EXPERIMENTAL STUDY IN NATURAL CONVECTION

The study of thermal and ventilation parameters, obtained in a transient, laminar solar chimney of reduced dimensions, (1 < m <3) m with a square collector (side = 2m) is presented. Experimental measurements has been made to determine the temperature of the absorber and the fluid in the collector, it is shown that at the entrance of the chimney, the temperature of the absorber decreases slightly while that of the fluid is maintained at a maximum level. Temperature differences were observed up to 32°C between the atmosphere and the fluid in April. A temperature variation at the absorber depending on the stack height is presented. Temperature measurements in the chimney, at various heights depending on the axial coordinate, show a variable temperature profile. It is, from these, shown that, in the selected interval of stack height, the average speeds of output increase linearly as a function of stack height. For a chimney of 3m in height and 20cm in diameter, a maximum speed of approximately 0.7 ms -1 was observed. The lack of appropriate equipment handicaps the velocity measurement at the chimney entrance. Thus, the results of simulations with the computer code COMSOL 5.1 has confirmed temperature values measured at the chimney entrance and after this, velocity values are determined.


INTRODUCTION
Solar chimneys are traditionally used for drying agrofood products, the renewal of air in silos, barns, greenhouses and houses and more recently for generating electricity. Less than 20% of the population in the Sahelian countries have access to electricity, Ministère de l'Environnement et du Développement Durable, (2011). However, energy is one of the challenges of urban growth in West Africa. This often concerns arrears agricultural countries with low industrial base and a population using wood and charcoal as major fuels. Nevertheless, their economies often heavily depend on imports of oil and gas. However, they have one of the largest solar deposits in the world, with an average intensity of sunshine ranging between 5.5 kWh.m 2 day -1 and 6.5 kWh.m -2 day -1 and duration from 3 000 to 3 500 hours per year. But it remains very low value in use, especially as regards thermal systems. However, these socio-economic and climatic conditions are favorable for the development of solar chimneys for the production of electrical energy or the drying of agro-food products.

Experimental models
The first prototype tower was built in Manzanares in 1981, thanks to the team of Professor Schlaich, Haaf et al., (1983); Haaf et al., (1984), in Spain. It worked from 1982 to 1989. In general, the collecting surface (collector) is made of a transparent cover (made of glass or plastic or a combination of both) held at a certain height from the ground, creates a greenhouse that is open to the outskirts but attached to a chimney in the centre. The thermal draught or "chimney effect" induces a depression and draws air from the outside through the sensing surface. The speed of air movement is determined by the temperature gradient within the stack. The kinetic energy gained by the fluid is converted into mechanical energy and then into electrical energy by turbines according to Figure 1. The output power profile, in this type of solar chimney without storage, is in perfect accordance with that of daily radiation. Note that this tower was used as an information provider Haaf et al., (1983);Haaf et al., (1984);Schlaich et al., (1995).   Table 2 summarizes the thermo-hydrodynamic performances obtained by Pasumarthi and Sherif, (1998b);Golder (2003); and Akbarzadeh et al., (2009);Zhou et al., (2007); Ferreira et al., (2006); and Maia et al., (2009 Amir and Hadi, (2015) brought an innovation to the design of a solar tower of 2.6m in height, with a cylindrical collector of 0.92m in height and 0.9m in diameter. They combine hubs that can track the sun's path, reflecting the radiative flux received on the glass walls of the collector. Top speed achieved then reaches 5.12 m.s -1 . This is very important as regards the dimensions of the prototype.

Theoretical models
Many theoretical models of solar chimney have been developed, but few comprehensive experiments were conducted on prototypes outside that of Professor Schlaich, Haaf et al., (1983) to confirm their validity. Note that five simple theoretical models given in literature were identified and compared with CFD results. Firstly, the model of Chitsomboon et al., (2001) is the only one which is in perfect agreement with CFD results but underestimates plant's performances. Secondly, the model of Schlaich et al., (2005) fits well with CFD results. Thirdly, the model of Ming et al., (2006) diverges from CFD results and overestimates plant's performances. Fourthly, the model of Zhou et al., (2009) also diverges from CFD results and overestimates plant's performances. Fifthly, the model of Koonsrisuk and Chitsomboon., (2009) gives reasonable solutions comparable to Schlaich et al., (2005).
Nevertheless, it was Mullett (1987) who began the development of theoretical models of solar chimney while expressing that the overall efficiency is proportional to the chimney height. Referring to data from Manzanares prototype Padki and Sherif, (1989) elaborated models liable to produce significant energy in the medium and long term. Yan et al., (1991) referring to a comprehensive analytical model used correlations allowing them to obtain the equations for air speed, air flow rate, output power and thermo fluid efficiency. Padki and Sherif, (1992) presented a brief presentation on the geometric effects and certain operating parameters on the performance of a solar chimney. In the late 90s, Pasumarthi and Sherif, (1998a) built a reduced prototype to study the effects of various geometric parameters related to air temperature, air velocity and output power of the solar tower. Besides, Pasumarthi and Sherif, (1998b) studied the performance of a collector by increasing the base of the collector and introducing a storage system. These two innovations allowed them to increase the total output power of the whole system. In addition, a brief analysis of the economic costs of the system is presented. Bernades et al., (1999) were the first to use the CFD code (Computational Fluid Dynamic) in solar chimney convective flow study. They presented a solution for the Navier-Stokes equations and energy for a steady laminar natural convection in providing for its thermohydrodynamic behavior, from the finite volume method in generalized coordinates. This method supplied a detailed view of geometric effects, and operational optimal characteristics. Kröger and Blaine, (1999) assessed the influence of ambient conditions and pointed that the air humidity could influence the driving force with the possibility of air condensation. During the same period Buys and Kröger, (1999) developed analytical relationships to determine the pressure difference due to friction phenomena and heat transfer in order to develop a radial flow between the roof and the collector. Gannon and Von Backström, (2000) used a solar collector model by coupling the temperature and mass flow to study the friction phenomena inside the chimney, the turbine and kinetic energy loss at the chimney outlet. Simulation results of this experimental work compared with those of a small central allowed to predict the performance of a large-scale plant.
During the same year they developed a onedimensional model of a compressible flow for the calculation of thermodynamic variables depending on the chimney height, rubbing against the wall chimney, additional losses, and the location variation. Chitsomboon (2001) also developed a mathematical model; the results were in agreement with those of Schlaich et al., (2005). Bernardes and Von Backström, (2010) undertook the study of the performance of two models of chimney. Larbi et al., (2010) presented an analysis of the energy performance of a solar chimney in the strongly sunny region of Adrar (south of Algeria); it may provide up to 140-200 kW of electricity throughout the year. Moreover, Chergui et al., (2010) have developed a CFD model compatible with the Benchmark of Vahl Davis, (1983) for natural convection. Sangi et al., (2011) demonstrate that pressure decreases along the collector and that its minimum value is located at the bottom of the chimney while the latter was considered positive and increasing throughout the flow in the collector by Pastohr et al., (2004). Tayebi and Djezzar,

EXPERIMENTAL STUDY IN NATURAL CONVECTION
(2012) present a numerical model of the laminar natural convection in vorticity-stream function in hyperbolic coordinates formulation. They prove that the airflow varies not only with the roof-floor distance of the collector, but also with the Rayleigh number; allowing the manufacturer to optimize the location of its turbine. In Egypt, El Haroun, et al., (2012) devised a mathematical model that could determine the power output and predict the effects of changes in environmental conditions. Tayebi and Djezzar, (2013) also established a digital model of a solar chimney with curved junction. They showed that maximum speed is achieved at the chimney entrance. koonsrisuk and Chitsomboon, (2013), Ming et al., (2013), Putkaradze et al., (2013), showed that the geometry of solar chimneys influence their performance. Ghalamchi, et al., (2013), Kasaeian, et al., (2014) proceeded to a theoretical performance evaluation of a solar chimney using the FLUENT CFD code and validated their study by using an experimental prototype. The parameters considered concerned the inclination of the collector, the diameter of the chimney and the height of the collector inlet. Penghua Guo et al., (2014) show that for a fixed stack diameter, there is a maximum value of the collector radius beyond which the output power remains substantially constant.
In conclusion these studies confirm the feasibility of the system. It is then necessary to conduct experiments based on local meteorological data, to optimize geometry settings in order to improve the overall efficiency, and finally, to locally produce technically well controlled systems that are affordable by rural populations.

2.
Description of our prototype An experimental prototype of solar chimney shown in Figure 2 has been designed at the Thermal Renewable Energy Laboratory of the University of Ouagadougou, Burkina Faso. It consists of a square collector 2m side topped with a chimney whose height can vary between 1m and 3m. The choice of a square collector is linked to the need for simplicity in the construction of such a prototype in rural areas. The chimney with 20cm of diameter is made of mild steel 1.5 mm thick and insulated with TRAPALU-40 which is a weather resistant insulator. The collector of the tower consists of a glass roof covering an absorber. The roof of the collector is composed of four (4) panes of soda-lime glass (SiO 2 72% silica, 14% sodium hydroxide, 10% lime, and 5% alumina magnesia oxide), a square meter of surface and 5 mm thick, forming an angle of 4 ° between the collector inlet and that of the chimney. The works of Ghalamchi, et al., (2015) show that in the case of small power systems using solar chimneys, a near zero slope is the most appropriate. The absorber is a zincaluminum sheet, thickness 0.35mm, painted matte black and isolated by 20cm polystyrene at its lower surface. Numerous other materials have been used as absorbers but do not reach the temperature level obtained with this material which has an average thermal conductivity of 140W / m.K, a density of 2.8 kg.m -3 . Figure 2: View of the solar chimney equipped with its temperature sensors 3.

Experimental Protocol
Our study was first to measure temperatures from the collector inlet to the center of the chimney and then at different heights in the chimney according to Figure 3. These measurements are made with type K thermocouples connected to two temperature recorders (MIDI LOGGER GL 200) shown in figure 2. The evaluation of speed at the inlet and outlet of the chimney is effected by means of an anemometer connected to a datalogger D02003, precision AP472S2 0.05 m. s -1 for a speed between 0.1 and 0.99 m.s -1 ; and 0.15 ms -1 for a speed between 1 and 5 ms-1; and finally we measure global solar radiation with a solarimeter KIMO SL 100 of 5% accuracy. In general, curves showed temperature levels all the more so higher as the global radiation was more important according to Figure 4. It should be noted that thermocouple 5 was located at the outlet of the collector, on the absorber, at the intersection point determined by the central axis of the tube and the absorber plane shown in Figure  3.  Zhou et al., (2007) show no changes up to this point which temperature is slightly lower but illustrates an important physical phenomenon according to Figure 3. The heat transfer between the fluid and the absorber reaches its maximum between the positions of sensors 4 and 5.
This result was also observed by Ghalamchi et al., (2015). Indeed, initially, the volume of incoming cold air heats up under the effect of intense solar radiation and as long as the fluid moves towards the center, the ratio of volume of the fluid in exchange surface increases, causing a slight drop in the temperature of the absorber.

Absorber temperature evolution as a function of chimney height
An increase in chimney height results in an intensification of convective motions within the prototype. The heat transfer coefficient by convection between the fluid and the absorber increases. Then we note temperatures decreasing along the absorber according to Figure  6. On the contrary, when the height increases, the convection intensifies, temperature gradients are lower between two neighboring points, leading to an almost hyperbolic profile halfway up as shown in Figure 9.

Study of flow speed 4.3.1 Study of the release speed as a function of stack height
The preheated air at the collector enters the chimney, the significant heat energy gained turns into kinetic energy causing a convective upward flow in the chimney. For April 14, 2015, the overall radiation between 10 o'clock and 16 o'clock is shown in Figure  10. Then we measure the output speed of the chimney over time, for heights ranging from 1 to 3m. Average speeds are calculated from hourly rates. It is then observed that the output speeds grow according to chimney height as indicated in Figure 11.  The representation of the average speed during a very sunny day for different heights, confirms the previous observations.

Determination of speed in the collector and
the chimney, using the digital code COMSOL 5.1 With no great precision measuring apparatus adapted to the geometry of the tower (e.g., flexibility and telescopic rod for measurements in the duct), it is proposed to complete this study by simulating the operating conditions of our solar chimney with natural convection. The thermo-physical conditions and the boundary conditions used are summarized in table 3. The speed distribution in Figure 14, corresponding to the temperature distribution of figure 13 is presented.

4.4.
Application to the production of electricity and solar drying Assuming average value for the coefficient of performance (0.50), one can roughly estimate the expected electric power with a small turbine placed at the entrance of the chimney, in small laboratory prototype built referring to Table 4. With: α = 0.5; ρ = 1.2; S = 0.0314m² Drying of fruits and vegetables is usually done around 60°C if we want to preserve the nutritional quality of products. Some trays can be placed at the bottom of the chimney where the fluid temperature can reach 72 ° C. This level of air temperature with speeds consistently above 0.1m.s -1 is satisfactory. Indeed, in general, failing to have the considered product drying speed curves, it is always recommended to keep a temperature of the air at its arrival on the product of 10°C above the maximum temperature permissible by the product. In fact, the more water the product contains, the more water evaporation enables the cooling of the product. Thus, at the beginning of drying, the temperature must be as high as possible to ensure rapid drying. This will require the air to have a temperature of at least 70°C at the beginning of drying. At the end of the drying, it is necessary to ensure that the product temperature does not exceed 65°C, i.e. when the humidity of the product has decreased considerably.

CONCLUSION
From the study, the following conclusions could be drawn: 1. At the entry of the chimney the temperature of the absorber decreases slightly while that of the fluid is maintained at a maximum level. 2. An increase in stack height results in increased convective motions in the prototype, with an increase in fluid temperature at the loss of that of the absorber. 3. The highest speeds are achieved at the entrance of the chimney but quickly decrease in order down the axis. Flow can be enhanced in this area with the aid of curved junctions. This confirms the choice in this area by several authors for the turbine installation. 4. The ideal place for drying racks corresponds to the area above the entrance of the chimney which is the relatively high temperature seat with maximum flow rates. These conditions are favorable to mass and heat transfer. This study confirms the feasibility of the system, and also highlights the need to optimize ventilation parameters for thermal power production and foodstuffs conservation.