International Journal of Energy, Information and Communications

Vol. 2, Issue 1, February 2011

Mathematical Modelling of Change of Temperature in Pulsating

Heat Pipes with Single Loops
Pranab K. Barua
Department of Energy
Tezpur University
Assam, India
[email protected]

D. Deka
Department of Energy
Tezpur University
Assam, India
[email protected]

U. S. Dixit
Department of Mechanical Eng.
Indian Institute of Technology
Guwahati, Assam, India
[email protected]

Abstract
Recovered heat that would have otherwise been wasted can serve useful purposes.
Efficiency of heat pipes used in heat recovery would depend on the mathematical
rules that govern the process of fall of temperature with respect to time. We therefore
need to establish the exact mathematical model followed by such a system. In this
article, we have shown that in pulsating heat pipes with single loops, temperature
decreases exponentially in time. It could not however be conclusively established that
for a fixed diameter of the pipe, fall of temperature depends on the length of the
evaporator section. Finally, it has been found that for a fixed length of the evaporator
section, temperature falls faster for smaller diameter of the pipe.
Keywords: Heat pipe heat exchanger, waste heat recovery.

1. Introduction
Waste heat is heat generated by fuel combustion or chemical reaction, and then wasted
even though it could possibly be reused for some useful purpose. If such waste heat could be
recovered and used for some useful purpose, that would always be of help in terms of
economics. The strategy of recovering this heat depends on the temperature of the waste heat
gases and the economics involved therein [1]. The heat pipe heat exchanger (HPHE) is a very
efficient lightweight compact waste heat recovery system. It is a self-contained passive
energy recovery device.
An HPHE consisting of a bundle of individual heat pipes with vaporizing and
condensing sections at the respective ends is unfit for large-scale needs in industrial
applications. However, modifications were added, and the HPHE has received much attention
since it was launched into industry at the beginning of the eighties [2].
The heat pipe, as a high efficiency heat transfer element, is widely used in the
electronics cooling industry and energy efficiency sectors. They can be embedded with
aluminium heat sinks to enhance cooling efficiency and/or compactness of cooling devices
[3]. Heat pipes are also widely used in energy recovery systems in domestic and industrial
applications, such as in some domestic appliances for improving the efficiency of the drying
cycles. However, the design of the heat recovery systems with heat pipe units is the key to
providing a heat exchanger system to work as efficient as expected [4].

33

International Journal of Energy, Information and Communications

Vol. 2, Issue 1, February 2011

The heat pipe heat exchanger has many advantages in comparison with conventional
heat exchangers, such as large quantities of heat transported through a small cross-sectional
area with no additional power input to the system, less pressure drop of fluid, advanced
maintainability, high reliability, simpler structure and smaller volume. Researches in this line
have been reported by Firouzfar and Attaran [5].
Akachi [6] proposed a loop type heat pipe with a check valve for directing the heat
carrying fluid flow. In his patent, he described 24 different embodiments of loop type heat
pipes. He later developed [7] loop type heat pipes without check valves. Eventually, based on
that idea two-phase thermal control devices called pulsating heat pipes (PHP) were developed
[8].
Mathematical models considering the heat transfer effects on operation of a pulsating
heat pipe with open end was proposed by Zhang and Faghri [9]. They further studied [10]
numerically the oscillatory flow in pulsating heat pipes. Numerical study was also done by
Lin et. al. [11].
Khandekar et. al. [12] experimented on a flat plate closed loop pulsating heat pipe
structures. Khandekar and Groll [13] constructed a closed loop pulsating heat pipe to study
internal hydraulics of the system. They found that gravity effects in systems with low number
of turns. Khandekar and Groll [14] found that internal diameter of the closed loop pulsating
heat pipe, volumetric filling ration of the working fluid, input heat flux, total number of turns,
operational orientation, and thermo-physical properties of the working fluid effect the
performance of a pulsating heat pipe. In the same year Khandekar et. al. [15] presented a
mathematical model of pulsating heat pipe which helps in predicting two-phase flow
parameters in each sub-section of the device.
Gravity assisted heat pipe has found numerous applications in heat recovery systems
([16], [17], [18]). Nearly one-fourth of the energy engines generate dissipating in the form of
exhaust loss energy [19]. If the exhaust gas enters into surroundings directly, it will not only
waste energy but also damage the environment more or less. Air from the carriage to be
heated is introduced into a heat pipe heat exchanger by a fan and warmed by the heat from the
exhaust gas there from. Then, the air flows into the carriage to keep comfortable temperature
[20].
In many circumstances heat otherwise wasted can be recovered and reused. Heat
pipes with excellent thermal merits have emerged as devices that are efficient in transporting
heat [21]. A heat recovery system based on a looped heat pipe was proposed by Lamfon et. al.
[22] to reuse waste heat from a gas turbine engine. Thermal performance of an ordinary heat
pipe has been studied extensively ([23], [24], [25], [26], [27], [28], [16], [29], [30], [31]).
The maximum heat transfer of a two-phase thermosyphon was investigated by Pioro
([32], [33]). Theoretical studies on the steady-state characteristics and stability thresholds on
closed two-phase thermosyphons were presented by Dorban [34]. Gross and Hahne [35]
investigated the heat transfer performance of closed thermosyphons over wide ranges of
pressure and inclination angles. Joudy et. al. [36] also reported improved heat transfer
performance with an internal wall that separated the liquid and vapour streams.
The heat transfer effectiveness of heat pipe heat exchangers was studied by Peretz [37],
who found that the HTE (heat transfer effectiveness) of the heat exchanger depends upon the
HTE of a single heat pipe and the number of rows parallel to the flow. The thermal
performance of heat pipe heat recovery system was investigated by Azad et. al. [27]. A model
for the system was developed to predict the temperature distribution in longitudinal rows of
the heat exchanger.

34

International Journal of Energy, Information and Communications

Vol. 2, Issue 1, February 2011

The effectiveness of gravity-assisted air-to-air heat exchangers was studied by

Wadowski et. al. [38]. Water-to-air gravity-assisted heat pipe heat exchangers were also
investigated ([26], [39]). The overall effectiveness of heat exchanger was calculated for a
wide range of parameters. The design procedure was presented by Azad and Geoola [24], and
models for heat transfer resistances in heat pipes were also given. Optimization of heat pipe
heat exchangers has been studied by Peretz and Horbaniuc [40].
It is well known that water is excellent as a working fluid for heat pipes for its high
latent heat, easy availability, and its high resistance to decomposition and degradation. Water
has been used particularly successfully in copper heat pipes for low temperature applications.
Problems of incompatibility for iron pipes have also been reported ([41], [42], [43], [28], [16],
[44], [45], [46]). The studies confirm that hydrogen is evolved inside iron-water heat pipes, so
that the condenser eventually becomes flooded with this noncondensable gas, making the heat
pipe inoperative.
Water was selected as the working fluid of the heat pipe system developed by Akyurt et.
al. [47]. Because of the problem of incompatibility with iron, copper was selected initially as
the container material. They developed a copper-water heat pipe system using the analysis
presented by Lamfon et. al. [22].
Traditionally, the application of a heat-pipe air-preheater for the dryer is unable to use
its waste heat with a closed system and does not employ the relative humidity in any drying
systems. In the application of closed-loop oscillating heat-pipe with check valves large
quantities of heat are transported through a small cross-section area. The closed-loop
oscillating heat-pipe with check valves is a very effective heat-transfer device invented by
Akachi et. al. [8]. It has a simple structure and fast thermal-response. The closed-loop
oscillating heat-pipe with check valves consists of a long capillary tube bent into many turns,
and the evaporator section, adiabatic section, and condenser section are located at these turns,
with the ends joined to form a closed loop. It incorporates one or more direction-control oneway check valves in the loop so that the working fluid can circulate in the specified direction
only. Miyazaki et. al. [48] studied the oscillating heat pipe with check valves. It was found
that the closed-loop oscillating heat-pipe with check valves has a high rate of heat transfer.
Pipatpaiboon [49] studied the effect of inclination angle, working fluid and the number of
check valves on the characteristics of heat transfer in a closed-loop oscillating heat-pipe with
check valves. It was found that the closed-loop oscillating heat-pipe with check valves
equipped with two check valves has the highest heat-transfer capacity. Rittidech et. al. [50]
studied the closed-ended oscillating heat-pipe air-preheater for energy thrift in the dryer. It
was found, from the experimental results, that the thermal effectiveness increases and the
closed-ended oscillating heat-pipe air-preheater achieves energy thrift. Wu et. al. [51] studied
the application of heat-pipe exchangers for humidity control in an air-conditioning system. It
was observed that this type of heat exchanger can be an advantageous replacement for a
conventional reheat coil, resulting in energy saving and enhancing the cooling capability of
the cooling coils with little or no external energy need.
The principle of operation is similar to that of an air-preheater by the closed-loop
oscillating heat-pipe with check valves, which is widely accepted as the most efficient heattransfer device for high heat-loads [8]. It has the capability of operating in any position and
has operational flexibility. At present, the closed-loop oscillating heat-pipe with check valves
does not sufficiently solve energy problems. So, an improvement of the heat-pipe airpreheater for the dryer is needed.

35

International Journal of Energy, Information and Communications

Vol. 2, Issue 1, February 2011

2. Objectives of the Study

To check for feasibility and efficiency of a heat pipe heat exchanger, it is important to
study the manner in which it helps in removing heat from a system concerned. It is obvious
that the higher the heat transfer rate, the more efficient the heat pipe is. However, first of all
we need to know the rate of heat transfer with respect to time in such a system. If the
differential equation governing heat transfer is established based on experimental data, the
question of efficiency could be answered properly. For that we need to establish the
mathematical model of fall of temperature with respect to time. So in our work, it has been
chosen to establish mathematical models of fall of temperature in the condenser section due to
natural cooling from a certain higher temperature to the room temperature with respect to
time for various diameters of the pipes and for various lengths of the evaporator sections for
single loops. This ultimately translates itself into the heat removal rate from the pipe. In the
present experimental work, the heat pipe has been selected to be of the pulsating type as we
are going to use water as the working fluid in both of its phases viz. liquid and vapor, inside
the same compartment. In other words, our setup is of a homogeneous flow model.
We propose to find answers to the following questions:
i) What is the general expression of falling temperature as a function of time?
ii) What is the effect of the length of the evaporator section of a heat pipe on the
cooling rate?
iii) What is the effect of the diameter of a heat pipe on the cooling rate?
3. Fabrication and Experimental Details
For efficient operation of a heat pipe, numerous parameters should be controlled. As
such, it makes it imperative to conduct experimental investigations as elaborately as it is
required. The first part of the experimentation requires analysis of performance characteristics
of a single loop pulsating heat pipe. Water has been used particularly successfully in copper
heat pipes for low temperature applications. We were interested to study how temperature
falls. For that we needed to fabricate simple heat pipes to serve our purpose only. Indeed, we
were not aiming for construction of efficient heat pipes. Water was selected as the working
fluid of the heat pipe system as was done in the case of Akyurt et. al. [47]. Because of the
problem of incompatibility with iron, they selected copper as the container material, and later
have developed a copper-water heat pipe system using the analysis presented by Lamfon et.
al. [22]. Hence for our case, copper was selected as the material for the pipes as it has a very
high thermal conductivity.
Copper pipes of 3 different diameters viz. inch (= 0.6354 cm.), inch (= 1.2708
cm.) and 5/8 inch (= 1.5885 cm.) were constructed. We shall however present in our
discussions the unit of the diameters in inch, because with such specifications only are they
available in the market. Each pipe was of the shape of a square.
The length of the
evaporator zone varied for each of the diameter types and was alternately selected as 30 cm,
35 cm and 40 cm respectively. We thus have fabricated 9 different heat pipe setups.
Each pipe contains two cutout sections of length 5 cm welded into the main setup. One
of them acts as water inlet while the other acts as an exit route for air pockets. The pipes were

36

International Journal of Energy, Information and Communications

Vol. 2, Issue 1, February 2011

partially filled with water and then the openings of the inlet and the outlet sections were
plugged with a sealing agent and an adhesive tape, the combination of which proved effective
enough.
To reduce loss of heat from the pipes baring the evaporator and the condenser sections,
glass wool was utilized as an insulator. Adhesive tape was used to bind the material with the
pipe. We have already stated that we are interested to establish mathematical models of fall of
temperature in heat pipes. The mathematical models are what we are interested in, and we are
not really interested on the operability of the entire heat exchanger set up. That is why we
have used insulating materials just for minimizing the heat loss without actually evaluating
the critical radius of insulation as it would not really affect the mathematical models. We have
used the software called Daisy Lab for acquisition of data. This software is connected to a
thermocouple to measure the temperature which gets displayed on the computer screen.

4. Statistical Tools to be Used for Data Analysis

We have mentioned in our objectives that we are interested to establish the general
expressions of falling temperature with respect to time in a heat pipe both for one single loop
in the evaporator section. In other words, we are interested in mathematical modeling of
falling temperature as a function of time. For that we would need to use certain statistical
tools to verify the possible acceptability of certain hypothesized mathematical functions
probabilistically. The hypotheses concerned would be framed only after we scrutinize the
observed data with the help of diagrammatic representations. Indeed, for our purpose, we
would need certain very standard statistical procedures to arrive at some data dependent
conclusions.
We would use the Method of Least Squares Estimation to fit the hypothesized curves
based on the observed data. The significance of the estimated values of the parameters
involved in the fitted equations would then be tested statistically using the standard
techniques included in Regression Analysis. After statistical validation of the estimated
values of the parameters, we would find the expected values of falling temperature in every
case concerned. Finally, to validate the functional patterns and to find numerically by what
part the model fits the data, we would take help of the Analysis of Variance for Linear
Regression.

5. Observed Average Temperatures

To fit a mathematical model regarding the general expression of falling temperature as a
function of time we would need sufficient amount of data. To arrive at a parameter dependent
general expression we need to collect time dependent data in equal interval of time on falling
temperature for different diameters of the pipes and for different lengths of the evaporator
section to verify whether that kind of a general expression really works for various
combinations of diameters and in lengths. We therefore collected data in time intervals equal
to 1 minute. In what follows, in Tables 1 to 3, considering diameters inch, inch and 5/8
inch respectively, we are going to compare the fall of observed average temperature for
different lengths of the evaporator section.

37

International Journal of Energy, Information and Communications

Vol. 2, Issue 1, February 2011

Table 1. Average temperature in the condenser section after minutes of

operation: Diameter 1/4 inch, for different Lengths

0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16

Temperature for
Length 30 cm
65
50
41.1
36.5
34.6
33.1
32
31.5
30.9
30.7
30.1
29.9
29.5
29.2

Temperature for
Length 35 cm
62
50
45.6
41.4
39.1
36.8
34.9
34.1
33
32.1
31.5
31.2
30.7
30.4
30
29.9
29.7

Temperature for Length

40 cm
60
50
46.2
43
39.9
37.9
35.6
34.6
33.1
32.3
31.1
30.8
30.4
30.1
29.9
29.7
29.5

Table 2. Average temperature in the condenser section after minutes of

operation: Diameter 1/2 inch, for different Lengths

0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

38

Temperature for
Length 30 cm
75
62
56.6
52.5
50
48
46.4
45
43.1
41.5
40.4
39
38.2
37
36.5
35.9
35.1
34.9
34.4
34.1
33.3
33
32.6
32.1
31.8

Temperature for
Length 35 cm
75
66
59
55.3
52.9
50
48.2
46.6
45.2
43.9
42.6
41.3
40
38.6
37.8
37
36.3
35.6
35
34.5
34
33.5
33.1
32.8
32.4
32.2

Temperature for
Length 40 cm
75
65
60.8
56.4
52.1
50
47.8
46
44.1
42.5
41.6
40
38.8
38.4
37.7
37
36.7
36
35.4
34.8
34.5
34.3
34
33.5
33.2
33

International Journal of Energy, Information and Communications

Vol. 2, Issue 1, February 2011

Table 3. Average temperature in the condenser section after minutes of

operation: Diameter 5/8 inch, for different Lengths

0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40

Temperature for Length

30 cm
80
73.8
67.8
62.7
60.2
58.4
56.8
55.2
54.2
53.1
52.1
51.2
50
48.5
47.3
46.1
44.9
43.8
42.9
42
41.3
40.6
40
39.3
38.6
37.9
37.2
36.6
36.1
35.7
35.2
34.9
34.5

Temperature for Length

35 cm
80
76
73.1
69.5
65.6
62.8
60.2
58.1
56.3
55.3
54.3
53.2
52.2
51.1
50
48.9
47.7
46.4
44.9
43.7
42.5
41.3
40.3
39.5
38.8
38.1
37.5
36.9
36.5
36.1
35.8
35.5
35.2
34.9
34.7

Temperature for Length

40 cm
85
80
76.8
74.1
72.3
70.7
69.2
67.6
65.8
64
62.1
60.3
59.1
58
57
56
55
53.8
52.6
51.3
50
48.9
48.1
47.3
46.4
45.3
44.3
43.2
42.2
41.6
40.8
40
39.5
39.1
38.6
37.9
37.2
36.5
36
35.6
35.2

39

International Journal of Energy, Information and Communications

Vol. 2, Issue 1, February 2011

6. Analysis of the Data

Numerical and statistical analysis of the data collected would now lead us to certain
conclusions related to heat transfer. This is the most important step in fitting data dependent
mathematical models in any situation. First, we would need certain parametric hypotheses
which we would test statistically or otherwise, based on the observed data. The following are
our hypotheses:
i. Hypothesis - 1: We hypothesize that temperature T decreases exponentially in time
following T = C + e , where and are parameters to be estimated, > 0 and < 0,
and C is the controlled temperature beyond which there would be no more cooling
possible.
ii.
Hypothesis - 2: We hypothesize that for a fixed diameter of the pipe, fall of
temperature possibly depends on the length of the evaporator section. This hypothesis
would not however be tested statistically. In this case, we would try to draw
conclusions based on the fitted curves only.
iii.
Hypothesis - 3: We hypothesize that for a fixed length of the evaporator section,
temperature falls faster for smaller diameter of the pipe. This hypothesis too would
not be tested statistically. In this case too, we would try to draw conclusions based on
the fitted curves only.
In effect, we would like to study whether feasibility of using heat pipes would be
dependent on: smallness of the diameter of the pipe used, and smallness of the length of the
evaporator section.
We shall now describe the statistical analysis performed on the data collected. We
would proceed as follows. The hypothesized equation being T = C + e , > 0 and < 0,
we have
loge (T-C) = log e + .
C in our case was 29.0 degrees centigrade, the room temperature. Now, ln and can be
estimated ([52], pp 350 352), using the Method of Least Square Estimation. It may be noted
here that for an exponential decay curve expressed as a probability density function of the
type f (t) = e - t, 0, t0, is better known as the instantaneous failure rate ([52], pp 319).
In our case, in the hypothesized decay curve T - C = e , the positive quantity (-) is
therefore the instantaneous failure rate.
In Tables 4 and 5, we are going to describe the Test of Significance ([52], pp 202) for
the Regression Parameters and Analysis of Variance ([52], pp 360) of the log linear fit
respectively, for pipe diameter inch and Length 30 cm.
Table 4. Test of Significance for the Regression Parameters:
Diameter inch, Length 30 cm
Parameters
Estimated Values
Calculated
Tabulated t ,
|t |
12 d.f.
Intercept loge
3.275531
25.45784
t0.05 =1.782
Slope
-0.33842
20.1175
t0.025=2.179
We would now like to test the null hypothesis H01: = 0.0 against the two sided
alternative hypothesis H11: 0.0. It can be seen that the calculated value of |t | (= 20.1175) is
very much larger than the tabulated value of t (= 2.179) for 12 degrees of freedom at 5% level

40

International Journal of Energy, Information and Communications

Vol. 2, Issue 1, February 2011

of significance ([52], pp 166). Hence we conclude that the hypothesis H01 is not acceptable.
Therefore the alternative hypothesis H11 is true. In other words, in this case 0.0.
We would like to test the null hypothesis H02: loge = 0.0 against the one sided
alternative hypothesis H12: loge > 0.0. It can be seen that the calculated value of |t | (=
25.45784) is very much larger than the tabulated value of t (= 1.782) at 5% level of
significance. Hence we conclude that the hypothesis H02 is not acceptable. Therefore the
alternative hypothesis H12 is true. In other words, in this case loge > 0.0.
Table 5. Analysis of Variance of the Log linear Fit:
Diameter inch, Length 30 cm
Source of
Degrees of
Sums of
Calculated
Tabulated
Variation
Freedom
Squares
Value of F
F0.05, (1,12)
Regression
1
26.05505
404.7118
4.7472
Error
12
0.772551
Total
13
26.82761
The table above shows that the effect due to regression on the total variations is very
highly significant ([52], pp 174) because the calculated value of the F statistic (=404.7118)
is very much higher than the tabulated value of F 0.05, (1, 12) (= 4.7472). We conclude that the
log linear equation fits the data very well. Indeed, it can be seen that the coefficient of
determination is given by (26.05505 / 26.82761)*100 = 97.12 ([53], pp 377). In other words,
97.12% of the variations are due to this mathematical relationship, while the rest 2.88% only
is due to randomness. We therefore conclude that for Diameter inch, Length 30 cm. we
have = 26.45727, and = -0.33842, and therefore
T = 29 + 26.45727 e-0.33842 .
For pipe diameter inch and Length 35 cm., the following are the analyses concerned.
Table 6. Test of Significance for the Regression Parameters:
Diameter inch, Length 35 cm
Parameters
Estimated Values
Calculated
Tabulated t , 15
|t |
d.f.
Intercept loge
3.266205
85.54
t0.05 =1.753
Slope
-0.22985
56.25
t0.025=2.131
It can be seen that for the calculated value of |t | is very much larger than the
tabulated value of t at 5% level of significance. Hence we conclude that the hypothesis H01 is
not acceptable. In other words, in this case 0.0. Similarly for loge the calculated value of
|t | is very much larger than the tabulated value of t at 5% level of significance. Hence we
conclude that the hypothesis H 02 is not acceptable. In other words, in this case log e > 0.0.
Table 7. Analysis of Variance of the Log linear Fit: Diameter inch, Length 35 cm
Source of
Degrees of
Sums of
Calculated
Tabulated F
Variation
Freedom
Squares
Value of F
0.05, (1,15)
Regression
1
21.55421
3188.82
4.5431
Error
15
0.10139
Total
16
21.65561

41

International Journal of Energy, Information and Communications

Vol. 2, Issue 1, February 2011

The effect due to regression on the total variations is very highly significant because the
calculated value of the F statistic is very much higher than the tabulated value of F0.05, (1, 15).
We conclude that the log linear equation fits the data very well. Indeed, it can be seen that the
coefficient of determination is 99.53. In other words, 99.53% of the variations are due to this
mathematical relationship, while the rest 0.47% only is due to randomness. We therefore
conclude that for Diameter inch, Length 35 cm, we have = 26.21167, and = -0.22985
and so
T = 29 + 26.21167 e-0.22985 .
For pipe diameter inch and Length 40 cm. the following are the analyses.
Table 8. Test of Significance for the Regression Parameters:
Diameter inch, Length 40 cm
Parameters
Estimated Values
Calculated
Tabulated t , 15
|t |
d.f.
Intercept loge
3.39498
121.0198
t0.05 =1.753
Slope
-0.25262
84.4749
t0.025=2.131
From the table above, it can be seen that in this case too we can conclude that the
hypothesis H01 is not acceptable. In other words, in this case 0.0. Similarly for loge also
we can conclude that the hypothesis H02 is not acceptable. In other words, in this case log e >
0.0.
Table 9. Analysis of Variance of the Log linear Fit:
Diameter inch, Length 40 cm
Source of
Degrees of
Sums of
Calculated
Tabulated F
Variation
Freedom
Squares
Value of F
0.05, (1,15)
Regression
1
26.03724
7136.017
4.5431
Error
15
0.054731
Total
16
26.09197
The table above shows that the effect due to regression on the total variations is very
highly significant because the calculated value of the F statistic is very much higher than the
tabulated value of F 0.05, (1, 15). We conclude that the log linear equation fits the data very well.
Indeed, it can be seen that the coefficient of determination is 99.79. In other words, 99.79% of
the variations are due to this mathematical relationship, while the rest 0.21% only is due to
randomness. We therefore conclude that for Diameter inch, Length 40 cm. we have =
29.81405, and = -0.25262, and hence
T = 29 + 29.81405 e-0.25262 .
For pipe diameter 1/2 inch and Length 30 cm, the analysis is as follows.

42

International Journal of Energy, Information and Communications

Vol. 2, Issue 1, February 2011

Table 10. Test of Significance for the Regression Parameters:

Diameter 1/2 inch, Length 30 cm
Parameters
Estimated Values
Calculated
Tabulated t , 23
|t |
d.f.
Intercept loge
3.525721
110.3385
t0.05 =1.714
Slope
-0.10462
45.8378
t0.025=2.069
Here too, it can be seen that we can conclude that the hypothesis H01 is not acceptable.
In other words, in this case 0.0. Similarly for loge too we can conclude that the
hypothesis H02 is not acceptable. In other words, in this case loge > 0.0.
Table 11. Analysis of Variance of the Log linear Fit:
Diameter 1/2 inch, Length 30 cm
Source of
Degrees of
Sums of
Calculated
Tabulated F
Variation
Freedom
Squares
Value of F
0.05, (1,23)
Regression
1
14.22907
2101.104
4.2793
Error
23
0.15576
Total
24
14.38483
The table above shows that the effect due to regression on the total variations is very
highly significant because the calculated value of the F statistic is very much higher than the
tabulated value of F 0.05, (1, 23). We conclude that the log linear equation fits the data very well.
Indeed, it can be seen that the coefficient of determination is 98.92. In other words, 98.92% of
the variations are due to this mathematical relationship, while the rest 1.08% only is due to
randomness. We therefore conclude that for Diameter 1/2 inch, Length 30 cm. we have =
33.978263, and = -0.10462, and therefore
T = 29 + 33.978263 e-0.10462 .
For pipe diameter 1/2 inch and Length 35 cm. the analysis is as follows.
Table 12. Test of Significance for the Regression Parameters:
Diameter 1/2 inch, Length 35 cm
Parameters
Estimated Values
Calculated
Tabulated t , 24
|t |
d.f.
Intercept loge
3.618465
165.3127
t0.05 =1.711
Slope
-0.10089
67.1939
t0.025=2.064
Here too, the conclusions are no different. In other words, in this case too 0.0, and log e >
0.0.
Table 13. Analysis of Variance of the Log linear Fit:
Diameter 1/2 inch, Length 35 cm
Source of
Degrees of
Sums of
Calculated
Tabulated F
Variation
Freedom
Squares
Value of F
0.05, (1,24)
Regression
1
14.88787
4515.015
4.2597
Error
24
0.079138
Total
25
14.967

43

International Journal of Energy, Information and Communications

Vol. 2, Issue 1, February 2011

The table above shows that the effect due to regression on the total variations is very
highly significant because the calculated value of the F statistic is very much higher than the
tabulated value of F 0.05, (1, 24). We conclude that the log linear equation fits the data very well.
Indeed, it can be seen that the coefficient of determination is 99.47. In other words, 99.47% of
the variations are due to this mathematical relationship, while the rest 0.53% only is due to
randomness. We therefore conclude that for Diameter 1/2 inch, Length 35 cm. we have =
37.28030, and = -0.10089, and
T = 29 + 37.28030 e-0.10089 .
The following analysis is for pipe diameter 1/2 inch and Length 40 cm.
Table 14. Test of Significance for the Regression Parameters:
Diameter 1/2 inch, Length 40 cm
Parameters
Estimated Values
Calculated
Tabulated t , 24
|t |
d.f.
Intercept loge
3.535868
87.34586
t0.05 =1.711
Slope
-0.09192
33.1023
t0.025=2.064
Here too, we reject the hypotheses H01 and H02. Hence 0.0, and log e > 0.0.
Table 15. Analysis of Variance of the Log linear Fit:
Diameter 1/2 inch, Length 40 cm
Source of
Degrees of
Sums of
Calculated
Tabulated F
Variation
Freedom
Squares
Value of F
0.05, (1,24)
Regression
1
12.35834
1095.76
4.2597
Error
24
0.27068
Total
25
12.62902
The table above shows that the effect due to regression on the total variations is very
highly significant because the calculated value of the F statistic is very much higher than the
tabulated value of F 0.05, (1, 24). We conclude that the log linear equation fits the data very well.
Indeed, it can be seen that the coefficient of determination is 97.85. In other words, 97.85% of
the variations are due to this mathematical relationship, while the rest 2.15% only is due to
randomness. We therefore conclude that for Diameter 1/2 inch, Length 40 cm. we have =
34.32479, and = -0.09192, and
T = 29 + 34.32479 e-0.09192 .
For pipe diameter 5/8 inch and Length 30 cm. the analysis is as follows.
Table 16. Test of Significance for the Regression Parameters:
Diameter 5/8 inch, Length 30 cm
Parameters
Estimated Values
Calculated
Tabulated
|t |
t , 31 d.f.
Intercept loge
3.782269
234.5462
t0.05 1.697
Slope
-0.06439
74.3402
t0.0252.042

44

International Journal of Energy, Information and Communications

Vol. 2, Issue 1, February 2011

Here too, it can be seen that we have to reject the hypotheses H01 and H02. So in this
case too 0.0, and log e > 0.0.
Table 17. Analysis of Variance of the Log linear Fit:
Diameter 5/8 inch, Length 30 cm
Source of
Degrees of
Sums of
Calculated
Tabulated F
Variation
Freedom
Squares
Value of F
0.05, (1,31)
Regression
1
12.40351
5526.461
4.1709
Error
31
0.069576
Total
32
12.47308
The table above shows that the effect due to regression on the total variations is very
highly significant because the calculated value of the F statistic (= 5526.461) is very much
higher than the tabulated value of F 0.05, (1, 31) ( 4.1709). In this case, we have taken the value
of F 0.05, (1, 30) as F 0.05, (1, 31) is not available in the table of F. We conclude that the log linear
equation fits the data very well. Indeed, it can be seen that the coefficient of determination is
given by (12.40351 / 12.47308)*100 = 99.44. In other words, 99.44% of the variations are
due to this mathematical relationship, while the rest 0.56% only is due to randomness. Hence,
it can be statistically concluded that the observed data on temperature follow the negative
exponential law. We therefore conclude that for Diameter 5/8 inch, Length 30 cm. we have
= 43.91557, and = -0.06439, and so
T = 29 + 43.91557 e-0.06439 .
For pipe diameter 5/8 inch and Length 35 cm. too the outcome of the analysis was
similar.
Table 18. Test of Significance for the Regression Parameters:
Diameter 5/8 inch, Length 35 cm
Parameters
Estimated Values
Calculated
Tabulated
|t |
t , 33 d.f.
Intercept loge
3.894716
248.3167
t0.05 1.697
Slope
-0.06541
82.4578
t0.0252.042
Here also we have found = 49.14209, and = -0.06541 acceptable.
Table 19. Analysis of Variance of the Log linear Fit:
Diameter 5/8 inch, Length 35 cm
Source of
Degrees of
Sums of
Calculated
Tabulated F
Variation
Freedom
Squares
Value of F
0.05, (1,33)
Regression
1
15.27195
6799.284
4.1709
Error
33
0.074122
Total
34
15.34607
The table above shows that the effect due to regression on the total variations is very
highly significant because the calculated value of the F statistic is very much higher than the
tabulated value of F 0.05, (1, 33). In this case, we have taken the value of F 0.05, (1, 30) as F 0.05, (1, 33)

45

International Journal of Energy, Information and Communications

Vol. 2, Issue 1, February 2011

is not available in the table of F. We conclude that the log linear equation fits the data very
well. Indeed, it can be seen that the coefficient of determination is 99.51. In other words,
99.51% of the variations are due to this mathematical relationship, while the rest 0.49% only
is due to randomness. We therefore conclude that for Diameter 5/8 inch, Length 35 cm. we
have
T = 29 + 49.14209 e-0.06541 .
For pipe diameter 5/8 inch and Length 40 cm. the analysis is:
Table 20. Test of Significance for the Regression Parameters:
Diameter 5/8 inch, Length 40 cm
Parameters
Estimated Values
Calculated
Tabulated t , 39
|t |
d.f.
Intercept loge
4.041256
244.6794
t0.05 1.684
Slope
-0.05266
74.0968
t0.0252.042
Here too, we found the null hypotheses not acceptable, and hence = 56.89776, and =
-0.05266 are statistically valid.
Table 21. Analysis of Variance of the Log linear Fit:
Diameter 5/8 inch, Length 40 cm
Source of
Degrees of
Sums of
Calculated
Tabulated F
Variation
Freedom
Squares
Value of F
0.05, (1,39)
Regression
1
15.92043
5490.335
4.0848
Error
39
0.113089
Total
40
16.03352
The table above shows that the effect due to regression on the total variations is very
highly significant because the calculated value of the F statistic is very much higher than the
tabulated value of F 0.05, (1, 39). In this case, we have taken the value of F 0.05, (1, 40) as F 0.05, (1, 39)
is not available in the table of F. We conclude that the log linear equation fits the data very
well. Indeed, it can be seen that the coefficient of determination is 99.29. In other words,
99.29% of the variations are due to this mathematical relationship, while the rest 0.71% only
is due to randomness. Hence, it can be statistically concluded that the observed data on
temperature follow the negative exponential law. We therefore conclude that for Diameter 5/8
inch, Length 40 cm. we have
T = 29 + 56.89776 e-0.05266 .
In what follows, in Figures 1 to 3, considering diameters inch, inch and 5/8 inch
respectively, we are going to compare the fall of average observed temperature for different
lengths.

46

International Journal of Energy, Information and Communications

Vol. 2, Issue 1, February 2011

Temperature in Degrees (Centigrade)

70
60
50
30 cm.

40

35 cm.
30

40 cm.

20
10
0
1

9 10 11 12 13 14 15 16 17

Time in Minutes

Figure 1. Fall of Expected Temperature in the condenser section for different

Lengths of the evaporator section: Diameter 1/4 inch

Temperature in Degrees (Centigrade)

70
60
50
30 cm.

40

35 cm.
30

40 cm.

20
10
0
1

11

13

15

17

19

21

23

25

Time in Minutes

Figure 2. Fall of Expected Temperature in the condenser section for different

Lengths of the evaporator section: Diameter 1/2 inch

Temperature in Degrees (Centigrade)

100
90
80
70
60

30 cm.

50

35 cm.

40

40 cm.

30
20
10
0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41
Time in Minutes

Figure 3. Fall of Expected Temperature in the condenser section for different

Lengths of the evaporator section: Diameter 5/8 inch

47

International Journal of Energy, Information and Communications

Vol. 2, Issue 1, February 2011

Tempearture in Degrees (Centigrade)

80
70
60
50

1/4 inch

40

1/2 inch
5/8 inch

30
20
10
0
1

9 11 13 15 17 19 21 23 25 27 29 31 33
Time in Minutes

Figure 4. Fall of Expected Temperature in the condenser section for different

Diameters of the pipe: Length 30 cm

Temperature in Degrees (Centigrade)

90
80
70
60
1/4 inch

50

1/2 inch

40

5/8 inch

30
20
10
0
1

9 11 13 15 17 19 21 23 25 27 29 31 33 35
Time in Minutes

Figure 5. Fall of Expected Temperature in the condenser section for

different Diameters of the pipe: Length 35 cm

Temperature in Degrees (Centigrade)

100
90
80
70
60

1/4 inch

50

1/2 inch

40

5/8 inch

30
20
10
0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41
Time in Minutes

Figure 6. Fall of Expected Temperature in the condenser section for

different Diameters of the pipe: Length 40 cm

48

International Journal of Energy, Information and Communications

Vol. 2, Issue 1, February 2011

In the Figures 4 to 6 above, considering lengths of the evaporator section 30 cm, 35 cm

and 40 cm respectively, we are going to compare the fall of average observed temperature for
different diameters.

7. Conclusions
We are now proceeding towards possible rejection of our three hypotheses stated earlier.
We would test the first hypothesis statistically. The other two would have to be checked based
on possible acceptance of the first hypothesis.
Our first hypothesis was that temperature T decreases exponentially in time following
T = C + e , where and are parameters to be estimated, > 0 and < 0, and C is the
controlled temperature, in our case the room temperature kept fixed using air conditioner,
beyond which there would be no more cooling possible.
In the following table, the expressions of e for different lengths of the evaporator
section and for different diameters of the pipes used are shown.
Table 22. Expressions for T C = e
Length
Diameter
inch
1/2 inch
5/8 inch

30 cm.

35 cm.

40 cm.

26.45727e-0.33842 .
33.97826e-0.10462 .
43.91557e-0.06439 .

26.21167 e-0.22985 .
37.28030 e-0.10089 .
49.14209 e-0.06541 .

29.81405e-0.25262 .
34.32479e-0.09192 .
56.89776e-0.05266 .

From Tables 4, 6, 8, 10, 12, 14, 16, 18 and 20, it is evident that in the log linear fit of the
data, in every case the values of were found to be very significantly different from zero, and
that in every case the values of log e were found to be very significantly larger than zero.
From Tables 5, 7, 9, 11, 13, 15, 17, 19 and 21 it is evident that for every log linear fit,
the effect due to regression was very significant. Indeed, in every case, the coefficients of
determination were found to be extremely high. We conclude that temperature decreases
exponentially in time.
Our second hypothesis was that for a fixed diameter of the pipe, fall of temperature
depends on the length of the evaporator section. We have mentioned earlier that we would try
to draw a conclusion based on fitted curves only.
From Figures 1, 2 and 3, it is not really clear that temperature falls faster or slower for
smaller lengths of the evaporator section. A row wise comparison of the values of the
instantaneous failure rates (-) particularly for diameter inch in Table 22 above, would
show this. It can be seen that for lengthwise changes the values of (-) are decreasing. For the
other two cases, the values of (-) have shown a decrease from 30 cm to 40 cm, though the
decrements are not uniform. In other words, from our data, it can not be conclusively said that
for a fixed diameter of the pipe, fall of temperature depends on the length of the evaporator
section. It may be that in the range of temperature that we have worked in, the picture is not
quite clear. Hence we would choose to remain inconclusive about the acceptability of this
hypothesis.
Our third hypothesis was that for a fixed length of the evaporator section, temperature
falls faster for smaller diameter of the pipe. We have mentioned that we would try to draw a
conclusion based on fitted curves only.

49

International Journal of Energy, Information and Communications

Vol. 2, Issue 1, February 2011

From Figures 4, 5 and 6, it is very clear that temperature falls faster for smaller
diameters of the pipes. A column wise comparison of the instantaneous failure rates (-) in
Table 22 above would clearly show this. We conclude that for a fixed length of the evaporator
section, temperature falls faster for smaller diameter of the pipe.

References
[1] D. A. Reay, Heat Recovery Systems, E. & F. N. Span, London, 1979.
[2] C. J. Tu, Gravity Heat Pipe Heat Exchanger and Its Application to Exhausted Heat Utilization, Zhejiang
University Press, Hangzhou, 1989.
[3] V. G. Pastukhov, F. Yu , C. V. Maidanik, M. A. Vershinin and M.A. Korukov, Miniature loop heat pipes for
electronics cooling, Applied Thermal Engineering, 23 (9), 2003, 11251135.
[4] S. Lin, John Broadbent, Ryan McGlen, Numerical study of heat pipe application in heat recovery systems,
Applied Thermal Engineering, 25, 2005, 127133.
[5] E. Firouzfar and M. Attaran, A Review of Heat Pipe Heat Exchangers Activity in Asia, Proceedings of World
Academy of Science, Engineering and Technology, Vol. 30, 2008.
[6] H. Akachi, Structure of a heat pipe, U.S. Patent, Patent Number 4921041, 1990.
[7] H. Akachi, U.S. Patent, Patent Number 5219020, 1993.
[8] H. Akachi, F. Polasek and P. Stulc, Pulsating heat-pipe. In: Proceedings of the 5th international heat-pipe
symposium, Melbourne, Australia, 1996, 20817.
[9] Y. Zhang and A. Faghri, Heat transfer in a pulsating heat pipe with an open end, International Journal of Heat
and Mass Transfer, Vol. 45, 2002, 755-764.
[10] Y. Zhang and A. Faghri, Oscillatory flow in pulsating heat pipes with arbitrary number of turns, 8th AIAA /
ASME, Joint Thermophysics and Heat Transfer Conference, St. Louis, Missouri, 2002 .
[11] S. Lin, J. Broadbent and R. McGlen, Numerical study of heat pipe application in heat recovery systems.
Available online since 19 August, 2004.
[12] S. Khandekar, M. Schneider, P. Schafer, R. Kulenovic and M. Groll, Thermo fluid dynamic study of flat plate
closed loop pulsating heat pipes, Micro scale Thermophysical Engineering, Vol. 6/4, 2002, 303-318.
[13] S. Khandekar, M. Groll, Pulsating heat pipes: study on a two-phase loop, 13th International Conference on
Thermal Engineering and Thermogrammetry, 2003.
[14] S. Khandekar and M. Groll, An insight into thermo-hydrodynamic coupling in closed loop pulsating heat
pipes, International Journal of thermal sciences, Vol. 43, Issue 1, 2004, 13-20.
[15] S. Khandekar, V. V. S. N. S. Manyam Sanka, M. Groll and Manmohan Pandey, Two-phase flow modeling in
closed loop pulsating heat pipes, 13th International Conference (13th IHPC), Shanghai, China, 2004.
[16] P. D. Dunn and D. A. Reay, Heat Pipes, 2nd ed., Pergamon, New York, 1978.
[17] S. H. Noie-Baghban. and G. R. Majideian, Waste heat recovery using heat pipe heat exchanger for surgery
rooms in hospital, Applied Thermal Engineering, 20, 2000, 12711282.
[18] K.A. Joudy. and A. M. Witwit, Improvements of gravity assisted wickless heat pipes, Energy Conversion and
Management, 41, 2000, 20412061.
[19] Z. S. Shang and P.L. Li, Practical Technique of Heat Pipes, Publisher of Chemical Industry, Beijing, 1988.
[20] Feng Yang , Xiugan Yuan and Guiping Lin, Waste heat recovery using heat pipe heat exchanger for heating
automobile using exhaust gas, Applied Thermal Engineering, 23 , 2003, 367372.
[21] M. Akyurt, M. Beirutty, O. M. AI-Rabghi, Y. S. H. Najjar and T. Alp, Heat pipes for heat recovery, Trans.
Mech. Eng., 1993, 221-234.
[22] N. Lamfon, M. Akyurt, O. M. Al-Rabghi and Y. S. H. Najjar, Thermophysical relationships for waste heat
recovery using looped heat pipes, Int. J. Energy Res., 18. 1994, 633-642.
[23] S. C. Arora, and S. Domkundwar, A Course in Heat and Mass Transfer, J. C. Kapoor, New Delhi, 1987.
[24] G. A. Asselman, and D. B. Green, Heat pipes. Philips Tech. Rev., 33, 1973, 104-113.
[25] E. Azad and F. Geoola, Design procedure for gravity-assisted heat pipe heat exchanger, J. Heat Recov. Sys., 4,
1984, 101-111.

50

International Journal of Energy, Information and Communications

Vol. 2, Issue 1, February 2011

[26] E. Azad and B. M. Gibbs, Analysis of air-to-water heat pipe heat exchanger, J. Heat Recov. Sys., 7, 1987,
351-358.
[27] E. Azad, F. Mohammadieh, and F. Moztarzadeh, Thermal performance of heat pipe recovery system, J. Heat
Recov. Sys., 5, 1985, 561-570.
[28] S. W. Chi, Heat pipe theory and practice, Hemisphere, Washington, DC, 1976, 242.
[29] B. S. Huang and J. T. Tsuei, Method of analysis for heat pipe heat exchangers, Int. J. Heat Mass Transfer, 28,
1985, 553-562.
[30] M. Shiraishi, K. Kikuchi and T. Yamanishi, Investigation of heat transfer characteristics of a two-phased
closed thermosyphon, J. Heat Recov. Sys., 1, 1981, 287-297.
[31] P. Stulc and L. L. Vasiliev, Heat pipe heat exchangers in heat recovery system. J. Heat Recov. Sys., 5, 1985,
415-418.
[32] I. L. Pioro, Maximum heat transfer capability two-phase thermosyphons, Heat Transfer Soy. Res., 15, 1983,
24-31.
[33] I. L Pioro., Correlation of experimental data on the maximum heat and mass transfer in two-phase
thermosyphons, Heat Transfer Soy. Res., 17, 1985, 75-83.
[34] F. Dorban, Steady-state characteristics and stability thresholds of a closed two-phase thermosyphon, Int. J.
Heat Transfer, 28, 1985, 949-957.
[35] U. Gross and E. Hahne, Heat transfer in closed thermosyphon over a wide range of pressures and inclinations,
Ger. Chem. Eng., 9, 1986, 292-299.
[36] K. A Joudy, T. A. Husain and P. M. Abdulmajeed, Improved heat transport of heat pipes with an internal
wall separating liquid and vapor streams, Energy Converg. Mgmt., 31, 1991, 141-148.
[37] R. Peretz, Heat transfer effectiveness of heat pipes, Heat Recov. Sys., 2, 1982, 165-172.
[38] T. Wadowski, A. Akbarzadeh and P. Johnson, Characteristics of a gravity-assisted heat pipe based heat
exchanger, J. Heat Recov. Sys. & CHP, 11, 1991, 69-77.
[39] E. Azad and F. Moztarzadeh, Design of air-to-water co-axial heat pipe heat exchanger, J. Heat Recov. Sys., 5,
1985, 217-224.
[40] R. Peretz, and B. Horbaniuc, Optimal heat pipe heat exchanger design, J. Heat Recov. Sys., 4, 1984, 9-24.
[41] M. Akyurt, Development of heat pipes for solar water heaters, Solar Energy, 32, 1984, 625-631.
[42] M. Akyurt, The heat pipe solar water heater, J. Eng. Appl. Sci., 3, 1986, 23-38.
[43] A. Basiulis, Compatibility and reliability of heat pipe materials, Int. Heat Pipe Conf. 2nd, Eur. Space Agency,
1, 1976, 357-372.
[44] D. A. Reay, The Perkins tube--A noteworthy contribution to heat exchanger technology, J. Heat Recov. Sys.,
2, 1982, 173-187.
[45] D. A. Reay, The European Community energy R & D program--Report on a recent contractors meeting on
heat exchangers and heat recovery, J. Heat Recov. Sys., 3, 1983, 245-271.
[46] S. Seshan and D. Vijayalakshmi, Heat pipes--Concepts, materials, and applications, Energy Convers. Mgmt.,
26, 1986, 1-9.
[47] M. Akyurt, N., J. Lamfon, Y. S. H. Najjar, M. H. Habeebullah, and T. Alp, Modeling of waste heat recovery
by looped water-in-steel heat pipes Int. J. Heat and Fluid Flow, 16, 1995, 263-271.
[48] Y. Miyazaki, F. Polasek and H. Akachi, Oscillating heat-pipe with check valves. In: Proceedings of the 6th
international heat pipe symposium, Chiang Mai, Thailand, 2000, 38993.
[49] N. Pipatpaiboon, Effect of inclination angle, working fluid and number of check vales on the characteristics of
heat transfer in a closed-looped oscillating heat-pipe with check valves, Proceedings of 1st international
seminar on heat-pipe and heat-recovery systems, Kuala Lumpur, Malaysia, 2004, 8387.
[50] S. Rittidech, W. Dangeton and S. Soponronnarit, Closed-ended oscillating heat-pipe air-preheater for energy
thrift in dryer, Appl. Energy, 81, 2005, 198208.
[51] Wu Xiao Ping, P. Johnson and A. Akbarzadeh, Application of heat-pipe exchangers to humidity control in
air-conditioning system, Appl. Therm. Eng., 6:5618, 1997.
[52] Irwin Guttman, Samuel Wilks and John Stuart Hunter, Introductory Engineering Statistics, John Wiley &
Sons, New York, 1982.
[53] Sheldon M. Ross, Introduction to Probability and Statistics for Engineers and Scientists, Academic Press,
London, 2004.

51

International Journal of Energy, Information and Communications

Vol. 2, Issue 1, February 2011

Authors
.

Pranab Kumar Barua completed his Bachelor of Technology in Mechanical Engineering

from Sikkim Manipal Institute of Technology, Gangtok, India, in 2007. He completed his
Master of Technology in Energy Technology from Tezpur University, India, in 2009.
Dhanapati Deka is an Associate Professor in the Department of Energy of Tezpur University,
Assam, India. He did his Bachelor of Science in Agriculture from Assam Agricultural
University, India in 1990, and Masters of Science in the same discipline from the same
university in 1992. Subsequently, he earned his doctorate from Tezpur University. His
research interests lie in the fields of Bio-energy, Biogas Technology, Biomass Gasification,
Energy and Environment.
Uday S. Dixit is a Professor in the Department of Mechanical Engineering of the Indian
Institute of Technology, Guwahati, Assam, India. He completed his Bachelor of Engineering
from the erstwhile University of Roorkee, in 1987. He did his Master of Technology in 1993,
and his PhD in 1998, both from the Indian Institute of Technology, Kanpur, in Mechanical
Engineering. His research interests lie in Design and Manufacturing covering mainly the
fields of Finite Element Modeling, Neural Network, Fuzzy Set Applications, and
Mechatronics.

52