The emergency shelters commonly used in humanitarian relief operations are mostly made of plastic sheeting, more specifically polyethylene. Polyester, polyvinyl chloride (PVC), and polycotton (60% polyester + 40% cotton) are also available. These tarpaulins are strong, flexible, water-resistant, and even waterproof. A standard sheet of plastic sheeting consists of a woven or braided black high-density polyethylene core and a layer of low-density polyethylene laminate on each side [8] , with a minimum weight of 200 g/m². Chemical additives such as calcium carbonate are incorporated into the core and outer layers to give the tarpaulin color, soften it, modify its opacity, and enhance its resistance to UV rays [9] . Plastic sheeting can have a significant negative impact on the environment in the absence of local collection and recycling infrastructures. Figure 1 shows photographs of two types of emergency humanitarian shelter commonly used in Burkina Faso.

The prototype emergency shelter ( Figure 2 ) is 5.0 m long, 3.5 m wide, and 2.4 m high. It covers a total floor area of 17.5 m², can shelter 5 people, occupies a volume of around 42 m³ and is in direct contact with the ground. The walls are made of compressed earth blocks (C.E.B.) stabilized with rice husk ash and supported by an embedded iron framework that holds the structure firmly in place. The blocks are self-locking, making the shelter removable. The shelter can, therefore, be dismantled and reassembled elsewhere, or, for example, when the people concerned return to their home localities. The roof is made of sheet metal but insulated from the inside by a false ceiling made from an eco-material of rice husks and kapok wool.

Some values for the thermo-physical properties of the materials (corrugated sheet metal, iron louvers, tarpaulin and straw mats) making up the shelter envelope were determined by characterization tests, while others were taken from the literature [13] - [18] . They are given in Table 1 below:

Measurements were taken of: air temperature and relative humidity (inside and outside) in both types of shelter, solar radiation, mean radiant temperature of interior walls and thermal comfort indices. Various measuring instruments were used: ambient temperature and humidity sensors, model “MSR 145 S”. The parameters of hygrothermal comfort and heat index were evaluated using an “87786 WBGT Logger” equipped with a black globe thermometer to measure average radiation temperature, an anemometer to measure air speed in accordance with standard EN 13779, a thermometer to measure air temperature and a hygrometer to measure relative humidity.

For measurement, the WBGT-meter is positioned (for a seated individual) at head, trunk and leg height, or 1.5 m above the floor. Internal and external building wall, roof and floor temperatures were measured using Graphtec Midi Logger GRAPHTEC GL240 data loggers fitted with J-type thermocouples. The temperature range measured by the thermocouples varies from −100˚C to +100˚C, with an accuracy of ± 0.5˚C. Solarimeters were used to measure the overall flux density of solar radiation entering the shelter.

To define the adaptive comfort range for natural ventilation, we drew on existing studies [19] - [23] , standards (NF EN 15251, RE2020) [24] , our in-situ surveys, feedback on the subjectivity of occupants’ perception of thermal comfort and acclimatization to the hot, dry tropical climate. For our study, we decided that a comfort temperature (an operative indoor temperature) of between 20˚C and 32˚C (20˚C ≤ Top < 32˚C) is realistic for the satisfaction of 80% of occupants. The pairs of points (outdoor temperature, indoor operating temperature) are plotted on a diagram. On the basis of this range, and for periods when the shelter is occupied, the number of hours during which the shelter’s inside temperature is outside the said range is counted, and defines the time during which comfort is not achieved (discomfort time). Then, the sum of the temperature differences (hour by hour) between the indoor air temperature and the maximum comfort temperature was calculated to take into account the discomfort degree-hours (DH), expressed in ˚C∙h, it is an index of thermal discomfort intensity [18] [19] .

The heat and moisture transfer equations between the shelter and its environment, including the occupants, are established by assimilating the interior air volume and wall surfaces (interior and exterior) to temperature nodes. For modeling and simulation, the simplifying assumptions are:

1) the thermophysical properties of building materials are assumed to be constant;

2) housing wall temperatures are assumed to be uniform;

3) the surfaces (internal and external) of the habitat are assimilated to gray bodies;

4) the celestial vault behaves like a black body;

5) interior air is assumed to be homogeneous and of uniform temperature.

$\begin{array}{c}{\rho }_{se}\cdot e_{se}\cdot c_{se}\frac{\text{d}{T}_{se}}{\text{d}t}=h_{ce}\left(T_{ae}-T_{se}\right)+h_{rc}\left(T_c-T_{se}\right)+h_{rs}\left(T_s-T_{se}\right)\\ +U_{se}\left(T_{si}-T_{se}\right)+{\alpha }_{se}G\end{array}$ (1)

The coefficients were calculated using the correlations of W.H Mc Adams- [25] .

$h_{ce}=5.7+3.8v$ (2)

$h_{rc}=\sigma {\epsilon }_{se}\left(T_c^2+T_{se}^2\right)\left(T_c+T_{se}\right)\left(\frac{1+\cos \beta }{2}\right)$ (3)

$h_{rs}=\sigma {\epsilon }_{se}\left(T_s^2+T_{se}^2\right)\left(T_s+T_{se}\right)\left(\frac{1-\cos \beta }{2}\right)$ (4)

$T_c=0.0552T_{ae}^{1,5}$ (5)

$T_s=T_{ae}+2$ (6)

${\rho }_{si}\cdot e_{si}\cdot c_{si}\frac{\text{d}{T}_{si}}{\text{d}t}=U_{si}\left(T_{se}-T_{si}\right)+h_{ci}\left(T_{ai}-T_{si}\right)+h_{rm}\left(T_{rm}-T_{si}\right)$ (7)

$h_{ci}=a{\left|T_{si}-T_{ai}\right|}^n+b$ (8)

$h_{rm}=\sigma {\epsilon }_{si}\left(T_{rm}^2+T_{si}^2\right)\left(T_{rm}+T_{si}\right)$ (9)

$T_{rm}=\frac{{{\displaystyle \sum }}_{p=1}^{N_{pi}}\left(S_{si,p}\times T_{Si}\right)}{{{\displaystyle \sum }}_{p=1}^{N_{pi}}{S}_{si,p}}$ (10)

${\rho }_{ai}{V}_{ai}{c}_{ai}\frac{\text{d}{T}_{ai}}{\text{d}t}=\underset{p=1}{\overset{N_{pi}}{{\displaystyle \sum }}}{h}_{ci,p}{S}_{si,p}\left(T_{ai}-T_{si,p}\right)+{\rho }_{ai}{c}_{ai}{\stackrel{\dot{}}{Q}}_v\left(T_{ae}-T_{ai}\right)+{\Phi }_{\mathrm{int}}$ (11)

The overall heat transfer balance in the shelter is the coupled system of equations (1), (7), (10) and (11), comprising all the balances for the walls, the interior air volume and the floor. The system of differential equations is then discretized at time step ∆t = 1 h using the finite-difference method according to the implicit scheme, and then translated into matrix form according to Equation (12).

$\left(\frac{C}{\Delta t}-A\right)\ast T\left(t+\Delta t\right)=\frac{C}{\Delta t}\ast T\left(t\right)+B\ast U\left(t+\Delta t\right)$ (12)

The resulting discretized system is solved by an iterative Gauss-Seidel method, assuming that at the initial time t0 of the simulation, all temperatures are equal to room temperature. At each time step Δt, heat transfer coefficients are calculated at time t + Δt. The temperature values are then calculated and compared with the previous values. If the residual (absolute error) between these values is greater than the chosen tolerance (ε_T = 0.1˚C), the calculated values replace the previous ones, and so on until the temperatures converge.

The study area is the town of Kongoussi, around 110 km north of Ouagadougou, the capital of Burkina Faso. Its geographical coordinates are: latitude 13.32˚ North, longitude 1.53˚ West, and altitude 330 m. The climate is tropical, hot, and dry, with a very hot period lasting around three months (March, April, May). Average daily maximum temperatures exceed 38˚C (min 30˚C and max 42˚C) for a typical 10-year period (2014-2023). Figure 3 shows the evolution of minimum (TaeMin), maximum (TaeMax), and average (TaeMoy) outdoor ambient temperature values for the typical year. April is the hottest month of the year, with average daily maximum temperatures in excess of 40˚C and minimum temperatures of 28˚C. The cool season lasts around two months (December, January), with average daily maximum temperatures below 33˚C. January is the coldest month of the year, with average minimum temperatures of 18˚C and maximums of 32˚C.

1) Experimental measurements on the humanitarian plastic sheeting tent (TBP)

The measurement campaign lasted 22 days (from October 27 to November 17, 2024). Figure 4 compares the simulated and measured temperatures of the air inside the plastic sheeting tent (TBP) for a seven-day measurement period. The minimum temperature is 18.2˚C (simulated value 21.6˚C), the maximum temperature is 45.4˚C (simulated value 42.7˚C), and the mean temperature is 29.6˚C (simulated value 30.9˚C). The mean absolute error (MAE) is 3.9˚C, the root mean square error (RMSE) is 4.2˚C, the mean relative error does not exceed 14.3%, and the coefficient of determination is 0.89. The linear regression line is shown in Figure 5 .

2) Shelter thermal damping and damping factor

Figure 6 shows, for the typical year, the evolution of emergency shelter temperatures as a function of time and the damping of the heat wave induced by the two shelters’ envelopes. We found that the inside air temperature of the plastic sheeting tent (TBP) was generally higher than that of the outside air during the day (6 a.m. to 6 p.m.). On the other hand, the prototype made from bio-sourced materials (PMB) induces a damping of up to 8.4˚C compared with the outside temperature; the annual average is 5.8˚C.

Figure 7 shows the temperature of the two shelters on the hottest day of the typical year. It also shows the damping of the thermal wave induced by the prototype made of bio-sourced materials in relation to the outside air temperature: the damping is 6.0˚C for the PMB.

The PMB prototype has a damping factor f ranging from 0.12 to 0.28, with an average damping factor of 0.14. It therefore generates a reduction in thermal amplitude ranging from 71% (f = 0.29) to 88% (f = 0.12), with an average reduction of 86% (f = 0.14).

3) Thermal phase shift of the shelters

Figure 8 shows the frequency, in number of days, of the thermal phase shifts of the two shelters.

For the plastic sheeting tent (TBP), the maximum thermal phase shift is 1 h (133 days of the year are concerned, i.e. 36%), while the prototype made from biosourced materials induces a thermal phase shift that varies between 3 h and 15 h, with an average of 5.6 h. Of the 365 days of the year, 61 days (17%) have a thermal phase shift of 3 h, 206 days (56%) have a phase shift of 4 h, 44 days (12%) have a phase shift of 4 h and 54 days (15%) have a thermal phase shift of 15 h. Figure 9 shows the thermal phase shift of the prototype on the hottest day, at 5 h.

1) Intensity of thermal discomfort in shelters

Figure 10 shows, for the warmest period of the year, the intensity of discomfort in shelters in terms of discomfort degree-hours as defined by RE2020. For the same length of occupancy over the whole year, the thermal discomfort intensity (in discomfort degree-hours) is 33,886˚C∙h for the plastic sheeting tent (TBP) and 10,513˚C∙h for the prototype made from bio-sourced materials (PMB). The PMB reduces thermal discomfort by 69% (hot period 55%).

We can see that the plastic sheeting tent presents a more intense thermal discomfort, whatever the month of the year ( Figure 11 ). The good comfort of the biobased prototypes is due to the good inertia of the biobased materials (clearly superior to that of the plastic sheeting) and to the insulation of the roof by the biobased thermal corrector (which insulates better than the straw mats of the tent).

2) Duration of thermal discomfort in shelters

To determine the duration of hot discomfort in shelters (temperature above 30˚C at night and 32˚C during the day), we counted the number of hours during which the indoor operating temperature exceeded these maximum comfort temperatures. The percentage of uncomfortable hours in relation to the total occupancy time of each shelter (8760 h in the year and 2208 h in the hot season) is given in Table 2 .

The plastic sheeting tent (TBP) has a relative discomfort time of around 62% of total occupancy time during the year. The prototype (PMB) reduces discomfort to 45% of annual occupancy time. We also note that it is only during the hottest months (March, April, May) that the PTB prototype registers higher discomfort times than the plastic sheeting tent.

The percentages of discomfort hours in relation to total occupancy time, day and night, are given in Table 3 . Over the year, the percentage of discomfort duration is 76% for the TBP and 30% for the PMB (61% comfort improvement) for the PMB. However, at night, the prototype slightly prolongs the duration of hot discomfort while considerably reducing its intensity.