The detailed process of bio composite implementation is described in Figure 1 . The material was produced by the contact molding process. This type of process keeps the pigs on the surface of the fiber. The average Young’s modulus of the raffia vinifera fiber used is 4.4 GPa [15] . The Young’s modulus of the epoxy resin used is 3.1 GPa in tension. The mass fraction of the woven fiber structure is 25% of the total mass of the composite. The epoxy resin was mixed with a hardener at 40% volume fraction. In accordance with ISO 8302:1991, the bio-composite sample plates were cut to dimensions ( $10\ast 10\ast 1$ cm³) and placed in molds. The resin was then cast. The process diagram is shown in Figure 1 .

The main equipment used for physical characterization is a digital milligram balance (0.001 g), a milliliter beaker (0.001 l), a millimeter caliper (0.001 m), distilled water and containers.

There are two approaches to determining the density of the biocomposite; that by direct weighing, dimension measurement then calculation of volume (apparent mass) [16] according to the ASTMD 792 standard. The expression of the apparent mass is given by Equation (1).

${\rho }_{ap}=\frac{m}{V_{th}}$ (1)

m: mass of the composite (kg); $V_{th}$: theoretical volume of the composite (m³).

As for the actual density ( ${\rho }_{real}$), it is found by direct weighing of mass, then embedding in paraffin then immersion in distilled water contained in a graduated volume (beaker), the actual density is given by Equation (2).

${\rho }_{real}=\frac{m_c}{V_{en}-\frac{m_{en}-m_c}{{\rho }_{pa}}}$ (2)

with $m_c$: mass of the composite (kg), $V_{en}$: volume of the coated composite (m3 ), $m_{en}$: mass of the coated composite (kg), ${\rho }_{pa}$: density of the paraffin (kg/m 3 ). Using the actual and volumetric density, the porosity of the composite is deduced by Equation (3).

$P=\frac{{\rho }_{real}-{\rho }_{ap}}{{\rho }_{real}}\ast 100$ (3)

where: $P$ is the porosity of the woven composite material.

It is noted ( $\tau$) and is done according to the ISO 62:2008 standard. The test consists of weighing the samples (ten), then immersing them in distilled water and carrying out successive weighings until saturation. The balance used is calibrated to the milligram. The expression for the absorption rate from [14] is given by Equation (4).

$\tau =\frac{m_f-m_i}{m_i}\ast 100$ (4)

where $M_f$: saturation mass (kg) and $M_i$: initial mass (kg).

The thermal properties of the composite material are determined by the asymmetric hot plane method with an insulated back face. The principle is to send a constant flow of heat into a heating element using a stabilized power supply whose power input is 220 V and the output varies from 0 to 30 V. The heating element with thermocouple K is in contact with the face of the sample and the temperature on this face of the sample is recorded by an 8-channel Picolog to which a temperature sensor is connected [17] . In this particular case, the temperature range was between the ambient temperature, which is considered as the initial temperature (32˚C), and the maximum temperature obtained in the case of this test, which is of the order of 50˚C. The Picolog recorder is connected to the computer which made it possible to interface the recording. This method is used by several authors such as [18] . The measurement uncertainties of the equipment are of the order of a thousandth. The unidirectional heat transfer model is given by Equation (5). Different from the rotating heat flow [19] .

$\left(\begin{array}{c}{\text{θ}}_{\text{c}}\\ {\text{Φ}}_{01}\end{array}\right)=\left(\begin{array}{cc}1& 0\\ {\text{C}}_{\text{hP}}& 1\end{array}\right)\left(\begin{array}{cc}1& {\text{SR}}_{\text{chs}}\\ 0& 1\end{array}\right)\left(\begin{array}{cc}{\text{A}}_{\text{s}}& {\text{B}}_{\text{s}}\\ {\text{C}}_{\text{s}}& {\text{D}}_{\text{s}}\end{array}\right)\left(\begin{array}{cc}1& {\text{S}}_{\text{s}}{\text{R}}_{\text{csi}}\\ 0& 1\end{array}\right)\left(\begin{array}{cc}{\text{A}}_{\text{i}}& {\text{B}}_{\text{i}}\\ {\text{C}}_{\text{i}}& {\text{D}}_{\text{i}}\end{array}\right)\left(\begin{array}{c}0\\ {\text{Φ}}_1\end{array}\right)$

$\left(\begin{array}{cc}{A}_s& B_s\\ C_s& D_s\end{array}\right)\left(\begin{array}{c}0\\ {\Phi }_1\end{array}\right)$ (5)

The following working conditions are permitted:

The hot flow propagates in one direction only;

The work materials are at ambient temperature;

There is no heat loss during the experiment;

The contact between the heating element and the insulation is resistance-free;

Polystyrene blocks are considered to be semi-infinite media.

The heat received by the sample from the heat source is given by Equation (6) [20] .

$\left(\begin{array}{c}{\text{θ}}_{\text{c}}\\ {\text{Φ}}_{01}\end{array}\right)=\left(\begin{array}{cc}1& 0\\ {\text{ρ}}_{\text{h}}{\text{c}}_{\text{h}}{\text{e}}_{\text{h}}\text{p}& 1\end{array}\right)\left(\begin{array}{cc}1& {\text{SR}}_{\text{chs}}\\ 0& 1\end{array}\right)\left(\begin{array}{c}{\text{θ}}_1\\ \text{E}\sqrt{\text{p}}{\text{θ}}_1\end{array}\right)$ (6)

Equation (7) represents the heat flux received by the polystyrene block $\left({\text{Φ}}_{02}\right)$from the element.

$\left(\begin{array}{c}{\text{θ}}_{\text{c}}\\ {\text{Φ}}_{02}\end{array}\right)=\left(\begin{array}{cc}1& {\text{SR}}_{\text{chi}}\\ 0& 1\end{array}\right)\left(\begin{array}{c}{\text{θ}}_2\\ {\text{E}}_{\text{i}}\sqrt{\text{p}}{\text{θ}}_2\end{array}\right)$ (7)

With $\text{E}$ and ${\text{E}}_{\text{i}}$respectively the thermal efficiencies of the sample and the insulator. The unit of thermal effusivity is $\left(\text{J}.{\text{m}}^{-2}.{\text{K}}^{-1}{\text{s}}^{-\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$2$}\right.}\right)$. Combining Equations (6) and (7) gives Equation (8) [21] .

${\text{θ}}_{\text{c}}\left(\text{z},\text{p}\right)=\frac{{\varphi }_0\text{S}}{\text{p}}\frac{1}{\frac{{\text{C}}_{\text{h}}\text{Sp}+\left(1+{\text{C}}_{\text{h}}{\text{pR}}_{\text{chs}}\text{S}\right)\text{ES}\sqrt{\text{p}}}{1+{\text{R}}_{\text{chs}}\text{SE}\sqrt{\text{p}}}+\frac{{\text{E}}_{\text{i}}\text{S}\sqrt{\text{p}}}{1+{\text{R}}_{\text{chi}}{\text{SE}}_1\sqrt{\text{p}}}}$ (8)

Equation (9) represents the inverse transform of this function for sufficiently long times (t → + ∞) i.e. in the Laplace domain+ when (p → 0).

$\Delta \text{T}\left(0,\text{t}\to +\infty \right)={\varphi }_0\text{S}\left(\frac{{\text{E}}^2{\text{R}}_{\text{chs}}+{\text{E}}_{\text{i}}^2{\text{R}}_{\text{chi}}}{{\left(\text{E}+{\text{E}}_{\text{i}}\right)}^2}-\frac{{\text{C}}_{\text{h}}}{\text{S}{\left(\text{E}+{\text{E}}_{\text{i}}\right)}^2}\right)+\frac{2{\varphi }_0}{\left(\text{E}+{\text{E}}_{\text{i}}\right)\sqrt{\text{π}}}\sqrt{\text{t}}$ (9)

The square root of time characterizes the heat variation within the source that generates it. The thermal effusivity of the sample can be pre-estimated from the one-dimensional model by calculating the slope α of the experimental curve between two times t and $\text{t}+\text{dt}$. It is presented in Equation (10).

$\text{ΔT}\left(0,\text{t}\right)=\text{f}\left(\sqrt{\text{t}}\right)$ (10)

In this way, over a time interval the curve is assimilated to a straight line. From then on, the calculation is carried out on the linear third of the curve. Equation (11) presents the slope of the straight line.

$\text{α}=\frac{\text{T}\left(\text{t}+\text{dt}\right)-\text{T}\left(\text{t}\right)}{\sqrt{\text{t}+\text{dt}}-\sqrt{\text{t}}}$ (11)

By identification with Equation (9), the slope is given by Equation (12)

$\text{α}=\frac{2{\varphi }_0}{\left(\text{E}+{\text{E}}_{\text{i}}\right)\sqrt{\text{π}}}$ (12)

The thermal effusivity of the sample can be deduced from Equation (12) by the Equation (13)

$\text{E}=\frac{2{\varphi }_0}{\text{α}\sqrt{\text{π}}}-{\text{E}}_{\text{i}}$ (13)

Over an infinitely small-time interval, the heat transfer 𝛿𝑞 crossing a probe allows to appreciate the thermal density, as noted in Equation (14).

${\varphi }_0=\frac{\text{δq}}{\text{dt}}$ (14)

Equation (15) shows the distribution that can be made of the heat transfer in case of an increase in temperature due to the flow.

${\varphi }_0\text{dt}={\text{q}}_{\text{e}}+{\text{q}}_{\text{i}}+{\text{q}}_{\text{s}}$ (15)

Here ${\text{q}}_{\text{e}}$ is the amount of heat passing through the sample; ${\text{q}}_{\text{i}}$ is the amount of heat passing through the insulating block and ${\text{q}}_{\text{s}}$ is the amount of heat passing through the probe. Joule’s law gives the amount of heat according to mass, specific heat and temperature in Equation (16).

$\text{δq}={\text{mC}}_{\text{p}}\text{dT}$ (16)

The heat flux dissipated by the probe also causes a rise in the temperature of the polystyrene block (insulation) and the material. Taking into account the definitions of Equations (15) and (16) we have Equation (17).

${\varphi }_0=\left({\text{ρC}}_{\text{p}}\text{e}+{\text{ρ}}_{\text{i}}{\text{C}}_{\text{p}}{}_{\text{i}}{\text{e}}_{\text{i}}+{\text{ρ}}_{\text{s}}{\text{C}}_{\text{p}}{}_{\text{s}}{\text{e}}_{\text{s}}\right)\frac{\text{dT}}{\text{dt}}$ (17)

By plotting the curve Δ T = f(t) and exploiting the linear part of the curve, the slope of the curve is defined by the relation Equation (18).

$\text{γ}=\frac{\text{dT}}{\text{dt}}=\frac{\text{T}\left(\text{t}+\text{dt}\right)-\text{T}\left(\text{t}\right)}{\text{dt}}$ (18)

Exploiting Equation (17) and the relation allows us to obtain the volumetric heat by Equation (19).

${\text{ρC}}_{\text{p}}=\frac{1}{\text{e}}\left(\frac{{\varphi }_0}{\text{γ}}-{\left({\text{ρC}}_{\text{p}}\text{e}\right)}_{\text{i}}-{\left({\text{ρC}}_{\text{p}}\text{e}\right)}_{\text{s}}\right)$ (19)

Inversion by the De Hoog algorithm will allow to find the theoretical temperature of the complete model. Figure 2 shows the principle diagram of the heat transfer.

Thanks to the Levenberg-Marquardt algorithm coded in MATLAB, we have an interval $\left[{\text{t}}_0;{\text{t}}_{\text{max}}\right]$on which the residual curve is centered around 0˚C, the values of $\text{E}$ and $\text{ρCp}$ (taking as initial value the pre-estimated values) which minimize the sum of the squares of the errors between the modeled value of the observed temperature ${\text{T}}_{\text{mod}}\left(\text{t}\right)$and the value obtained experimentally ${\text{T}}_{\text{exp}}(\text{t}$).

All the experimental tests have been carried out with the Douala ambient temperature (32˚C) and 68% relative humidity

The compression test is carried out according to the NBN B15-220 standard [7] .

Figure 3 gives an image of the testing machine. The measuring comparators have an uncertainty of one millimeter. The test was carried out on 15 samples.

The test is performed using two springs; the first gives the value of the compression of the ring ∆x of stiffness (K = 121740 N/mm) and the second the compression of the sample (∆e). The graph F(∆e) is plotted to obtain the Young’s modulus of the composite in compression. The force exerted on the sample is obtained using E Equation (20)

$F=K\Delta x$. (20)

Subsequently, we calculate the Young’s modulus in compression by Equation (21)

$E=\frac{e\Delta F}{Ll\Delta e}$ (21)

The expression ∆F/∆e is given by the slope of the curve F(∆e). Hence:

e: thickness of the sample, L: length of the sample and l: width of the sample.

The experimental curves are plotted by converting the video of the comparator values variations into an image as presented by [8] .

This test assesses the potential to withstand shocks without deteriorating or breaking. The test is carried out on a Charpy type machine (test specimen on two supports). The samples are cut to meet the requirements of ASTM E23. The test was carried out on 15 samples. The uncertainty of reading the test dial is 1 degree. Figure 4 gives the principle of Charpy test.

The experimental protocol consists in dropping a hammer of known mass (m) located at the end of the arm of a pendulum of length (L) rotating in a vertical plane around a horizontal axis. The pendulum is dropped from its initial position at zero speed onto a sample. Without the sample, it describes i; with the sample it describes another angle θf due to the percussion and the rupture of the sample upon contact with the mass. The resilience (K) (J/cm²) is obtained by the ratio of the absorbed energy to the cross-section of the notched area on the sample. It is given by Equation (22):

K = W/S (22)

The absorbed energy is obtained using Equation (23)

$w=w_i-w_f=P{h}_i-P{h}_f=P\left(h_i-h_f\right)=mgl\left(\cos {\theta }_f-\cos {\theta }_i\right)$ (23)

With l: arm length (32 cm), $w_i$: Potential energy of the forward pendulum (joule);

$w_f$: Potential energy of the pendulum after (joule), m: mass of the projectile (1 kg), hi: initial fall height, h_f: final fall height, S: section of the test piece (cm²).

Hardness testing may seem abstract and difficult to understand for some. Mastering the vocabulary and reading the hardness value demystifies the subject. The hardness test is defined by 3 criteria, letters and numbers. The hardness value is always placed before the test criteria. The Brinell test performed consists of applying a load for a determined time by a ball, and measuring the diameter of the imprint. The Brinell number is the result of the ratio between the applied load and the surface area of the spherical cap of the imprint. The Brinell test standard [10] : ISO 6506 and ASTM E10 were followed to perform the test. In the specific case of this work, we used a durometer to obtain the hardness value directly by reading. Figure 5 represented the hardness test principle.

Abrasion is used to determine the loss of material due to friction with another material. In this work, sandpaper was glued to the engine plate (6); in order to imitate the effect of the bottom of a boat rubbing on the beach. Abrasion is measured here using the tribometer opposite. Figure 6 presents the pin disk tribometer.

The experimental protocol implemented to determine the specific wear of the material was developed by [8] . The wear is obtained by Equation (24):

$w=\frac{1}{2\pi rn}\times \frac{1}{Fm}\times \frac{\Delta m}{\rho }$ (24)

∆m: mass variation (initial mass before the test minus final mass after the test), ρ: sample density, n: number of revolutions (engine speed is 1380 rpm) during operation for 5 seconds it makes 115 revolutions or a linear speed of 43.332 m/s (722.2 km/h or 389.1164 knots), r: the friction radius is 60 mm, Fm: The friction force due to a normal load of 55 N is given by modeling the steel bar.

Determination of the friction force involves modeling the steel bar carrying the axle, this given by Equation (25).

$Fm=\frac{3EI}{L^3}{\Delta }_{\max }$ (25)

The quadratic moment is given by Equation (26)

$I=\frac{l{e}^3}{12}$ (26)

∆max: displacement given by the comparator, E: Young’s modulus of steel 20500 Mpa, I: quadratic moment, L: length of the pawn arm (320 mm).

The average density obtained from this an average value of 0.5 g/cm³ with a relative error of 0.05 g/cm³. As for the actual density test the porosity, it is 9.8%. The low density of the composite is 0.45 g/cm³ with a relative error of 0.045 g/cm³, however, the apparent density has important in the implementation of insulation materials, however this density depends on the quantities and nature of the starting materials used., the low density of this composite is due to the higher fraction of fibers which corresponds to the result of a recent publication which shows that the density decreases with the increase in fibers [22] . This composite is lighter than particle-reinforced composites such as doum and polypropylene particles [23] and many woven-reinforced composites such as flax fiber woven [2] .

The mixture of the highly absorbent vinified raffia fiber [24] with the hydrophobic epoxy gives a resulting absorbent biocomposite resulting composite, the same observation have been done with the random vinifera fiber epoxy [25] . The composite absorbs after 100 days a quantity of water worth 12.61% of its initial weight with a standard deviation of 1.13. This absorption rate can decrease if the implementation process is improved or changed. Compared to wood in Table 1 , to other materials used in shipbuilding. This material has the lowest water absorption rate. In addition, this material is less absorbent than other composites made from raffia fiber. The analysis of the absorption kinetics shows the increase in water uptake by the material over time. The experimental curve of the absorption kinetics shows that the water absorption is rapid in the first days due to the high cellulose content and the pigs on the surface of the raffia fiber. This same phenomenon is observed in other works where the composite was made using epoxy resin, including [25] and [18] . This absorption behaves and the use of epoxy shows that it is a good thermal insulator typical of many composites [26] . Compared with the local shipbuilding timber shown in Table 1 , the developed bio composite exhibits better physical properties. The absorption behavior is depicted in Figure 7 .

One of the reasons for bio-composites in the maritime field is their contribution to thermal insulation. Figure 8 shows the result of the heat transfer test through the developed bio composite.

It is observable on the curve at the interpolations equivalent to 100 seconds that the resulting temperature exchange is 18˚C, however the real time range used for the calculation is (50 to 90 seconds) in this interval the interpolation error is minimal (minimal residual). The two curves in Figure 8 show the evolution of the heat flow through the bio composite per unit of time. All the values obtained are presented in Table 2 . The effusivity is obtained by observing the increase in temperature as a function of the square root of time; this is presented in Figure 9 .

In the interpolation interval (place of minimum error) all thermal parameters were calculated. All values obtained are summarized in Table 2 .

The similarity of the values in the table to those of the work on the determination of the thermal properties of a bio composite for construction [28] shows that the bio composite developed can be used in insulation. Likewise, these values are close to those of the hybrid composite obtained from raffia particles and resin [18] . In addition, these values are very low compared to those used in high energy efficiency buildings [29] . In addition, the shapes of the curves obtained are similar to those of other works [17] and [30] which used the same process to develop their composite. In addition, this composite offers better overall thermal insulation than some tropical woods used in shipbuilding [31] . Its thermal conductivity is also lower than that of some bio-composites for a given range of humidity values [32] and [33] , as presented in Table 3 .

Generally, Table 3 shows that the global thermal conductivity of insolate material is lower than 1 (W/m K). Like other material, this raphia vinifera/epoxy woven composite material can help for boat insolation in sub-Saharan areas.

Boat hulls are very often subjected to strong compressions due to the weight of the boat on the one hand and the water pressure on the other hand. The compression test therefore makes it possible to know the Young’s modulus in compression and the limit load that the composite can support without crushing. The samples subjected to the compression test did not undergo complete destruction. However, beyond the force value of 25 KN the deformation was no longer perceptible. The average value of Young’s modulus is E = 22.86 GPa with a standard deviation of 0.2. This is much higher than those obtained for the tensile and bending tests are. This test clearly shows the material's ability to work in compression. In other words, it is potential to be also used for the construction of the main structures (hulls) of boats. Table 4 clearly illustrates the possibility of using this composite in shipbuilding by comparing its compressive stress with those of the local woods used. Figure 10 represented the curve of compression test.

The contact due to the movement of the tide between the boats at the quay creates a shock that can lead to damage to the material used to build its hull. The resilience test was carried out experimentally on the material given in Figure 11 . The value of the resilience coefficient of the different samples.

The graph in Figure 11 gives the average value of all the points in this diagram, which is 1.238 J/Cm² with a standard deviation of 0.385. The value of the composite resilience is less than that of the composite with glass fiber and polymer matrix whose values are between (6.4 - 8.5 J/Cm²) [35] . The resilience of woven reinforcement composites like other mechanical properties depends on the nature of the fibers used., for the same implementation conditions, PBO (poly(p-phenylene-2,6-benzobisoxazole), aramid, basalt, glass, and Carbon have the highest resilience value in decreasing order and natural fibers have the lowest resilience [36] .

Boat hulls collide with rocks or other boats due to maneuvering errors. The Brinell hardness test performed using the Durometer is used to assess the material’s resistance to penetration and impact. Figure 12 shows the result of all tested samples.

The average hardness value of all composite samples is given by 233.04 BH30-2.5/187.5-15s with a standard deviation of 0.25 and variance, 233.04 represents the Brinell hardness value in N/mm²; B represents test type (Brinell); H is hardness; 15s is load application time and 30 give approximate equivalent value in Kgf of the applied test force (30 Kgf = 294.2 N).

Generally, Brinell hardness measures the depth of the imprint left by a 23 mm diameter ball, with a mass of 1 kg, thrown at 50 cm. The hardness of this material is higher than bio-sourced composites made from raffia, coconut, canarium and palm nut shell particles whose values are between (100.67 and 136.48 MPa) [37] . This shows that weaving could increase the hardness of a composite. This is the case of woven natural fibers of flax, hemp, banana, sisal, hemp, etc., all of whose mechanical properties improve compared to those of particle-reinforced composites [38] .

For wood, this test measures the resistance to punching and depending on the value obtained, the wood can be class A (soft), B (semi-hard), C (hard) and D (very hard). Compared to local woods used in shipbuilding, the developed material is admissible because its hardness is much higher than that of some of these woods. Table 4 is an illustration of this.

Friction between the hull and the beach generally causes material removal. The experimental setup we used to idealize this phenomenon assumes a boat rubbing the beach at a speed of 389.11 knots. Figure 13 illustrates the material removal by abrasion test curve.

The average abrasion (wear) rate is 0.005349 and can be considered minimal. Compared to the composite based on PEEK (polyetheretherketone) matrix and reinforced with various synthetic fibers whose friction coefficient values vary between (0.6 and 1.2) [39] . In addition, for certain materials from alloys such as carbon, silicon, manganese, iron, copper, etc., the friction coefficient is between (0.1 and 0.5) [40] . It should be noted from the results in the literature that this composite cannot be used for braking applications. However, this is a satisfactory result for the desired use, because a low wear rate means minimal material tearing. This composite could be used to build hulls. Compared to the local wood used, this class is admissible. Table 4 illustrates this. In addition to having physical, thermal and mechanical properties comparable to local wood used in shipbuilding, this composite made from woven raffia fibres cast in epoxy resin is a better thermal insulator, lighter and harder than many resin-based composites, as shown in Table 5 .