2p/24 h1.
Substituting Eq. (3) into Eq. (1), we get

thermal mass is also assumed to be uniform, i.e. Tm. This means

that the thermal conduction process within the materials is much
faster than thermal convection at surface. At daytime, the indoor
air temperature is kept constant, i.e. Tset; see Fig. 1a. At nighttime,
the ventilation rate, qv , is constant; see Fig. 1b.
2.1. Daytime
The heat balance equations for thermal mass and room air are
Mcm

@T m
ho Ao T m  T o hi Ai T m  T i 0
@t

Q cl hi Ai T m  T i E 0

(1)

(2)

where Qcl is the heat removed by air-conditioning equipment, i.e.

the cooling load.
We assume that the outdoor temperature can be expressed by
Fourier analysis as the sum of sinusoidal components of periods 24,
12, 8, 6 h, etc. We consider the main sinusoidal component of
period 24 h.
T o To DTo sinvt

(3)

vt

@T m
lo li T m lo To li T i lo DTo sinvt
@vt

(4)

where t Mcm =rcp qv is the time constant based on a reference

ventilation rate qv , which is chosen to be the night ventilation ow
rate. lo ho Ao =rcp qv is the non-dimensional exterior convective
heat transfer number, and li hi Ai =rcp qv the interior convective
heat transfer number.
Two convective heat transfer numbers li and lo measure the
relative strength of interior and exterior convective heat transfer at
the thermal mass surfaces. A small thermal resistance (large
interior or exterior convective heat transfer number) represents
the interior or exterior convective heat transfer is very effective
compared to the ow mixing in the room. In such situations, the
thermal mass is considered to be in thermal equilibrium with the
room air or outdoor.
The general solution of Eq. (4) is
T m vt

lo
li
T
T
lo li o lo li set
lo DTo

q sinvt  b1
lo li 2 v2 t 2
C 1 elo li =vtvt

(5)

2054

L. Yang, Y. Li / Energy and Buildings 40 (2008) 20522058

where Tset is the indoor temperature set point, which is maintained

by the air-conditioning system (Ti = Tset = constant at daytime),
and C1 is a constant. For the thermal mass, the phase shift
b1 = tan1(vt/(lo + li)) with a value between 0 and p/2, i.e. 0 and
6 h. Here the warm climates are dened as To  T set , however, the
analysis can also be extended to the cold climates (i.e. To < T set ).
Substituting Eq. (5) into Eq. (2) and we have the cooling load

mass.
R vt2

jt R vvtt21

jQ cl jdvt

(9)

vt1 jQ cl0 jdvt

2.2. Nighttime

9
8
>
>
=
<
li lo
li lo
Q cl vt rcp qv T E
To  T set q DTo sinvt  b1 C 1 li elo li =vtvt
>
>
l

l
2
o
i
;
:
lo l v2 t 2

(6)

where T E E=rcp qv is the temperature rise due to the heat source.

A negative/positive Qcl means that cooling/heating is needed.
Eq. (6) shows that the cooling load is the sum of three parts. The
rst is the mean load Q cl , which comprises the steady state cooling
load due to steady heat source, and the heat gain or loss through
convective heat transfer due to the temperature difference
between mean outdoor and indoor air temperature (as thermal
conduction through the thermal mass is assumed to be innite).
The second part is the periodic uctuating load DQ cl with its
amplitude depending on the outdoor air temperature swing DTo ,
the convective heat transfer number (li and lo), and the time
constant t. b1 is the phase lag of the cooling load swing with
respect to the outdoor air temperature; and the third part (the
exponent term) is due to initial condition and it decays to zero as
time or exterior convective heat transfer increases, or time
constant decreases.
The cooling load increases as heat source (TE), or temperature
differences between indoor and mean outdoor air To  T set , or
increase. On the other
uctuation of outdoor air temperature DT
hand, the time constant t and the convective heat transfer
coefcient with surrounding air (li and lo), are also important
parameters affecting the cooling load. Undoubtedly, the peakcooling load can be reduced with heavy thermal mass, i.e. large
time constant t. The phase shift of cooling load is b1 = tan1(vt/
(lo + li)) with a value between 0 and p/2 (i.e. 0 and 6 h).
When there is no thermal mass, i.e. t = 0, we can easily obtain
the cooling load Qcl0:


li lo
li lo
To  T set
DTo sinvt
Q cl0 vt rcp qv T E1
lo li
lo li
(7)
The effect of thermal mass can be seen from the lack of the
time constant terms in Eq. (7). A cooling load ratio j = Qcl/Qcl0 is
dened here to represent the ratio of cooling load in building
with thermal mass to the load in building without thermal mass
at daytime.

The two heat balance equations for the thermal mass and room
air are
Mcm

@T m
ho Ao T m  T o hi Ai T m  T i 0
@t

rcp qv T o  T i hi Ai T m  T i 0

(10)

(11)

Let li li =1 li , we have

vt

@T m
lo l0i T m lo l0i To lo l0 DTo sinvt (12)
@vt
The general solution of Eq. (12) is

lo l0i
DTo sinvt  b2
T m vt To q
2
lo l0i v2 t 2
0

C 2 elo li =vtvt

(13)

where C2 is a constant. The phase shift b2 tan1 vt =lo l0i .

The solution for indoor air temperature is
Tvti To

l0i
lo l0i

DT sinvt l0i q
li o
2
lo l0i v2 t 2
0

DTo sinvt  b2 l0i C 2 elo li =vtvt

(14)

2.3. Matching conditions

We assume that in 24 h, the daytime is from t1 to t2. The thermal
mass temperatures at t2 and t1 + 24 should be continuous.
At t = t2, we have

q
2
T E1 li lo =lo li To  T set li lo = lo li v2 t 2 DTo sinvt  b1 C 1 li elo li =vtvt
Q cl
j

Q cl0
T E1 li lo =lo li To  T set li lo =lo li DTo sinvt

The cooling load ratio is a time-dependent parameter. The

smaller the ratio, the more energy saved. The cooling load ratio
decreases when the heat source E, temperature difference between
indoor and outdoor air To  T set , or the outdoor temperature
swings DTo decrease.
To understand the total energy conservation potential, the total
cooling load ratio is introduced to represent the ratio of total
cooling load in building with thermal mass to that without thermal

(8)

0
lo li
DTo sinvt 2  b2 C 2 elo l i =vtvt2
To q
0 2
2
2
lo li v t

lo
li
lo DTo
T q
T
lo li o lo li set
2
l l v2 t 2
o

lo li =vt vt2

 sinvt 2  b1 C 1 e

(15)

L. Yang, Y. Li / Energy and Buildings 40 (2008) 20522058

2055

Fig. 3. Proles of cooling load ratio and total cooling load ratio as a function of the
temperature differences between indoor and outdoor air To  T set : (a) cooling load
ratio; (b) total cooling load ratio.

Fig. 2. The temperature proles of indoor air and thermal mass with different mean
outdoor air temperatures: (a) To 308:15 K; (b) To 303:15 K; (c) To 297:15 K.

and at t = t1 + 24, we have

lo l0i
DTo sinvt 1 24v  b2
To q
2
lo l0i v2 t 2

3.1. Outdoor air temperature and cooling load

C 2 elo li =vtvt1 24

lo
lo li

To

lo DTo
T q sinvt 1
lo li set
2
lo li v2 t 2
li

 b1 C 1 elo li =vtvt1

(mean outdoor temperature To and temperature swing DTo ), the

interior and exterior convective heat transfer numbers (li and lo),
and as well as the time constant t.
For simplicity, we let t1 = 8:00, t2 = 18:00, v = p/12, room
volume V = 32.4 m3 (4 m  3 m  2.7 m), lo = 30 (exterior convective heat transfer coefcient ho = 15 W/m2 K), li = 10 (interior
convective heat transfer coefcient hi = 5 W/m2 K), t = 100 (common concrete brick), DTo 7:5 K, E = 240 W (20 W/m2),
Tset = 297.15 K (24 8C), and the reference night ventilation rate is
1
given by qv =V 1 h . Then the phase shift b1 is about 2.21 h at
daytime and b2 = 2.68 h at night. The variation of cooling load can
be obtained by giving different outdoor climates, i.e. To 308:15 K
(35 8C), To 303:15 K (30 8C); To 297:15 K (24 8C, the reference
case).

(16)

The two constants C1 and C2 are determined from Eqs. (15) and
(16). The expressions are lengthy and are not given here.
3. Results and discussion
As in Eqs. (8) and (9), the control parameters for the cooling load
ratio include indoor air setting Tset, the outdoor air temperature

The temporal proles of indoor/outdoor air temperature and

thermal mass temperature are obtained for three different mean
outdoor air temperatures; see Fig. 2. The maximum temperature of
thermal mass at daytime occurs at around 16:00, then reach
another maximum at 19:00 after the air-conditioning is turned off
and if the outdoor air temperature was higher than indoor
temperature during this period. Heat is released from the thermal
mass at nighttime and thermal mass achieves its minimum
temperature at around 4:00 in the following day. At nighttime the
indoor air temperature reaches its minimum at around 3:00 while
the minimum outdoor air temperature occurs at around 2:00. Also
there is a gap for indoor air temperature between daytime and

2056

L. Yang, Y. Li / Energy and Buildings 40 (2008) 20522058

Fig. 4. Proles of cooling load ratio and total cooling load ratio as a function of
outdoor air temperature swing: (a) cooling load ratio; (b) total cooling load ratio.

nighttime at the time 8:00 and 18:00 when the air-conditioning

system is on or off. The gap is smaller when the mean outdoor air
temperature is smaller.
Fig. 3 shows the cooling load ratio and total cooling load ratio as
affected by To  T set . The higher the temperature difference, the
larger the cooling load ratio and total cooling load ratio. In Fig. 3a,
the cooling load ratio is less than one for most of the hours due to
heat storage. After the heat storage capacity of the thermal mass
achieves its maximum at around 16:00, it begins releasing its heat,
leading to an increase of the cooling load and the cooling load ratio
becomes greater than one after 16:00. The slightly excessive
energy consumption at later hours is tiny, which can be offset by
the energy saving at earlier hours. Hence the total cooling load
ratio can be less than one. The total cooling load ratio decreases as
To  T set difference reduces. The maximum energy saving of close
to 30% is obtained when To  T set 0 K. It is noted that when
To  T set 0 K, the cooling load ratio is negative at early hours as
the night ventilation reduces the indoor air temperature to be
below the setting temperature of 24 8C. In this case the heating is
needed rather than cooling.
Large diurnal temperature variations benet thermal mass.
Fig. 4 shows the effect of three different temperature swings of 5,
7.5, and 10 K. The cooling load ratio and total cooling load ratio
decrease as the temperature uctuation increases, as the
amplitude of cooling load is proportional to the outdoor
temperature swing as shown in Eq. (6). When the swing is larger,
the reduced amplitude in the case with thermal mass is larger than
that without thermal mass; see Eq. (8). The effect of indoor heat
gain can also be shown (not done here). It is obvious that the
cooling load is dominated by the heat gain when the daytime heat
gain is large; hence the relative contribution of thermal mass is
reduced.

Fig. 5. Proles of cooling load ratio and total cooling load ratio as a function of time
constant: (a) cooling load ratio; (b) total cooling load ratio.

3.2. Time constant/convective heat transfer and cooling load

The time constant t is a very important parameter for
describing the thermal properties of thermal mass. The magnitude
of the time constant determines the heat storage capacity of
thermal mass and phase shift of the peak-cooling load. The effect of
the time constant t on the cooling load ratio and total cooling load
ratio is shown in Fig. 5.
The cooling load ratio and the total cooling load ratio decrease
rst and then increase as the time constant increases. When t = 0,
both ratios equal to one. When the time constant t increases, the
phase shift of thermal mass for both day and night time increases
as time constant increases, hence the phase shift for the cooling
load increases correspondingly, see Eq. (6). The peak cooling load is
delayed and the energy saving is realized. On the other hand, since
the amplitude of cooling load with thermal mass (see Eq. (6)) is
inversely proportional to the time constant, the cooling load ratio
and total cooling load ratio decrease as the time constant increases.
That is why the total cooling load ratio reduces when t < 1000 in
Fig. 5. However, there is a slight increase in the total cooling load
ratio when t > 1000 and increases to innity. This may be due to
the fact that at nighttime, the heat release process of thermal mass
also slows down due to the very large time constant but relatively
smaller night ventilation, which may lead to that the thermal mass
is kept at high temperature at night, and hence slightly increased
cooling load next day. Hence due to the phase shift approaches the
maximum value of p/2, i.e. 6 h for the time constant to approach
innity; there is minor increase for the total cooling load ratio
when the time constant is greater than 1000.

L. Yang, Y. Li / Energy and Buildings 40 (2008) 20522058

2057

Fig. 6. Proles of cooling load ratio and total cooling load ratio as affected by the
interior convective heat transfer number: (a) cooling load ratio; (b) total cooling
load ratio.
Fig. 7. Proles of cooling load ratio and total cooling load ratio as affected by the
exterior convective heat transfer number: (a) cooling load ratio; (b) total cooling
load ratio.

It is obvious that when there is a mismatch between the two

convective heat transfer numbers, an increase of any one of them
will not make any signicant impact. Figs. 6 and 7 show that
neither small nor large convective heat transfer numbers reduce
the cooling load ratio. For interior convective heat transfer number
li, the zero value of this number means there is no heat transfer
between indoor air and thermal mass. In this case, there is no
thermal mass in building and the cooling load ratio or the total
cooling load ratio are equal to 1; see Fig. 6. A zero exterior
convective heat transfer number lo means there is no heat transfer
between outdoor air and thermal mass. In this case, only interior
thermal mass contributes.

Acknowledgments
The work was supported by a grant from the Research Grants
Council of Hong Kong SAR, China (Project No. 7154/05E). The
authors thank Dr. Pengcheng Xu for his assistance in mathematical
derivations. The work is a part of the International Energy Agency
(IEA) Annex 44 project on Integrating Environmentally Responsive
Elements in Buildings.

References
4. Conclusions
The effect of thermal mass on cooling load reduction in
buildings is studied in detail using a very simple building model,
which allows us to examine the hourly benets in using thermal
mass and night ventilation. The cooling load ratio and the total
cooling load ratio are introduced to represent the effect of thermal
mass on cooling load reduction. Our analysis quantied the
dependence of the cooling load on the thermal properties of
thermal mass, including the time constant t and the convective
heat transfer factors (both interior number li and exterior number
lo), the outdoor air temperature, and as well as the indoor heat
gain. Our results show that only appropriate amount of thermal
mass in terms of both thermal properties and convective heat
transfer together with suitable outdoor climates will benet most.
The present work provides a simple model for designing thermal
mass and night ventilation.

[1] Y. Wu, Chinese building energy conservation: existing situation, problems and
policy, in: Presentation on the International Conference on Sustainable Development in Building and Environment, Chongqing, China, 2003.
[2] ASHRAE, ASHRAE HandbookHVAC Applications, American Society of Heating,
Refrigerating and Air-Conditioning Engineers, Inc., Atlanta, 1999.
[3] H. Asan, Y.S. Sancaktar, Effects of walls thermophysical properties on time lag and
decrement factor, Energy and Buildings 28 (2) (1998) 159166.
[4] C.A. Balaras, The role of thermal mass on the cooling load of buildings. An
overview of computational methods, Energy and Buildings 24 (1) (1996) 110.
[5] E.H. Mathews, P.G. Rousseau, P.G. Richards, C. Lombard, A procedure to estimate
the effective heat storage capability of a building, Building and Environment 26
(2) (1991) 179188.
[6] K.A. Antonopoulos, E.P. Koronaki, Apparent and effective thermal capacitance of
buildings, Energy 23 (3) (1998) 183192.
[7] J.D. Balcomb, Heat storage and distribution inside passive solar buildings, in:
Proceedings of the Second International PLEA Conference, 1983, pp. 547561.
[8] M.D. Ruud, J.W. Mitchell, S.A. Klein, Use of building thermal mass to offset cooling
loads, ASHRAE Transactions 96 (pt. 2) (1990) 820829.
[9] P. Burberry, Practical Thermal Design in Buildings, Batsford Academic and
Educational Ltd., London, 1983.
[10] B. Givoni, Man, Climate and Architecture, 2nd ed., Applied Science, London, 1981.

2058

L. Yang, Y. Li / Energy and Buildings 40 (2008) 20522058

[11] K.A. Antonopoulos, E.P. Koronaki, Thermal parameter components of

building envelope, Applied Thermal Engineering 20 (13) (2000) 1193
1211.
[12] M.S. Sodha, J. Kaur, R.L. Sawhney, Effect of storage on thermal performance
of a building, International Journal of Energy Research 16 (1992)
697707.
[13] E. Shaviv, A. Yezioro, I. Capeluto, Thermal mass and night ventilation as passive
cooling design strategy, Renewable Energy 24 (3/4) (2001) 445452.

[14] M. Brown, Optimization of thermal mass in commercial building applications,

Journal of Solar Energy Engineering, Transactions of the ASME 112 (4) (1990)
273279.
[15] T. Kusuda, J. Bean, Comparison of calculated hourly cooling load and indoor
temperature with data for high mass building tested in an environmental
chamber, ASHRAE Transactions 87 (1981) 12321240.
[16] D. Burch, S. Malcolm, K. Davis, The effect of wall mass on the summer space
cooling of six test buildings, ASHRAE Transactions 90 (1984) 521.