Accepted Manuscript

Title: Modelling and measuring of the thermal properties of

insulating vegetable fibers by the asymmetrical hot plate
method and the radial flux method: kapok, coconut, groundnut
shell fiber and rattan

Author: J.C. Damfeu P. Meukam Y. Jannot

PII: S0040-6031(16)30018-1
DOI: http://dx.doi.org/doi:10.1016/j.tca.2016.02.007
Reference: TCA 77444

To appear in: Thermochimica Acta

Received date: 17-7-2015

Revised date: 25-11-2015
Accepted date: 20-2-2016

Please cite this article as: J.C.Damfeu, P.Meukam, Y.Jannot, Modelling and measuring
of the thermal properties of insulating vegetable fibers by the asymmetrical hot plate
method and the radial flux method: kapok, coconut, groundnut shell fiber and rattan,
Thermochimica Acta http://dx.doi.org/10.1016/j.tca.2016.02.007

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
Modelling and measuring of the thermal properties of insulating vegetable fibers by the
asymmetrical hot plate method and the radial flux method: kapok, coconut, groundnut
shell fiber and rattan

J.C. Damfeu*a, P. Meukama and Y. Jannotb,c

a
Laboratory L3E Energy, Water and Environment. National Advanced School of Engineering,
Yaoundé, Cameroon, P.O. Box 8390 Yaounde
b
Université de Lorraine, LEMTA, UMR 7563, Vandoeuvre-lès-Nancy, F-54500, France
c
CNRS, LEMTA, UMR 7563, Vandoeuvre-lès-Nancy, F-54500
*corresponding authors:damfeuclaude@yahoo.fr

Highlights:

 The thermal effusivity is estimated for different fibers with transient model
 The thermal conductivities are measured with a steady state regime
 The thermal conductivities are in good agreement with two models
 Experimental results are shown to be in good agreement with the completed and
simplified models and with the radial flux method

Abstract
This article presents two methods dedicated to the determination of the thermophysical
properties of natural fibers grown in almost all countries of the world. Experimental
measurements were carried out on four dry natural plant fibers of low density: kapok fibers;
groundnut shell fibers; rattan fiber and coconut fibers. The thermal effusivity has been
estimated by a transient method (the asymmetrical hot plate method). With experimental
measurement of the specific heat capacity by Differential Scanning Calorimeter (DCS) which
made it possible to deduce the thermal conductivity of such fibers. The thermal properties
estimated by this method are then compared to the results obtained by radial flux flow
method, a steady state method that directly measures the thermal conductivity. The
experimental results compared by these two methods are in good agreement (relative error
<5%). The thermal conductivities of kapok fibers (λ = 0.045 W m-1K-1) and coconut fibers (λ
= 0.055 W m-1K-1) obtained all showing that they can be used as substitutes for synthetic
insulating materials such as polyester fibers (λ = 0.045 W m-1K-1) or glass wool (λ = 0.04 W
m-1K-1). Also the thermal conductivities of kapok and coconut fiber measured correspond with

1
the results obtained by other measuring methods in the literature. This can then help us infer
that the quadrupole 1D model developed to estimate the thermal properties of bulk fiber is
valid and can allow a good estimation of thermo-physical properties. Therefore, the thermal
conductivity of groundnut shell fibers and cane fibers are estimated in the order of λ = 0.093
W m-1K-1 and λ = 0.072 W m-1K-1 respectively.

Keywords: transient regime, radial flux method, natural fiber, thermal properties

Nomenclature
T temperature (°C) cp specific heat capacity (J kg-1
K-1)
E thermal effusivity (J m-2°C-1 s-1/2) λ thermal conductivity (W m-1
K-1)
a thermal diffusivity (m2 s-1) h Convective heat loss coefficient (W m-
2°
C)
Rc thermal contact resistance (°C W-1) θ Lapalce transform of temperature (W m-
2°
C)
Φ Laplace transform of heat flux p Laplace
parameter
ρ density (kg m-3) φ heat flux dissipated in the heating element
(W)
 Heat flux density (W m-2) Bi Biot number
e thickness (m)
Subscripts
i insulating blocks h heating element

2
 1. Introduction

With the global warming which is increasingly felt on the planet, one of the solutions to slow
down the growth of this phenomenon is to reduce the emission of greenhouse gases. In the
building sector, greenhouse gases (GHG) emissions is mainly due to the high energy
consumption through the use of devices providing a good thermal comfort in the interior of
the building, but which strongly emit carbon dioxide (CO2). This high consumption rate
observed is due either to poor insulation of the building or the use of materials with high
coefficient value of thermal conductivity. One of the challenges would be to significantly
reduce energy consumption by manufacturing insulating building materials or low grey
energy insulation materials capable of optimally isolating a building [1].
One of the most used materials nowadays for insulating a building is glass fibers. However,
these fibers are not always originally renewable thus, raises the problem of their disposal at
end of life cycle. Glass fibers fabrics also pose health and safety problem. For example, they
cause skin irritations during manual wrought-handling, transportation or processing. One
recommendable ecological alternative to glass fibers would be the use of natural plant fibers.
Some author’s studies [1, 2] have shown that building materials made from plant fibers are a
perfect solution to the problem of environmental impact of materials (gray energy and
emissions) and the reduction of energy consumption of the building. In Europe for example,
one of the most valued vegetable fibers are hemp fibers. Research [3, 4] conducted on this
fiber recently has helped determine its mechanical, thermal and water properties. Samri [5]
and Evrard [6] have shown that concretes containing hemp has very good hygrothermal
performance (λ = 0.046 W m-1 K-1) that help provide a good thermal comfort.
Regarding the Africa, there is a wide variety of natural plant fiber such as cane; kapok;
groundnut shell; rattan; wood; coconut fibers; bamboo; millet pods; palm tree fibers or the
sawdust. In Cameroon, these fibers are available in large quantities in the national territory
but are not valued. Most of these fibers litter the environment and are sources of pollution or
if incinerated, generate large amounts of greenhouse gases responsible for climate change. If
recycled, it could be valorized by incorporating them as aggregates with local building
materials. Several authors have focused on properties of local materials incorporated with
natural fiber insulation. Bal et al [7] has incorporated into the lateritic soil, millet pods in
different proportions. He studied the influence of millet pod content on the thermal
conductivity of the composite material lateritic-millet, but his study, lack prior estimate of

3
thermal properties of millet pods that are not known in the literature review. Meukam [8] was
incorporated into the lateritic soil, ayous fibers. He showed that this mixture optimizes the
insulation capacity of the manufactured composite material. But his study did not equally
estimate the thermal conductivity of ayous fiber. Asangwing [9] incorporated into the lateritic
soil, oil palm fiber to the proportions ranging from 1 % to 12 %. He determined the thermal
conductivity of the material produced not by an experimental method, but with an electronic
device which indicates the thermal conductivity from the application of electrodes of the
device on the manufactured material. Also, in his study, the thermal properties of palm fibers
are not estimated and are not found in the literature review. Mekhemerche [10] made a
composite material by incorporating clay (37 %), sand (40 %) and 3% of Algeria palm fibers.
He arrived in a compressive strength of 3.24 MPa. He showed that bricks made with a
maximum of 3% of palm fibers can improve the thermal insulation of a building. But his
studies also lack prior knowledge of mixture of thermal conductivity of loose fiber. In all
these studies, thermal and mechanical properties of the fibers incorporated as aggregates to
local building materials (lateritic mostly) are not determined and are not known in the
literature review. However, the addition of these fibers improves the insulation capacity of the
composite material used. Therefore, prior knowledge of the thermal and mechanical
properties of fibers incorporated is indispensable since its knowledge may predict the
insulating character of a composite material when incorporated with different aggregates.

Usually, the main thermo-physical parameter that classifies material as an insulating material
is its thermal conductivity. There are several methods for determining the thermal
conductivity with the most frequently used being:
- The hot wire method [11, 12]
- The guarded hot plate method [13, 14]
- The hot-disc method [15, 16]
- The three-layer method [17]
Among these methods, the only one that a priori may identify the thermal conductivity of the
bulk fibers is the hot wire method. The other methods are used for samples, having a specific
geometric shape. The samples studied here is that of bulk and light fiber, a hot wire with a
very small diameter (less than 1 mm) would not be very suitable for measuring the
temperature increase along the wire through the fiber samples. If the contact between the
fibers and the thread is not perfect, it could lead to an erroneous value of the estimated
thermal conductivity.

4
The purpose of this study is the experimental determination of the thermal conductivity of
plant fibers: kapok fibers, coconut fibers, groundnut shell fiber and rattan fibers. The choice
of these fibers is due to the fact that they are the most available fibers and some are even
cultivated in households. Knowing their thermal conductivity will justify the reason of their
future usage to manufacture insulating composite material or for the building of cold rooms to
be used for food stuffs conservations for example.
Two methods are then proposed for measuring the thermal conductivity of these plant fibers.
The first methods: the asymmetrical hot plane method (new experimental device) which is a
transient method, which allows the estimate of the thermal effusivity E from the experimental
temperatures. The experimental measuring of the mass specific heat using Differential
Scanning Calorimeter helps to deduce the thermal conductivity of these fibers. The second
method: the radial flux method is a steady state method that provides a direct measurement of
the thermal conductivity. Calculating the relative differences between the two measurement
methods allow us to validate the experimental results of the conductivity of the fibers and
compare to those of the literature review.

2. Fibers preparation and experimental testing

2.1 Fibers preparation
2.1.1 Kapok fibers
Kapok fiber is a natural cellulosic fiber which grows on the kapok plan. It consists of
unicellular fibers such as cotton but they are seven times less dense [18] than the latter and
have a buoyancy which may have twenty times its weight. Kapok (Fig 1.a) studied in this
paper was taken from the region in the Far North of Cameroon.

Fig 1: Natural fibers kapok: a) kapok fruit; b) fibers of kapok

Several authors were interested in the kapok fibers. Voumbo et all [19] measured the thermal
conductivity of kapok fibers using the box methods [20, 21]. It conductivity values obtained
varied between 0.03 W m-1 K-1and 0.04W m-1 K-1. Pang Cui et all [22] studied a heat transfer
for different temperatures through kapok fibers. The value of the thermal conductivity of
kapok fibers used for his experiments was 0.0486W m-1 K-1. Faming et al [23] after his study
proposed kapok fibers could be used jackets instead of duck fibers from ducks.

5
The process for obtaining kapok fibers for the measurement of thermal properties is as
follows:
- the fibers englobing the grain matures (Fig 1.a) at this stage, they are separated from
the kapok fruit cob and treated by removing solid grains and the result is a light-
weighted fiber as shown in (Fig 1.b)
- these fibers (of dry density 17.24 kg m-3) are then dried in an open air before being
placed for 48 hours in a vacuum chamber where the pressure can be reduced below 1
Pa
- once they are removed from the vacuum drying chamber, they are packed in a sealed
plastic bag in order to keep them at zero water content.

2.1.2 Groundnut shell fiber

Groundnut shells studied in this article come from the West Region of Cameroon. Indeed,
Cameroon Oil mill generates an average of 3,000 tons of groundnut shells [24]. During
burning of these shells, very large amounts of carbon dioxide are emitted into the atmosphere
contributing to global warming. It is therefore advisable to thermally recycle and valorize
these groundnut shells. In Senegal and Gambia [25], for example, brick made from groundnut
shells (bio char) are used as fuel and as an alternative solution to deforestation.

Fig 2: groundnut: a) groundnut; b) groundnut shell fiber

Before the thermophysical properties are measured, the process of obtaining groundnut shell
is as follows:
‐ the harvested groundnut pod from humid soil is dried in ambient air for several weeks.
this process help remove soil from groundnut.
‐ the groundnut as shown in Fig 2.a, are then separated from the shell by the use of a
husking machine
‐ the groundnut shell fibers are then wash with a tap water in order to remove all the
soil. The washing process stop until having a Ph of 7. Next, they are dried in an
electric drying oven at 70°C for 48 hours
‐ finally, the peelings are gathered and ground by a mechanic mold. The chopped
peelings are gathered (5-20 mm in length; maximum thickness of 0.3 mm and 147.92
kg m-3of dry density) as seen in Fig 2.b dried with a vacuum drying chamber for 48
hours.
‐ after these 48 hours, they are wrapped in a sealed plastic bag.

6
2.1.3 Coconut fibers
Coconut fibers studied in this article come from the Central region of Cameroon. In civil
engineering, coconut fibers are used as reinforcing material [26]. According to Li Z.et all [27]
who created the coconut fiber board (CFB) has shown that the fibers can replace building
materials such as tiles, bricks, polywood. It also showed that these fibers can provide a good
thermal comfort. Momohar et al [25] using a method in steady state (ASTM C518) showed
that the experimental conductivity of coconut fibers at 21.8 °C lies between 0.05009 W m-1 K-
1
and 0.05758 W m-1 K-1 for density varying between 40 and 90 kg m-3. Fig 3 shows the
coconut fibers.

Fig 3: coconut: a) coconut husk; b) coconut fibers Treated

Before measuring the thermophysical properties, the procedure for obtaining coconut fiber is
as follows:

‐ coconut is harvested once they are mature

‐ coconut fibers englobbing the nut, as shown in Fig 3.a are then separated from the
nuts, and then dried for several days in air. These dried fibers are then soaked in water
for several days in order to remove ash content and impurities.
‐ these soaked fibers are wash with a tap water until having a Ph of 7
‐ a rotation machine (285 tr/min) containing again a tap water, is used to mix all of the
fibers and facilitate their separation. Fibers obtained as shown in Fig 3.b
‐ once formed into small fibers (160 mm maximum in length;0.2-0.4 mm in diameter
and 46.76 kg m-3in dry density), they are finally dried in a vacuum drying chamber
for 48 hours
‐ after random selection, the fibers are packed in a sealed plastic bag.

2.1.4 Rattan fibers

Rattan fibers are physically similar to bamboo. Unlike bamboo, cane stems are more solid, but
most of the species need support and can’t grow up-right. In forest where Cane grows, its
economic value can help protect forest land by offering an alternative means for loggers to
switch from felling forest trees to harvesting rattan canes instead. Cane fibers are mixed with
mud to build houses in certain villages in south Cameroon as could be seen in Fig 4.b. The

7
cane studied in this write up is from Limbe (South West Region of Cameroon). Before
carrying out the experimental measurements, they were first of all cut into thin slices as shows
in Fig 4.a, (5-10 mm in length; 0.27 mm in diameter and 70.58 kg m-3 in dry density) and then
dried in a vacuum drying chamber for 48 hours.

Fig 4: Rattan: a) rattan fibers; b) mud house+fiber rattan

2.2 Experimental device

The experimental device set-up for measuring the thermal effusivity is show in Fig 5. It is an
asymmetrical hot plate device that allows the temperature recording at the center of the
heating surface.

Fig 5: View of the experimental device

2.2.1 Calibration process of the heating element

Temperature at the center of the heating element was measured with type K thermocouple
made with two wires of diameter 0.005 mm. These wires were stuck at the center of heating
element MINCO HK 5178. Fig 6 shows the scheme of the calibration process.

Fig 6: Scheme of calibration process of the heating element

The thermal conductivity of insulating material polystyrene measured by the center plate
method [26] is 0.032 W m-1 K-1 and its volumetric heat capacity is 4800 J m-3 K-1. With these
data, using a hot plate symmetrical method that minimizes the function

(Tc(t))mod is
n
F    (Tc )exp (ti )  (Tc ) mod el (ti )  where
2
the inverse Laplace transform of
i 1

 c ( z, p) given by:
0 Bi
c ( z, p)  (1)
2 p  h ch eh B p  D
i i
2
Where:
h, ch and eh are the density; the specific heat capacity and the thickness of the heating
element respectively.

8
Bi and Di are the coefficients of the quadrupolar matrix of insulating material
The estimated surface of the heating element is 0.010141 m2 and the estimated thermal
capacity of the heating element per unit area is 590.6 J m-2 K-1

2.2.2 Experimental program

A calibrated heating element made of a plane resistance inserted between two polymide films
was stuck to the flat face of an insulating material with a surface of 150 mm150 mm and of
perfectly known thermal effusivity. On the section of the insulating material not covered
(50mm50mm) with the heating element, another sample of polystyrene, of equal thickness
as the first was stuck to it, but this one has a rectangular hole with the same dimensions as the
heating element and a depth of 5 cm. The processes used to measure the thermal effusivity
were:
- the fibers were inserted with the hand into the rectangular space in small heaps and
pressed until they filled all the space,
- before temperature was recorded, the fibers with which the rectangular hole was filled
were removed and put in a sealed dry plastic bag, whose mass is known,
- these fibers were weighed and dried for 24 h with the help of a vacuum drying
chamber.
- after the 24 h, these fibers were weighed once again to obtain the dry mass
mf,Knowing the apparent volume Va of a rectangular hole (10 cm 10 cm 5 cm) and
the apparent mass ma, their apparent density ρapp can easily be determined.
mf
 app  (2)
Vapp

- lastly, The fibers were rapidly re-inserted intothe rectangular spacethen covered with a
plastic film to keep moisture close to zero. The rectangular hole was covered with an
insulating aluminum block (see Fig5) with a thickness of 40 mm and cross sectional
area of 130 mm130mm, in order to keep moisture of the air close to zero.
This experimental set-up is placed on an isothermal block ofthicknessof20 mm and a section
of 200 mm200mm. The temperature (T0) of the isothermal block was assumed uniform. This
hL
hypothesis is validated if the Biot number Bi  is lower than 0.1 [27]. Considering h=10
 Al
W m-2 K-1, the thermal conductivity of the blocks Al=200 W m-1 K- 1it implies that Bi=0.01 so
that the temperature of the aluminum blocks may be considered as uniform.

9
A flux step was then sent into the heating element by means of the electric generator and the
temperature Th(t) at the center of heating element was recorded.

3. Physical modelling
The modelling was based on the following assumptions:
- Initially, the system has a uniform temperature Ta (ambient temperature)
- Polystyrene is an insulating material. So the contact thermal resistance at the interface
of the heating element/polystyrene will be neglected;
- the fibers are dried, no mass transfer occurs

3.1 3D model
If T(x,y,z,t) is the temperature through the fibers, the heat transfer equation is:

Fig7: Schema of the modeled 3D system

 2T ( x, y, z , t )  2T ( x, y, z , t )  2T ( x, y, z, t ) 1 T ( x, y, z , t )
   (3)
x 2 y 2 z 2 a t
Initial condition is:
t  0, T ( x, y, z, 0)  Ta (4)

The boundary conditions (Fig 7) may be written as:

T (0, y, z , t )
x0 0 (5)
x
T (x, 0, z , t )
y0 0 (6)
y
z  e 0  x  c et 0  y  d
T (x, y, e, t ) (7)
  h1 (T (x, y, e, t )  Ta )
z
T (c  a, y, z, t )
x ca   h(T (c  a, y, z, t )  Ta ) (8)
z
T (x, y, 0, t ) Th (x, y, t )  T (x, y, 0, t )
z0   (9)
z Rc
T (c, y, z, t ) T (c, y, z, t )
xc   i i (10)
x x

10
 T (x, d, z, t ) T (x, d, z, t )
yd   i i (11)
y y
Tc ( x, y , t ) T ( x, y , 0, t ) T ( x, y, 0, t )
z0 0   h ch eh   i i (12)
t z z
Where:
 is the fibers’ thermal conductivity,
i is the polystyrene thermal conductivity,
ei is the thickness of the insulating material,
Ti is the temperature at the center of insulating material block,
Th is the temperature at the center of heating element,
2c and 2d are respectively width and length of the rectangular space in which the fiber were
filled,
h1 is the convective heat transfer coefficient on the top of the fibers,
h is the convective heat transfer coefficient on the lateral sides of the insulating material,
0 is the heat flux density produced in the heating element.
Equation (7) formalizes the continuity of the heat flux losses on the top of the fibers. Equation
(9) formalizes the continuity of the heat flux produced at the interface resistance Rc (thermal
contact resistance) between the heating element\fibers. Equations (10) and (11) formalize the
continuity of the heat flux on each side of interfaces fibers/polystyrene.
Using the method of separation of variables, it is not possible to solve equation (3) with its
boundary conditions through an analytical method, due to the fact that temperature T(x,y,z,t),
Th(x,y,t) and Ti(x,y,z,t) are not equal and moreover there is no continuity in the thermal
effusivity on these interfaces. The only way to solve it is through the numerical method.
Using the value of the thermal conductivity obtained by estimation from the simplified model,
a quarter of the apparatus was simulated with the help of the software Comsol. Fig 8 shows a
simulation from the thermo-physical properties of a groundnut shell fiber obtained from a
simplified model (=0.097 W m-1 K-1, cp=218.0089 103 J m-3 K-1).

Fig 8: Schema of the quarter of experimental device solve by Comsol for groundnut shell:
e=5cm, =0.097 W/m/ K, cp=218.0089 103 J m-3 K-1

By considering the simulated values of temperature (T (t ))comsol (step time 0.5 s) performed

using the software Comsol, firstly by considering the lateral heat losses on the sides of the

11
polystyrene block h=0 (1D transfer) and the convective heat coefficient on the top of fibers
h1=10 W m-2°K-1, the next step h=10 W m-2°K-1 (3D transfer) and h1=10 W m-2°K-1. The
T (t ) h 0  T (t ) h 10
relative difference was calculated. The time during which 1 % of this
T (t ) h 0
relative difference was observed was tmax=450 s. Thus, the time range for which a good
estimation of the thermal effusivity can be obtained by the 1D transfer assumption of heat is
[0, 450 s]. This result may be validated by observing the residues curves of 1D model.

3.2 1D model
A simplified model may be established by considering the supplementary hypothesis that, the
heat transfer remains 1D at the center of system during the experiment. With this hypothesis,
the temperature at the center of heating element depends only on z and t (see Fig 9).

Fig 9: Experimental device in sectional view

Using the quadrupole formalism presented by Maillet et al [19], and neglecting the convection
lateral heat loss on lateral faces of insulating material and considering the heating element as a
thin system, the following matrix relations can be written:
- when the heat flux density (  01 ) leaving the heating element through the fibers is

considered:

 c   1 0   1 SRc   A B   1 1 0 

 h1  (13)
      
  01    h ch eh p 1   0 1   C D   0 1   1 

- when the heat flux density leaving (  02 ) the heating element through the insulating

materials is considered:
 c   Ai Bi  0 
    (14)
  02   Ci Di    2 

where:
 sh(qe) 
 A B   ch(qe)
q 
p
  with q
 C D    qsh(qe) 
ch(qe) 
a

and:

12
  sh(qi ei ) 
 Ai Bi   ch(qi ei )
i qi 
p
  with qi 
 Ci Di   ai
 i qi sh(qi ei ) ch(qi ei ) 

with:
-  c ( z, p)  L(T ( z, t )) is Laplace transform of the temperature difference T(z,t)-Ta
-  0 is the Laplace transform of total heat flux density produced in the heating element

- Rc is the thermal contact resistance at the interface heating element/fibers (°C W-1)
- p is the Laplace parameter
- a and ai are respectively thermal diffusivity of fibers and thermal diffusivity of
polystyrene (m s-2).
The total heat flux density in the Laplace space is:
 0  1   2 (15)

After combining relations (13), (14) and (15), the final result is given by relation (16):

0 1
c ( z, p)  (16)
p  h ch eh ( A  Bh1 ) p  Dh1  Di
A  Bh1 Bi
Applying the inverse Laplace transform by use of De Hoog algorithm [29] at the relation (16),
numerical value of temperature model Tmod el (t ) through the fibers was calculated.

3.3 Simplified model

A simplified 1D model may also be written with the hypothesis that the heat transfer remains
1D at the center of a system and added to that, both fibers and insulating materials are semi-
infinite medium. Within these hypotheses, one can write:
- when the heat flux density (  01 ) leaving the heating element through the fibers is

considered:

c   1 0   1 
     (17)
  01    h ch eh p 1   E p1 
- when the heat flux density leaving (  02 ) the heating element through the insulating

material is considered:

  c   Ai Bi    2 
     (18)
  02   Ci Di   Ei p 2 

where:

13
E is the thermal effusivity of fibers (J m-2 C-1s-1/2)
Ei is the thermal effusivity of the insulating material (J m-2 C-1s-1/2).
The equations (17) and (18) lead to:

0 1
c ( p)  (19)
p  h ch eh p  ( E  Ei ) p

For sufficiently long times (p0):

20
T (0, t  )  t (21)
( E  Ei ) 
Thermal effusivity may be evaluated with simplified 1D model from numerical calculation of
the slope α(t) of the linear part of curve T(t)=f( t ).
20
E  Ei (22)
 

3.4 Experimental measurement method of specific heat capacity

Before each experiment, the fibers were dried in a vacuum drying chamber (PFIEFER
VACUUM PMZ 01300) for 48 hours. The specific heat capacity cp of each fiber and of the
inner tube was measured with a Differential Scanning Calorimeter SETERAM DSC3. The
fibers’ densities were derived from weighing a known volume of dried fibers (relation 2). The
results are presented in Table 1.

Table 1: Measured values of dry densities and dry mass specific heat capacity

For four products, mass heat capacity as a function of temperature varies linearly as shows in
Fig10.

Fig 10: Mass heat capacity of fiber as a function of temperature

4. The thermal effusivity estimation method

The thermal effusivity was measured using an asymmetrical hot plate method (see Fig 5).
Estimation of the parameter E is achieved by a completed model taking as initial value, the
value pre-estimated from the simplified model with experimental temperatures. The values of

14
thermal contact resistance Rc and convective heat coefficient h1are fixed respectively to 10-10
°
K W-1 and 10 W m-2 K-1. The theoretical curve Tmod el (t ) is calculated by relation (16) using

the De Hoog algorithm. The Levenberg-Marquart algorithm [30] integrated in the Matlab
code to estimate value of the thermal effusivity which minimizes the sum of quadratic errors
n
    Texp (ti )  Tmod el (ti ) 
2
between the experimental curve Texp (t )  T (0, t )  Ta
i 1

(temperature recording) and theoretical curve. The estimation has been done on a time
interval [t0, tmax] such as the residues are small and perfectly centered around 0°C, which
validates our 1D model. With the value of the thermal effusivity and volumetric heat capacity
ρcp, the thermal conductivity λ is then deduced from relation (22).
E2
 (22)
cp

5. Experimental study and discussion

The experimental studies focused on four natural fibers samples:
- a very light fiber: kapok
- light fibers: coconut fibers
- heavy fibers: groundnut shell fibers
- heavy fibers: rattan fibers.
5.1 Estimation from experimental 1D model
Three measurements were made for each sample of fiber and the mean values retained. Fig
11.a shows the experimental curve and the slope curve (simplified model) and Fig 10.b the
experimental curve and simulated curve with the pre-estimated parameter values of the
thermal properties of groundnut shells (completed model).

Fig 11: Temperature curve for groundnut shells. e=5cm, ei=5cm, i=0.032 W m-1K-1,
icpi=48000 J m-3 K-1. a) Simplified model; b) Completed model

It may be noted that unlike high density material, such as lateritic soil powder for example
[31], the temperature is sensitive to the volumetric heat capacity of the probe for a long
duration and for the low density materials. As concerns groundnut shell fibers, for example
(Fig 11), during the first 25 s after having recorded the temperatures, it became only sensitive

15
to the thermal capacity of the sensor and the contact resistance sensor interface/fibers. This
long duration could be due to an estimation problem from the 1D model if the 3D effects
appear very quickly. If we observe Fig 11.a, we notice that heat transfer in the center of the
fibers is unidirectional (residues curves are centered at 0 °C) within the time interval of [5 s;
20 s] for the simplified model and the time interval of [50 s; 400 s] for the completed model.
An account of maximum time of (400 s) through which the estimation can be done this time is
slightly below the maximum time estimated by Comsol’s simulations (tmax= 450 s). This
could be confirmed if only the quadripolar developed model does not take into account the
side convective losses on the sides of insulating material (h= 0), but only on the convection
(h1=10 W m-2 K-1) which may occur on the surface of the fiber in contact with ambient air
during the temperature recording time. For example, with groundnut shell fibers, an estimate
ranging from 100-150 s for the simplified model is equal to E = 145.41 J m-2 °C-1 s-1/2 and =
0.097 W m-1 K-1 while between 200-300 s from the completed model is equal to E = 143.910
J m-2 °C-1 s-1/2 and = 0.095 W m-1 K-1. The standard deviation between the pre-estimated
values and the estimated values are less than 4 % which is quite acceptable. This observation
remains valid for coconut fibers and rattan fibers (see Table 2). The difficulty is even greater
with the case of kapok fibers which is very light (apparent density of 17.24 kg m-3). Fig 12
shows the evolution of the temperature at the center of kapok fibers placed in the open air for
3 days. If we observe the experimental temperature and the residues curve of simplified
model, we will notice that the thickness of the layer of the kapok is felt for a very long time
unlike that of groundnut fibers which is short. For example, the time range within which the
assumption of semi-infinite medium is valid is 100-144 s for kapok (residues centered within
this time range: Fig 12.a) against 25-300 s for groundnut shells (residues centered within this
time range: Fig 11.a). Fig. 12.b shows that the completed model does not minimize the
differences between the experimental temperatures and model temperature. This can be
justified by the fact that the recording of temperatures is done on kapok fiber exposed to
moisture of the air for 3 days (apparent density 18.24 kg m-3, water content Xe= 5.66%)
meanwhile the quadripolar model does not take into account mass transfer phenomena.

Fig 12: Temperature evolution of kapok fibers. e=5 cm; ei=5 cm; i=0.184 W m-1 K-1,
icpi=48000 J m-3 K-1. a) Simplified model; b) Completed model

16
We then pulled these kapok fibers under vacuum drying chamber for 48 hours (apparent
density 17.1 kg m-3; water content Xc = 0 %.) As shown in Fig. 13 this process improves the
residues which are then centered at zero degree for up to 300 s (Fig. 12.b). The estimation
could then be better with the simplified model (100-225 s) and much better with the
completed model (10-300 s). Above 300 s, (time in which the heat flux is no longer
unidirectional at the center of the heating element: the residue curve is no longer centered at 0
°
C) only a 3D model can allow a good estimation of the thermal effusivity.

Fig 13: Temperature evolution for dry fibers of kapok. e=5 cm; ei=5 cm; i=0.184 W m-1 K-1,
icpi=48000 J m-3 K-1. a) Simplified model; b) Completed model

Table 2 presents a review of experimental results of the thermal properties of fibers estimated
with the simplified and the completed models.

Table 2: Estimated values

One can note a good agreement between the simplified and the completed model: the standard
deviation is less than 5 % that is quite acceptable.

5.2 Estimation from the radial flux model

In order to compare these values of the thermal conductivity to other values obtained by
another experimental method, the thermal conductivities of the fibers were measured by the
radial flux method which is a steady state method [32]. It consists of two coaxial tubes: an
inner tube of stainless steel of inner radius 4 mm; outer radius 4.5 mm with a height of 380
mm and an outer brass tube of inner radius 39 mm; outer radius 40 mm and a height of 426
mm. The centering of the inner tube is provided by two hollowed out PVC foams with
thickness 8 mm, internal diameter 9 mm and outside diameter 78 mm. These two discs were
placed above and below the tubes respectively; they help to maintain the fibers in the annular
space between the tubes and to isolate it from the moisture of air. The outer cylinder was
cooled by circulating water through a pipe wrapped around its outer surface in order to
maintain a constant uniform temperature. Two type K thermocouples of 0.005 mm in
diameter are placed at mid-height respectively on the inner tube of stainless steel to measure
T1(t) and on the outer tube of brass to measure T2(t).

17
The experimental procedure was as follows: the fibers were filled into the annular space
between the two tubes (see Fig 14). It involves passing a current of high intensity through the
inner tube of stainless steel and the temperature was recorded by an ALMENO 2390-5
apparatus until the temperature T1(t) and T2(t) reach the stationary values. An advantage of
this method is that with this method it does not suffer the problem of side convective losses:
no 3D effect can occurs.

Fig 14: Experimental device for radial flux method: a) general view, b) schematic view

The modelling was based on the following assumptions:

- the fibers are dry: no mass transfer occurs
- the temperature gradients along z are negligible
- the radial temperature gradient is negligible in the thickness of the inner tube
The heat transfer equation at the steady state regime is:
1   T 
r 0 (23)
r r  r 
The heat flux at a distance can be written:
 T  dT
  A(r )      2 rL (24)
 r  dr
The conservation of heat flux between the outer radius of the stainless steel inner tube of R1
(m) and the inner radius of the outer brass tubeR2 (m) allows:

2 rL
R2 2T
dr

R1
r
 
T1

dT (25)

This last relation leads to:

 R2 
 ln 

  R1 
(26)
2 L(T1  T2 )
By assuming that all the power supply by the electric generator is totally transformed into heat
by Joule effect in the resistance heater. The heat flux is:
  RI 2 (27)
where: R is the electrical resistance of the stainless steel tube calculated from the relation
(27):

18
 L
R (28)
 ( R12  Ri 2 )
where: Ri (m) is an inner radius of the inner tube and φ(μ cm) electrical resistivity [33]
given by:
  71.3  1.261 103 T1  7.76 104 T12 (29)

By combining relations (27) and (28), we finally have:

 R2  2
 ln I
 2 2  R1 
(30)
2 ( R1  Ri 2 )(T1  T2 )

The relation (30) helps to determine the thermal conductivity of the fibers once the steady
state regime is reached. Another advantage of this method is that the fibers can be over-loaded
to the maximum around the annular space between the two tubes thereby increasing the
number of contact points between the fibers and the hot tube. The results of the thermal
conductivity of the fibers obtained by this method are shown in table 3. We also calculated the
standard deviations between experimental values of thermal conductivity obtained by the
method of radial flux and those derived from the estimate of the thermal effusivity of the
completed model and the simplified model.

Table 3: Measured values and estimated values of thermal conductivity of fibers

For each of the sample fiber used, an additional mass of dried fiber was added to the annular
space by strongly pressing them to those contained already in the annular space. The purpose
of this is to find out if any addition of fibers changes the value of the thermal conductivity.
We noticed that once the mass of the fibers introduced into the annular space between the two
pipes ensures a perfect contact with the hot tube, the conductivity value remain constant. This
is consistent with the conclusions made by Agoudjil et al [35] findings. The results analysis in
Table 3 shows good agreement between the experimental results of thermal conductivity
obtained by the method of radial flux and that gotten from estimating asymmetrical hot plate
method. The deviations observed are less than to 3 % (except for kapok fiber), which is
acceptable. As concerns kapok and its literature review [19,20,21,22], we can say that due to
their very light density, only a steady state method like radial flux method presented in this
section or the three-layer cylinder method [36] can allow a good estimate of its thermal
conductivity. Added to that, its thermal conductivity is close to marketed fibers [37]: polyester

19
fibers (= 0.045 W m-1 K-1); hemp wool or glass wool (= 0.05 W m-1 K-1) and therefore may
serve as an alternative to these synthetic insulators. There is a correlation between the results
from coconut fiber and that of the literature review [37]: coconut wool (= 0.05 W m-1 K-1).
The experiments on the thermal characterization of groundnut shell and rattan fibers being
rare in the literature review, due to the results obtained for the two other fibers which are in
agreement with the literature review, it can be concluded that the thermal conductivity of
groundnut shell fibers and rattan fibers are around 0.09 W m-1 K-1 and 0.07 W m-1 K-1
respectively.

6. Conclusion
The asymmetrical hot plate method used in this article is a simple device which helps to
estimate the thermal effusivity. The experimental results obtained by this method compared to
the radial flux method, shows that the quadrupolar 1D model used can give a good estimate
for plant fibers with high density. Direct measurement of the thermal conductivity achieved
with radial flux method, has helped to validate that got from the estimation of E and
volumetric heat capacity obtained by Differential Scanning Calorimeter. The experimental
results for kapok and coconut fibers that correlate with those of literature review permitted us
to validate 0.09 W m-1 K-1 and 0.07 W m-1 K-1 as the thermal conductivities of rattan and
groundnut shells fibers respectively. In tropical countries such as Cameroon, as far as kapok
fibers are concerned, their use as a substitute to conventional insulation can be less expensive
than imported insulators. This can lead to significant economic savings on the investment cost
for constructing cold rooms for storing of basic food items for example. Introducing new
building materials aggregated with fibers of low thermal conductivity is important for
reduction of energy consumption and consequently contributes in the reduction of emission of
greenhouse gases, therefore to climate changes.

Acknowledgements
A special thank is send to research teams of LEMTA INPL-Nancy, Lorraine University-
France for their support to the expertise of characterization methods used in this work.

20
References
[1] L. Arnaud, V. Cerezo, D. Samri, Global approach for design of building material
containing lime and vegetable particles, In. The sixth International symposium on cement and
concrete, 1261-1265
[2] F. Collet, M. Bart, S. Moisette, L. Serres, Conductivité thermique des matériaux à faibles
impacts environnementaux-Modélisation par méthode autocoherente, VII colloque
interuniversitaire Franco-Québécois (CIFQ), Thermique des systèmes, Saint-Malo, France, pp
449-454
[3] V. Cerezo, Propriété mécanique thermique et acoustique d’un matériau à base de
particules végétales. Approche expérimentale et modélisation thermique. Thèse de doctorat,
École Nationale des Travaux Publics de l’Etat, 2005
[4] F. Collet, Caractérisation hydrique et thermique de matériaux en génie civil à faibles
impacts environnementaux. Thèse de Doctorat de l’Institut National des Sciences Appliquées
de Rennes, 2004
[5] D. Samri, Analyse physique et caractérisation hygrothermiques des matériaux de
construction: Approche expérimentales et modélisation numérique. Thèse de Doctorat, École
Nationale des Travaux Publics de l’Etat, 2008
[6] A. Evrard, Transient hydrothermal behavior of lime-hemp materials, Thèse de Doctorat,
École Polytechnique de Lovain, Belgique, 2008
[7] H. Bal, Y. Jannot, S. Gaye, F. Deumerie, Measurement and modelisation of thermal
conductivity of a wet composite porous medium: laterite based bricks with millet waste
additive, Construction and Buildings Materials 41 (2011), 586-593
[8] P. Meukam, Valorisation des briques de terre stabilisée en vue de l’isolation thermique de
bâtiments. Thèse Doctorat Spécialité Energétique Université Cergy Pontoise, Novembre 2004
[9] V. Asangwing, Optimization of formulations of stabilized earth blocks using a local
natural fibers for home construction, These de Doctorat, ENSET, Douala, 2010
[10]A. Mekhermeche, Contribution à l’étude des propriétés mécaniques et thermiques des
briques en terre en vue de leur utilisation dans la restauration des krous sahariens. Mémoire de
magister. Université de KasdiMerbah, Ouargla, 2012.
[11] P. Anderson, Thermal conductivity of some rubers under pressure by the transient hot
wire method, J. Appl. Phys (1976) 47 2424-6.
[12] R. Coquart, B. Baillis and D. Quenard, Experimental and theorical study of the hot wire
method applied to low-density thermal insulators, Int. J. Heat Mass Transfer (2006) 49 4511-
24.

21
[13] D. Salmon, Thermal conductivity of insulations using guarded hot plates including recent
developments and source reference materials, Meas. Sci. Technol (2001) 12 89-98.
[14] J. Huan, Sweating guarded hot plate test method, Polym. Test (2006), 25 709-16.
[15] S.E. Gutafsson, Transient plane source techniques for thermal conductivity and thermal
diffusivity measurements of solid materials, Int. Commun Heat Mass Transfer (1991), 17 473-
84.
[16] Y. He, Rapid thermal conductivity measurement with a hot disk sensor: Part I.
Theoritical considerations, ThermochimicaActa (2005), 436 122-9.
[17] Y. Jannot, A. Degiovanni and G. Payet, Thermal conductivity measurement of
insulationg materials with a three layers device, Int. Jour. of Heat Mass transfer 52 (2011)
1105-1111.
[18] N. Dujardin, Un matériau biosource de choix: caractérisations et applications, 25ème
journée scientifique de l’environnement, 18-20 février 2014
[19] M. L. Voumbo, A. Wereme, S. Gaye, M. Adj and M. Sissoko, Charaterization of
thermophysical properties of kapok, Research Journal of Applied Sciences, Engineering and
Technologies, (2010), 2(2): 143-148.
[20] P.S Ngohe-Ekam, P. Meukam, G.Menguy, P. Girard, Thermophysical characterization of
tropical wood used as building materials: with respect to the basal density, Construction and
building materials 20 (2006) 929-938
[21] A. Noumowe, P. Meukam, Y. Jannot and R. Duval, thermophysical and mechanical
characterization of stabilized clay bricks for building thermal insulation, Materials and
structures Vol. 36, August-september 2003, pp 453-460
[22] C. Pend and W. Fumei, An investigation of heat flow through kapok insulating material,
Fibres and textiles (2009), Vol 2 N0 1, pp: 88-92.
[23] W. Fumei, Comparaisons of thermal and evaporative resistances of kapok coats and
traditional down coats, Fibres and textiles (2010), Vol 18 N0 1(78), pp: 75-78
[24] I. Fonfatawouo, A. BoubaOumarou, Politique nationale, stratégie et pla d’action pour
l’efficacité énergétique dans le secteur de l’énergie au Cameroun, ARSEl, Mars 2014
[25] L. Thioune, Le biocharbon dans la politique énergétique du Sénégal, Journal Vie N0 11,
Mai-Juin 2009.
[26] A. Majid, Coconut fibre-A versatile material and its applications engineering, Second
International Conference on Sustainable Construction Materials and Technologies, June 2010
ISBN 978-1-4507-1490-7.

22
[27] Z. Li, L. Wang and X. Wang, Flexural characteristics of coir fibre reinforced
cementitious composites, Fibres polymers (2006) 7(3), 286-294.
[28] K. Manohar, D. Ramlakhan, G. Kochhar and S. Haaldar, Biodegradable fibrous thermal
insulation, J. Braz. Soc. MechSciEng (2006) ISSN:1678-5878 28(1):45-47.
[29] Y. Jannot, B. Remy and A. Degiovanni, A centered hot plate method for measurement of
thermal properties oh thin insulating materials, Measurements Sciences and Technology, 21,
20100
[30] S. Bories, M.Prantt, Transferts de chaleur dans les milieux poreux, Tech. Ing (1995), B8
(250).
[31] D. Maillet, S. Andre, J.C. Batsale, A. Degiovanni and C. Moyne, Thermal Quadrupoles,
Wiley, New York, 2000.
[32] F. R. De Hoog, A improved method for numerical inversion of Laplace transforms,
(1982). Soc. Ind. App. Math 3 357-66.
[33] P.E Gill and W. Murray, Algorithms for the solution of the non linear least-squares
problems, J. Numr. Anal, Vol 15, N0 5 (1978), pp:977-992.
[34] C.Y. Ho and T.K.Chu, Electrical resistivity and thermal conductivity of nine selected
AISI stainless steels, Cindas report 45, Iron and steel Institute, 1000 sixteenth street N.W.
[35] B. Agoudjil, A.Benchabane, A.Boudenne, L. Ibos, M. Fois, Renewable materials to
reduce heat loss: characterization of plate palm wood, Energy and Buildings, 43 (2011), 491-
497
[36] Y. Jannot, A. Degiovanni, Thermal properties measurement of dry bulk materials with a
cylindrical three layers device review of Scientific Instruments (2013) 84, 094901.
[37] Agence locale de l’énergie. www.ale-lyon.org

23
List of Figs:
Fig 1: Natural fibers kapok: a) kapok fruit; b) fibers of kapok
Fig 2: Groundnut: a) groundnut nut; b) groundnut shell fibers
Fig 3: coconut: a) coconut husk; b) coconut fibers treated
Fig 4: Rattan: a) rattan fibers; b) mud house+fiberrattan
Fig 5: View of the experimental device
Fig 6: Scheme of calibration process of the heating element
Fig 7: Schema of the quarter of experimental device solve by Comsol for groundnut shell:
e=5cm, =0.097 W/m/ K, cp=218.0089 103 J m-3 K-1
Fig 8: Experimental device in sectional view
Fig 9: Mass heat capacity as a function of temperature
Fig 10: Temperature curve for groundnut shells. e=5cm, ei=5cm, i=0.032 W m-1K-1,
icpi=48000 J m-3 K-1. a) Simplified model; b) Completed model
Fig 11: Temperature evolution of kapok fibers. e=5 cm; ei=5 cm; i=0.184 W m-1 K-1,
icpi=48000 J m-3 K-1. a) Simplified model; b) Completed model
Fig 12: Temperature evolution of kapok fibers. e=5 cm; ei=5 cm; i=0.184 W m-1 K-1,
icpi=48000 J m-3 K-1. a) Simplified model; b) Completed model
Fig 13: Temperature evolution for dry fibers of kapok. e=5 cm; ei=5 cm; i=0.184 W m-1 K-1,
icpi=48000 J m-3 K-1. a) Simplified model; b) Completed model
Fig 14: Experimental device for radial flux method: a) general view, b) schematic view

24
Fig 1: Natural fibers kapok: a) kapok fruit; b) fibers of kapok

Fig 2: groundnut: a) groundnut; b) groundnut shell fiber

Fig 3: coconut: a) coconut husk; b) coconut fibers Treated

Fig 4: Rattan: a) rattan fibers; b) mud house+fiber rattan

25
 Fig 5: View of the experimental device

Fig 6: Scheme of calibration process of the heating element

26
 Fig7: Schema of the modeled 3D system

Fig 8: Schema of the quarter of experimental device solve by Comsol for groundnut shell:
e=5cm, =0.097 W/m/ K, cp=218.0089 103 J m-3 K-1

27
 Fig 9: Experimental device in sectional view

Fig 10: Mass heat capacity of fiber as a function of temperature

28
 15
Texp
Slope
Residus x10
10

Temperature (C)
5

-5
0 5 10 15 20 25 30 35
rac(t)

a)
15
Texp
Tmodel
10 Residus x10
Temperature(C)

-5

-10
0 200 400 600 800 1000 1200
t(s)

b)
Fig 11: Temperature curve for groundnut shells. e=5cm, ei=5cm, i=0.032 W m-1K-1,
icpi=48000 J m-3 K-1. a) Simplified model; b) Completed model

29
 15
Texp
Tmodel
10 Residus x10

Temperature(C)
5

-5

-10
0 100 200 300 400 500 600
t(s)

a)
15
Texp
10 Slope
Residus x10

5
Temperature (C)

-5

-10

-15
0 5 10 15 20 25
rac(t)

b)
Fig 12: Temperature evolution of kapok fibers. e=5 cm; ei=5 cm; i=0.184 W m-1 K-1,
icpi=48000 J m-3 K-1. a) Simplified model; b) Completed model

30
 15
Texp
Slope
10 Residus x10

Temperature (C)
5

-5

-10
0 5 10 15 20 25 30
rac(t)

a)
15
Texp
Tmodel
10 Residus x10
Temperature(C)

-5

-10
0 100 200 300 400 500 600 700 800 900
t(s)

b)
Fig 13: Temperature evolution for dry fibers of kapok. e=5 cm; ei=5 cm; i=0.184 W m-1 K-1,
icpi=48000 J m-3 K-1. a) Simplified model; b) Completed model

a) b)
Fig 14: Experimental device for radial flux method: a) general view, b) schematic view

31
List of tables:
Table 1: Measured values of dry densities and dry mass specific heat capacity

T(0C) 20 30 40 50 60
GS  (kg m-3) 170.92
cp (J kg-1 K-1) 1259 1310 1363 1415 1470
cp (J kg-1 m-3) 215.18 103 223.90 103 232.96 103 241.85 103 251.25 103
Coconut  (kg m-3) 46.76
cp (J kg-1 K-1) 1180 1228 1278 1329 1381
cp (J kg-1 m-3) 55.17 10 3 57.42 103 59.75 103 62.14 103 64.57 103
Kapok  (kg m-3) 17.1
cp (J kg-1 K-1) 1261 1307 1360 1413 1469
cp (J kg-1 m-3) 21.56 103 22.34 103 23.25 103 24.16 103 25.11 103
Rattan  (kg m-3) 70.58
cp (J kg-1 K-1) 1280.5 1341.5 1414.5 1466.1 1526.9
cp (J kg-1 m-3) 90.65 10 3 94.68 103 99.90 103 103.47 103 107.76 103

GS: Groundnut Shell

32
 Table 2: Estimated values

Fibers Simplified model Completed model

E cp  E cp  D1 D2
-2 °C-1 s-1/2 -2 °C -1 s-1/2
Jm J Kg-1 m-3 W m-1 K-1 Jm J Kg-1 m-3 W m-1 K-1 % %
GS 145.419 218008.90 0.097 143.910 218008.90 0.095 1.037 2.061
kapok 31.453 21049.38 0.0477 30.879 21049.38 0.0463 1.824 2.935
coconut 56.952 55923.96 0.0582 56.459 55923.96 0.0576 0.865 1.030
Rattan 82.202 93851.69 0.0710 83.336 93851.69 0.0741 1.360 4.183
GS: Groundnut shell, D1 standard deviations between the completed and the simplified model for thermal effusivity, D2
standard deviations between the completed and the simplified model for thermal conductivity

33
 Table 3: Measured values and estimated values of thermal conductivity of fibers

Fibers Radial lux SM CM

T1 T2 T I     D1 D2
°C °C °C A Μ  cm W m-1 K-1 W m-1 K-1 W m-1 K-1 % %

GS 26.6 20.5 6.5 5.45 71.88 0.093 0.097 0.095 4.123 2.105
kapok 33.4 20.4 11.7 5.45 72.20 0.042 0.0477 0.0463 10.638 7.284
coconut 30.1 205 9.10 5.36 72.04 0.058 0.0582 0.0576 0.429 1.724
Rattan 31 23.3 7.7 5.46 72.08 0.072 0.0710 0.0741 1.291 2.702
GS: Groundshell, SM: Simplified model; CM: Completed model; D1: Sstandard deviation between radial flux and Simplified
model; D2: standard deviation between radial flux and Completed model

34