Performance evaluation and energy conservation potential

of earth-air-tunnel system coupled with non-air-conditioned building

Rakesh Kumara>*, S. Ramesh b , S.C. Kaushik a
aCentre for Energy Studies, Indian Institute of Technology, New Delhi-110016, India
1
Municipal Corporation of Delhi, New Delhi-110001, India
Received 16 October 2002; received in revised form 28 November 2002; accepted 31 January 2003

Abstract
A numerical model to predict energy conservation potential of earth-air heat exchanger system and passive thermal performance of
building has been developed. This model improves upon previous studies by incorporating e3ects of ground temperature gradient, surface
conditions, moisture content and various design aspects of earth-air-tunnel (EAT). The model is based on simultaneously coupled heat
and mass transfer in the EAT and is developed within the scope of numerical techniques of Finite Di3erence and FFT (Matlab). The model
is validated against experimental data of a similar tunnel in Mathura (India), and is then used to predict the tube-extracted temperature
for various parameters such as humidity variations of circulating air, air9ow rate and ambient air temperature. The model is found to be
more accurate in predicting tube extracted temperature variations along the length (error range 1 : 6%).
These results are further used to study the thermal performance of a non-air-conditioned building. Cooling potential of 80-m earth
tunnel is found adequate (19 kW) to maintain an average room temperature 27.65C. However, auxiliary energy load of 1:5 kW for winter
season is required in achieving comfort conditions with EAT system a3ecting an average room temperature of 24.48C. The present
model can be easily coupled to di3erent greenhouse and building simulation codes.

Keywords: Earth-air heat exchanger; Soil temperature; Cooling/heating capacity; Auxiliary energy load; ACH (air change rate); Space cooling

1. Introduction

The use of earth to air heat exchangers has gained a

signiAcant ground in heating and cooling of buildings [1,2]
due to considerable increase in energy consumption and inherent expenditure of buildings. An earth-air-tunnel (EAT)
system suitably provides heating and cooling energy loads
of building based upon seasonally varying inlet temperature and tunnel-wall temperature which depends on ground
temperature. It was reported by Bharadwaj et al. [3] that
the temperature of the ground varies annually up to a depth
of 4 m and remains constant beyond this depth. The performance of an EAT system depends upon the temperature and
moisture distribution in the ground as well as on the surface conditions. Various simpliAed models presented in the

literature [4-11], report a reasonable prediction of tubeextracted temperature. However, these models are characterized by limited scope of assumptions, as they do not
consider the latent heat transfer between air and pipe-wall
and hence cannot predict the humidity variations of tunnel
air. Non-inclusion of latent heat not only a3ects temperature
calculations but also limits the scope of determining comfort conditions of the building signiAcantly. Furthermore,
most of the models ignored heat transfer phenomenon in the
earth, both sensible and latent heat transfer, or considered
an axially symmetric heat 9ow into the ground with constant pipe-wall temperature. Incorporating these factors, in
the model accommodates natural thermal stratiAcation in the
ground, which leads to more accurate predictions. Whereas,
constant wall temperature excludes any possibility of determining the long-term dynamic performance of an EAT
system.
Therefore, this paper presents a more accurate transient
model based on coupled and simultaneous transfer of heat
and mass into the soil and the circulating air. Natural
thermal stratiAcation is accounted in the ground. The soil

R. Kumar et al /Building and Environment 38 (2003) 807-813

808

Nomenclature
speciAc heat capacity of ground
speciAc heat of air
radial incremental step
isothermal di3usivity of moisture in vapor
form
thermal moisture di3usivity
isothermal moisture di3usivity
heat transfer coeKcient
thermal conductivity of pipe

Cp
Ca
dr
Du; vap

DT
h
Kp

K
I
Nr
Nl
Nt
Pa
R
r
Ta

boundary conditions are applied at the ground surface. Also,

the analysis of inter-related phenomenon of humidity variations and temperature magnitude helps in accounting for
latent and sensible heat transfer and thus predicting pipe outlet temperature and comfort conditions of the building more
accurately.
The proposed model is validated with a wide range of
experimental data taken from similar tunnel in Mathura
(India). The extensive experimental investigations on existing EAT system for the summer and winter seasons
were carried out by Sodha et al. [5,10]. The present model
consisting of transient axisymmetric EAT system is capable
of predicting the humidity and ground temperature variations accurately. Finally, the proposed model is coupled
with building model to predict the thermal performance of
the building and comfort conditions.

Finite di3erence techniques are used to determine temperature at subsequent nodes. The temperature at the (n + 1 )th
node, T(n + 1;i;j), may be calculated from the knowledge
of T(n; i;j) after the time interval AT. Where i denotes the
radial node, j denotes axial node and n is the index for time.
The procedure is repeated to obtain the temperature distribution after any desired number of time increments. However,
the selection of parameters is done under constraints to avoid
the problem of numerical instability. It limits the choice of
AT, when Mx and My are Axed. Boundary conditions incorporate energy balance equations. Heat transferred into pipe
wall may be equated to heat loss by the air as follows:

C(Ta(n;j) - T(n;
C =

2. Mathematical formulation
The earth to air heat exchanger is modeled as a transient
axi-symmetric system. The coupled simultaneous heat transfer equation is as follows:

8^

_/

9y

9r

r 9r
r

9r
u

dy V

(1)

v a p 9

9 8
r 9r

18/

dT\
dr

9h

9y

(3)

2%8yNl
hc=r+1=Kplog(R=r)

where C represents the overall thermal conductance of the

complete earth-air heat exchanger system. Undisturbed temperature variation is considered beyond the zone of pipe in9uence at all time.

T(n;i;yk) =

(4)

T(1;i;j) = T0(r);

oy

and the mass transfer equation is given by the relation

9h
dt

9Ta(y)

where yk is the large distance before or beyond the

pipe-length.
The initial conditions that are required to calculate the
initial solution are given as

9y

gPm

thermal conductivity of the soil

moisture heat of vaporization
number of radial shells
number of length increments
number of time increments
density of air
outer radius of pipe
radial distance
moist air temperature

9 /
dv \

9T
dv

(2)
"" T~ \Dr T~
9y V dy
r 9r u
9r
where, h is the humidity content of soil (kg-moisture/kg-soil)
and K the thermal conductivity of the ground.

where T0(r) is the undisturbed soil temperature at a given

radius. Knowledge of undisturbed soil temperature distribution is required for calculation of heat transfer between the
moist earth and the pipe system [3]. The variation of the
temperature depends on the soil conditions viz. ordinary dry
sunlit soil, blackened and glazed soil, etc. [3]. The present
study involves dry sunlit conditions for calculating undisturbed temperature variations in the ground. The di3usivity
values taken in this study were obtained from [13]. The rate

R. Kumar et al /Building and Environment 38 (2003) 807-813

of heat lost by air to ground is given as

(5)
where m, the mass 9ow rate of air is in Kg/s, C^v is the speciAc heat of air in Jkg^ 1 K" 1 , Ta(n;Nl) is the temperature
at the outlet and Ta(n; 1) is the temperature at the inlet of the
earth-air heat exchanger. On integrating heat 9ux for 24 h,
the cooling/heating potential of EAT system is determined.
The room temperature distribution is governed by onedimensional heat conduction equation [5]. The resultant heat
9ux was incorporated in building code [6] in Matlab using
FFT.
The constant parameters are as follows:
K p = 1:87 W / m - K ,
K = 0:5 W=m - K;
Cp = 2000 J=Kg - K;
C a =1000 J / K g - K ,
lg = 5 4 0 x 4 . 2 x 1000 J=Kg;
pa = 1 Kg=m3;
pm = 1000 Kg=m3;
ps = 2050 Kg=m3:
A room size o f 4 m x 4 m x 4 m with occupancy of
4 people was considered. An air change rate of 5 h^ 1 was
taken for a window-area of 2 sq m. The predicted results of
the 80-m length of tunnel with an inlet area of 0:53 m2 and
an air 9ow velocity of 4:9 m=s, were matched with experimental data with similar surface areas regardless of tube
inlet geometry.

3. Results and discussion

The accuracy of the presented model is veriAed on the
basis of a comprehensive set of experimental data. The performed simulations extend over validation and parametric
study of tunnel with cooling/heating loads, ground temperature, humidity and building conditioning. Fig. 1 shows the
validation of hourly variations in predicted tube extracted
temperature against the experimental results as well as theoretical results obtained by Sodha (1983). There is an excellent agreement between the predicted and experimental
data within an error of 3.9% for 20-m length, which progressively reduced to 1.6% (error range 0.35C) for 80-m
tunnel length compared to an error of 6% as per theoretical
prediction by Sodha et al. Taking account of temperature
distribution in ground and progressive heating of ground
layers near pipe surface are resulting in better estimation
of tube extracted temperature in the present analysis. The
latter model neglected progressive heating of ground layers
due to heat transfer from circulating air to pipe and hence

809

erroneous percentage is large. Fluctuations in air-temperature

at the pipe outlet follow the 9uctuations in the ambient
temperature but are reduced greatly in amplitude. The outlet
air temperature for 80 m earth-tunnel varied in the range of
23.8-27.9C. The performed simulations included four different tunnel lengths of 20, 40, 60 and 80 m while the pipe
radius, air9ow rate and the soil depth remain unchanged. It
was determined that a reduction of tunnel-length from 80 to
20-m increased the average temperature at the pipe's outlet
by 3.6C.
For further validation of the new model, earth temperature
varied for a continuous application of tunnel system a3ecting pipe outlet temperature. Fig. 2a presents the long-term
dynamic behaviour of the pipe outlet temperature showing
similar trends as predicted by modeling results of Sodha et
al. and Santamouris et al. [12]. The constant air circulation
experiment is run for 5 days (120 h). As can be seen from
the Fig. 2a, there is a gradual increase in the tube-extracted
temperature due to continuous heating of ground considering its poor conductivity in transmitting heat. Constant usage of tunnel for 120 h caused a rise of nearly 1.5-2C in
outlet temperature.
Diurnal 9uctuations in soil temperatures as predicted by
the new model are shown in Fig 2b. The soil temperature
near pipe entry at 3.00 P.M. reached 37.3C and decreased
exponentially to nearly 24.7C within a short length (12 m)
from the entry point. The soil temperature after 24 m is not
a3ected by 9uctuations in inlet temperature. Hence longer
pipe lengths are recommended for realizing more cooling
energy from EAT system.
Figs. 2(c) and (d) show the predicted cooling and heating capacity of EAT system. Results of the present model
are compared with the simulation of Sodha et al. It is
evident from the Agure that the 80-m long portion of tunnel
has a daily average cooling potential of 456 kWh (which
corresponds to a mean of 19 kW) and heating potential of
296 kWh. While, no cooling was observed at 5.00 A.M.,
the maximum e3ective cooling load of 43 kWh could be
realized at 3.00 P.M. In winter (December 30), a maximum
heating load of 27 kWh was supplied by EAT system. Also,
the circulating tunnel air did not receive any gain from the
earth between 12.00 P.M.-19.00 P.M. since outdoor temperature exceeded over soil temperature. Further, it required
an extra heating load of 35 kWh for maintaining comfort
conditions (averaged to 5 kWh each hour) during these 7 h.
In Fig. 1, the validity of the model was shown by matching
tube extracted temperature (Tunnel outlet temperature) with
practical experimentation results. We can also substantiate
our validation by matching tube temperature along length
with theoretical results obtained by Sodha et al. [3] for such
a variation. The temperature of tube air was predicted along
length and validated for summer and winter season (Fig. 3a)
for a 10 m tunnel length. An error range of 6 : 5% is noted
in accurately predicting tube temperature along length.
Based upon the procedure developed for coupling
earth-tunnel with building, simulation curves for building

R. Kumar et al /Building and Environment 38 (2003) 807-813

810

Experimental
Sodha et al. (1983)
Predicted results

35

Experimental
Sodha et al. (1983)
Predicted results

34
33

31
<D

30

u 29
% 28
27
26
25
10
15
Time(Hrs)

(a)
60-m EAT

20

10
15
Time(Hrs)

(b)
35

Experimental
Sodha et al. (1983)
Predicted results

80-m EAT

20

Experimental
Sodha et al. (1983)
Predicted results

34
33

32
5- 31
3 30
S. 29
^

28
27
26

10
(c)

15

20

Time(Hrs)

(d)

10
15
Time(Hrs)

20

Fig. 1. (a) Validation of tube extracted temperature at 20-m tunnel length. (b) Validation of tube extracted temperature at 40-m tunnel length. (c)
Validation of tube extracted temperature at 60-m tunnel length. (d) Validation of tube extracted temperature at 80-m tunnel length.

conditioning were generated against various tunnel parameters as shown in Fig. 3b-d. The EAT system was found
to be very e3ective in dampening the 9uctuations in ambient
air temperature.
Fig. 3b shows simulations carried out for determining
e3ect of air9ow rate on room-air conditioning for three 9ow
rate values (1.2,2.58,4 kg=s). It is calculated that a decrease
of air9ow rate from 2.58 to 1:2 kg=s lowers the average
room temperature from 27.65C to 25.9C. In general terms,
a decrease in the air 9ow rate increases the contact period
of air with the pipe-wall thus lowering the temperature at
pipe's outlet.
The e3ectiveness of tunnel performance against length is
shown in Fig. 3c. It is evident from the Agure that smaller
lengths (20, 40, and 60 m) could not suKciently reduce
the indoor temperatures in the late night and early morning
hours. Results predicted an average room temperature of
32.1C for 20-m length, which progressively decreased for
80-m tunnel-length to 27.65C.

Finally, a comparison is made between the cooling potential for various tunnel-radii (0.41, 0.48, 0.52, 0.58, 0:7 m)
in Fig. 3d. For di3erent radii, a change in the boundary
conditions is accounted due to variations in the undisturbed
temperature of the ground. An increase of tunnel radius by
0:17 m increased the average room air temperature by 1.5C.
The results show an increase in pipe-outlet temperature (reduced drop in temperature) with increasing radius above a
critical value. Increase of pipe radius in general represents a
reduction in heat transfer coeKcient. However, pipe-outlet
temperature decreased until a critical radius value of 0:52 m
was reached opposing the above-mentioned trend because
the increased heat transfer area a3ected tunnel temperature
more than reduced heat transfer coeKcient. Kaushik et al.
[6] described the same trend however, within the scope of
assumptions, over-estimated the value of critical radius. The
discrepancy is believed due to constant tunnel-wall temperature assumed in [6] irrespective of ground temperature
distribution.

R. Kumar et al. /Building and Environment 38 (2003) 807-813

43

lull

i>

jre (ideg

lull

39

ro

[3

37

(D
Q.

<u 35

33

lulu

31

A A A \
\ \ \ i'^
k ' l >l P KT >

41

\ /

Sc/

X/

50

(a)

25
24

(b)

Hourly Cooling potential of 80m EAT on 15 June, Mathura

40

Cool ing poten

"T

4-

40

60

Length (m)
Heating potential of 80m EAT
Station: Mathura: 30 December
Sodha etal. (1983)
Simulated results

\ J V\

30

25
20

15 z
[

10

Sodha et
Simulated

20

100
Time (Hrs)

35

-3 nn p M
12.00 P.M.
1 n nn A M
6.00 A.M.

Vv

[>

37 r
36
35 L
5
34
33 t-i
"> i\
31
30
8 29 ^ \\
E 28
h- 27 T
26
rat ure (degC)

45

Simulated results
Sodha
Santamouris

811

jn

~^\^J

>.
10

(c)

15

20

Time (Hrs)

ature) Tg
10

15

Time (Hrs)

Fig. 2. (a) Constant air circulation experiment for 5 days. (b) Variations in soil temperature with time along tunnel wall. (c) Cooling potential of EAT.
(d) Heating potential of EAT.

The daily cooling potential of EAT system averaged

to 19 kW. However, a single room additionally required
an auxiliary energy of 1:5 kW in maintaining the average
temperature of 24.48C.

4. Conclusion
A transient implicit numerical model based on coupled
simultaneous heat and mass transfer equations has been developed to describe the thermal performance of earth-air
heat exchanger system. The numerical model incorporates
greater number of parameters than the previous models,
such as humidity variations of circulating air, natural thermal stratiAcation of the ground, latent and sensible heat
transfer, and ground surface conditions, etc. The results indicated very good agreement between the measured results
and model predictions.
Plots determined the e3ect of tunnel length, mass 9owrate and pipe radius on tube extracted temperature and
were validated against practical data. These can be em-

ployed for selection criterion while designing a building

with an EAT system. Earth-air-tunnel system dampened
the 9uctuations in ambient air temperature to a great extent. Variations in pipe outlet temperature reduced from
9.75C with 20-m EAT system to 4C with 80-m EAT
system which is important for room air conditioning.
Longer tunnel lengths were eKcient in more cooling energy
savings.
The long-term dynamic performance of pipe outlet temperature was studied under constant air circulation experiment. However, a signiAcant rise in pipe outlet temperature
was sensed for smaller lengths. Intermittent application of
tunnel system is suggested for achieving comfort conditions
in such cases as conArmed by the results in Fig 3c.
The daily cooling potential of EAT system averaged to
456 kWh and heating potential to 296 kWh. However, a single room further required an auxiliary energy of 1:5 kW for
achieving comfort conditions as per winter climate prevalent on December 30 (Mathura, India).
The analysis also determined the optimum radius for
room-air conditioning. The simulation predicted cooling

R. Kumar et al. /Building and Environment 38 (2003) 807-813

>

42
40

Sodhaetal. (1983)
simulated results (Summer)
Sodhaetal. (1983)
Simulated results (Winter)

Temperal:ure (degC)

812

. 4Kg/s
2.58Kg/s
1.2Kg/s

.A

Tunnel length=80m

/--' \

Ambienttemeperature

38
36

34
32
30

28
26
4

10

Length (m)

(a)

7 7 1 ~ -

15

20

Time (Hrs)

(b)
0-m EAT
60-m EAT
40-mEAT

0.41m radius

Air mass flow rate = 2.58Kg/s 20-mEAT Ambientair temperature

/\

^j

40

4 8

^ 1 T 0.58m radius
A
0.7m radius
\ 4 0 a m b i e n t temperature

o" 38
B. 36

rat

42
41
40
39
o 38
<U 37
0) 36
35
34
<u 33
Q.
E 32
<u 31
30
29
28

Air mass flow rate = 2.58Kg/s

I 32
|

.-/"

30
28
26

10
(c)

34

15

Time (Hrs)

". '. TTTT'T' i i

10

20

15

20

Time (Hrs)

(d)

Fig. 3. Validation of tube temperature along tunnel length. (b) Indoor-air temperature 9uctuations with various mass 9ow rates. (c) Showing the room
air temperature with various tunnel lengths. Room air temperature variations with various tunnel radii.

energy for di3erent radii (0.41, 0.48, 0.52, 0.58, and 0:7 m).
For 0:52 m radius, the cooling energy was found to be
maximum. Increasing pipe radius above a critical value and
air9ow rate increased the outlet temperature, which significantly a3ected the cooling/heating potential of earth air
tunnel.
The numerical predictions of pipe outlet temperature for
various tunnel characteristics are in agreement with the
experimental data to a high accuracy. Results can be applicable in greenhouse and di3erent types of buildings
with earth-air-tunnel systems for the evaluation of passive
thermal performance.

Acknowledgements
We are thankful to the Ministry of Non-Conventional
Energy Sources for its Anancial help in carrying out the
research work. The authors are grateful to Professor M.S.
Sodha and Professor N.K. Bansal for providing the experimental data.

References
[1] Antinucci M, Fleury B, Lopezd Asian J, Maldonado E, Santamouris
M, Tombazis A, Yannas S. Passive and hybrid cooling of
buildings-state of art. Int. J. Solar Energy 1992;11:251-72.
[2] Puri VM. A Anite element analysis of earth tube heat exchangers.
ASAE Paper No. 84-4537, 1986.
[3] Bharadwaj SS, Bansal NK. Temperature distribution inside ground
for various surface conditions. ASHRAE 1983;16(3-B):183-92.
[4] Puri VM. Feasibility and performance curves for intermittent earth
tube heat exchangers. ASAE paper No. 84-4537, 1986.
[5] Kaushik SC, Sodha MS, Bansal PK, Bhardwaj SC. Solar thermal
modelling of a non-airconditioned building: evaluation of overall
heat 9ux. Energy Research 1982;6:143-60.
[6] Kaushik SC, Kumar GS. Performance evaluation of an earth
air tunnel for space heating of a non air-conditioned building.
International Journal of Ambient Energy 1994;15(3):112.
[7] Chandra M, Puri JS. Cooling and heating of buildings with buried
earth-tube heat exchanger: an experimental study. Journal of the
Solar Energy Society of India 2000;10(1):1-10.
[8] Singh SP. Optimization of earth-air-tunnel system for space cooling.
Energy Conversion and Management 1994;35(8):721-5.
[9] Bansal NK, Sodha MS, Bhardwaj SS. Performance of earthair-tunnel system. International Journal of Energy Research
1983;7(4):333.

R. Kumar et al. /Building and Environment 38 (2003) 807-813

[10] Sodha MS, Sharma AK, Singh SP, Bansal NK, Kumar A. Evaluation
of an earth-air-tunnel system for cooling/heating of a hospital
complex. Building and Environment 1985;20(2):1 15-22.
[11] Tzaferis A, Liparkis D. Analysis of the accuracy and sensitivity
of eight models to predict the performance of earth-to-air heat
exchangers. Energy and Buildings 1992;18:35-43.

813

[12] Santamouris M, Argiriou, Vallindras M. Design and operation of a

low energy consumption passive solar agricultural greenhouse. Solar
Energy 1994;52(5).
[13] Incropera FP, Dewitt DP. Textbook on fundamentals of heat and
mass transfer, 4th ed. New York: Wiley, 1996.