Effect of sand and crosswind on the performance of solar chimney power plant

A solar chimney power plant (SCPP) is suitable to be deployed in desert areas, but there are few studies on the impact of desert wind and sand. In this paper, numerical simulations for the Spanish prototype SCPP were conducted to investigate desert environmental parameters, including sand movement, crosswind, and sand deposition, on the output performance of SCPP. The results indicate that the main erosion area of the turbine blade in the SCPP is the leading edge of the blade pressure surface. The erosion area of the suction surface is much less than that of the pressure surface, mainly concentrated on the leading edge. The output power of SCPP is reduced by 73.3% because of the crosswind and it decreases by 16.9% after the windbreak wall is installed. The sand and dust deposition would affect the transmissivity of the solar radiation through the collector cover. In a dusty environment for 1 month, the system output power decreased by 12.9%, and the collector efficiency decreased from 34% to 23% due to the sand deposition. These results show that the sand has a considerable negative effect on the performance of SCPP in ways of sand erosion and deposition. Sand removal is necessary for SCPP.

efficiency of the Carnot cycle, and the output power is greatly affected by the ambient temperature. To get considerable output power, the size of SCPP should be large, which brings huge land consumption and construction difficulties. Therefore, it is difficult for the large-scale commercialization of SCPP.
In 1982, the world's first SCPP was constructed in La Mancha, Spain. 2 The Spanish prototype plant was set in a desert, where the climate is dry, solar radiation is high, and day-night temperature difference is large. However, the test results show that the installation had some bad distractions: (a) The wind velocity is usually over 20 m/s and sometimes the highest wind speed could reach 30 m/s. (b) The test place was at Calina, the notorious dust haze of central Spain. (c) There was a gravel plant near the test place, which would cause dust clouds at the installation. Haaf 3 pointed out that solar radiation, ambient temperature, crosswind, and sand have influences on the performance of the SCPP.
An SCPP is suitable for a desert with abundant solar energy, such as the west-north region of China, with good solar radiation and a cheap area, which is a good choice for the SCPP. Dai et al. 4 proposed to build SCPPs in remote rural areas in northwest China. They chose three counties in the Ning Xia Hui Autonomous region, namely, Yinchuan, Pingluo, and Helan, where solar radiation is better as the construction pilot areas of SCPP, and studied the plant size, the solar radiance, and the ambient temperature on the system power output. Cao et al. 5 presented to construct a 5-MW sloped SCPP for Lanzhou, a city in Northwest China. They found that the sloped SCPP has low power efficiency, but it is environmentally friendly and easy to manage, which is meant to solve the pollution and environmental problems in northwest China. Guo et al. 6 proposed a comprehensive theoretical model considering the hourly variation of solar radiation. Based on this model, they designed four 100-MW SCPPs in Hami, Xinjiang, to get the most cost-effective one. According to the results, to meet the power demand of the whole Xinjiang region, only 14% of unused land in the Hami region is needed to construct SCPPs. In 2019, Guo et al. 7 took Yinchuan 10-MW SCPP in northwest China as an example and carried out a technical and economic analysis of SCPP to study the feasibility of building SCPP in China. The research shows that SCPP is economically competitive with wind power and solar photovoltaics. The results of the economic analysis highlight the economic feasibility of building SCPP in China under the conditions of abundant solar energy and land resources and favorable loans. In addition to the above theoretical ideas, a relatively large operational solar chimney power station exists in the Jinshawan Desert, Wuhai, Inner Mongolia, China. 8,9 This plant was operated and connected to the grid in October 2010. However, due to the limited altitude of the airport, the chimney height of the Jinshawan power plant is only 53 m, which leads to very low power generation efficiency, and its output power is only 3 kW.
Since SCPP is greatly affected by environmental factors, the influence of the desert environmental factors must be considered when constructing SCPP in desert areas. Solar radiation is the only energy source of the system, and there are many studies on the effect of solar irradiance on SCPP. The effects of ambient crosswind on the performance of SCPP have also been investigated.
The study conducted by Zhou et al. 10 and Zhou and Xu 11 shows that the ambient crosswind could influence SCPP in three ways: (1) by heat loss from the collector roof to the environment; (2) by blowing the indoor heated air to the outside of the collector instead of up through the chimney; (3) by producing a suction effect through the chimney outlet. The first and second effects have negative effects on the system, while the third effect may increase the power output of the system. Pretorius 12 points out that ambient wind can increase the negative pressure at the chimney outlet, which increases the output power of the plant. Serag-Eldin 13 studied the ambient wind effects on the SCPP by using computational fluid dynamics (CFD). The numerical results show that the influence of the ambient wind on SCPP cannot be ignored, and the system performance will be reduced in a strong wind environment. Under weak wind conditions, the performance of SCPP would be significantly reduced. The negative effect can be relatively alleviated by lowering the inlet height of the collector, but it will lead to a decrease in the available pressure drop of the system. Numerical simulation was used by Ming et al. 14 to examine the influence of the ambient crosswind on the SCPP performance. The results show that the weak environmental crosswind will worsen the flow field in the collector and reduce the system output power. If the ambient crosswind is strong enough, the mass flow rate and output power of the system can be increased. Taking the size of the Spanish prototype SCPP as an example, when the ambient wind at the chimney outlet is greater than 15 m/s, the output power increases slightly with the increase in wind speed. Later, Ming et al. 15 tried to weaken the negative effect of the ambient wind by setting a blockage at the collector inlet and verified the positive effect of blockage by using CFD simulation. Zhu et al. 16 designed three kinds of windshield layouts to reduce the influence of the ambient wind on the performance of solar chimney air clean towers. The effects of the three kinds of windshield layouts were compared by numerical simulation. Jafarifar et al. 17 mainly focused on the positive effect of the crosswind on SCPP and tried to make use of the positive effect of the environmental wind on SCPP to compensate for the negative effect of solar radiance weakening. Using the Orkney Islands in the United Kingdom as an example, they used numerical simulations to study how strong the ambient winds would be to achieve this goal. Numerical results show that the system efficiency can be increased by more than 50% under strong and stable crosswinds. RahimiLarki et al. 18 studied the effect of ambient wind on the performance of a tilt-chimney SCPP by using a combination of experimental studies and numerical simulations. The results show that the inclined chimney has a negative effect on the system performance under low-speed wind conditions. Arzpeyma et al. 19 proposed a novel chimney outlet setting to eliminate the adverse effects of ambient wind on the chimney outlet airflow. In the novel SCPP, the oblique angle at the chimney outlet is adjusted to decrease the throttling effect of the ambient wind, increasing the efficiency of an SCPP. Later, by using numerical simulation, they investigated the effect different oblique angles had on the performance of an SCPP. 20 The results showed that the efficiency of SCPP decreased as the oblique angle changed from 27°to 45°. They point out that a suitable oblique angle relies on the wind velocity and needs to be increased as the wind velocity increases.
Haaf 3 studied the influence of dust deposition on the light transmittance of different collector cover materials as early as 1984 than with solar radiation and environmental wind. The existing studies show that researchers pay little attention to the influence of sand and dust. Sand and dust affect SCPP in two ways. On the one hand, in the actual operation of SCPP, sand particles in motion and impeller friction, collision, rebound, wear, and deposition affect the aerodynamic performance of the turbine impeller system. On the other hand, sand and dust will deposit and cover the collector cover plate, resulting in the reduction of solar radiation intensity entering the collector.
Chaichan et al. 21 built a small solar chimney test device in Baghdad, Iraq in 2018, and exposed SCPP to the natural environment to accumulate dust, to explore the influence of dust and pollutant deposition on the performance of SCPP. The comparative test results of cleaning and dust accumulation of collector cover show that dust deposition reduces the solar radiation entering the system, resulting in lower system efficiency. Thus, it is necessary to clean the collector cover to maintain the higher efficiency of the system.
Although SCPP is suitable for desert areas, at present, there are few studies on the impact of desert wind and sand. In addition, there is more focus on the effect of sand deposition on the collector cover plate than on the impact of sand on the key component of SCPP (turbine). This is not conducive to the popularization of SCPP. Therefore, the influence of desert sand movement should be considered in the design of SCPP.
In this paper, the influence of the desert environmental factors on the SCPP performance was explored by using numerical simulation. The main contents are as follows: (a) the influence of sand erosion on the SCPP turbine was analyzed; (b) the influence of the sand deposition was studied; (c) the effects of ambient crosswind were analyzed and the feasibility of wind-proof wall to weaken the negative influence of environmental crosswind was also explored.

| Governing equations of fluid motion
The governing equations for the airflow in SCPP can be written as Continuity equation: Momentum equation: The Boussinesq model was adopted in this simulation as where ρ 0 is the density of ambient air, T 0 is the ambient temperature, and β is the thermal expansion coefficient. Energy equation: The RNG k ε − model is used to simulate the turbulent flow and the equations for the RNG k ε − model are as follows:

| Particle motion model
The particle tracking in CFD simulation is based on the Eulerian-Lagrangian methodology. In the dispersed phase model, solid particle trajectories are characterized in Lagrangian coordinates. When the density of the fluid is much lower than that of particles, the pressure gradient force can be ignored. The density of the airflow in SCPP is much smaller than that of the sand particles, so the effect of pressure gradient force on the particles is not considered. According to Newton's second law, the motion equation of particles is as follows: in which m p is the mass of the particle;  V f and  V p are the fluid and particle velocities; ρ f and ρ p are the fluid and particle densities;  g is the acceleration of gravity; p is the drag force per unit solid mass; and F D is the inverse of particle relaxation time, which is given as where d p is the particle diameter, μ is the dynamic viscosity, and C D is the coefficient of drag force, which is given as where Re s is the particle Reynolds number, which is given as

| Particle erosion model
The erosion rate E f is defined as the mass of removed material per unit of area per unit of time, which is given by where A f is the face area; ṁπ is the particle mass flow rate that collides with the face; and e r is the erosion ratio. In Pereira et al., 22 the erosion rates are expressed in penetration ratio E r , and the relationship of E f and E r is as follows: where ṁp is the mass flow rate of the sand and ρ e is the density of the wall material. The e r is calculated by Oka model 23,24 as where α is the impact angle; V p and d p is the particle impact velocity and particle diameter, respectively; V ref and d ref is the reference velocity and particle reference diameter; k 2 and k 3 is velocity and diameter exponents, respectively; and e 90 is the reference erosion ratio at 90°i mpact angle, which is given as The function of the impact angle is where H V is the wall material Vickers hardness in unit of GPa and n 1 and n 2 are angle function constants. The coefficient relationships for perpendicular and parallel velocity components are given by 25 The parameters involved in the above model in this paper, which come from ANSYS Help 19.0 and Oka et al., 23 are given in Table 1.

| Dust deposition CFD model
The generalized equation governing the entire system can be expressed as where ϕ is the generalized dependent variable, which can represent 1, v, T , k, and ε of the continuity equation, Navier-Stoke equation, and energy equation; Γ ϕ is the universal diffusivity and S ϕ is the universal source term. The blocking effect of the collector roof on the ground long-wave radiation is the key aspect governing the heating function of an SCPP. As the discrete ordinate (DO) radiation model can be used to model semitransparent and opaque walls, this model is adopted to solve the radiative transfer equation. The radiative heat transfer equation is given by where  r is the position vector,  s is the direction vector,  s ′ is the scattering direction vector, α is the spectral absorption coefficient, n is the refractive index, σ s is the scattering coefficient, σ is the Stefan-Boltzmann constant, I is the spectral intensity, ϕ is the phase function, and Ω′ is the solid angle.

| Sand erosion simulation
(1) Geometry model The main dimensions of the physical model are from the Spanish prototype, and the detailed parameters are shown in Table 2. In the sand erosion simulation, the turbine model was taken from the model of Tingzhen et al., 26 which was used in the previous study. 27 (2) Boundary conditions and mesh distribution The main boundary conditions are shown in Table 3. The inlet and outlet of SCPP were set as pressure-inlet and pressure-outlet boundary, respectively. The turbine wall, chimney wall, ground wall, and collector wall were set as a reflected wall. Particles rebound at these walls and their momentum changes. The chimney outlet adopts an escape boundary, and the calculation of the particle trajectory is terminated. According to Li et al., 29 the sand particle mass concentration ranges from 10 3 to 10 6 μg/m 3 . In this paper, the particle mass concentrations were selected as 8.3 × 10 3 , 10 4 , 10 5 , and 10 6 μg/m 3 and the particle diameters were selected as 5, 10, 20, 30, 50, 70, 90, and 110 μm.
A multiblock mesh is applied to model the computational domain. This mesh consists of three different parts that are interconnected by interfaces between the turbine zone and the domain before and after the turbine. The mesh distribution is shown in Figure 2. As shown in Figure 2, unstructured mesh (tetrahedral mesh) is adopted for the turbine zone, while the meshes of the other zones in the computational domain are structural hexahedral grids. To get a grid-independent solution, the grid independence test was conducted, in which the numbers of grids are 2,016,679, 2,908,499, 4,840,590, 6,338,979, respectively. Numerical results indicated that the volume flow rates at the chimney outlet are 669.55, 643.78, 638.29, and 636.09 kg/s, respectively. According to this analysis, the grid number of 4,840,590 is selected in this paper. (

3) Validation
The simulation model of SCPP was validated in a previous study. 27 In this paper, the erosion simulation model was verified by the 90°elbow erosion simulation in Pereira et al. 22 The elbow model is shown in Figure 3 and the parameters in the validation simulation case are shown in Table 4.
The present simulation result and that of Pereira et al. 22 were compared. From Figure 4A, the maximum erosion position occurs in the area where the elbow is biased toward the outlet, and the erosion area is Y-shaped. The maximum penetration ratio calculated by the present model is 3.16 × 10 −5 m/kg, while that of the Gabriel model is 7.90 × 10 −5 m/kg.
As shown in Figure 4B, the location and erosion ratio calculated in this paper are basically consistent with Gabriel's results, which proves the accuracy of the numerical erosion model in this paper.

| Ambient crosswind simulation
The geometrical model and boundary conditions of the ambient crosswind simulation are shown in Figure 5. Arzpeyma et al. 20 have studied the influence of the external computational domain size on the accuracy of numerical simulation. According to their calculation results, the length and width of the computational domain in this paper are set to 400 m and the height to 250 m, which can ensure the accuracy of the simulation. 14 The windbreak wall is an annular thick wall around the collector inlet, with a radius of 124 m and a wall height of 2 m, equal to the height of the collector inlet.
The boundary conditions and computational mesh of the crosswind simulation cases are as follows: The inlet boundary is set as velocity inlet, and its temperature is the ambient temperature (298.15 K). The ambient crosswind speed follows the logarithmic law of the wind speed profile, and its velocity with the height is as follows: where v 0 is the wind speed at H 0 above the ground; v is the wind speed at H above the ground; n is the wind shear index. Based on the fitting of measured wind speed data in a typical desert environment, the wind velocity distribution as a function of height is given by 30 The inlet boundary: v w = = 0, u H = 1.5216 0.4226 , which was compiled into UDF (user-defined function).  27 The ambient ground outside the collector is assumed to be an isothermal wall and its temperature is 298 K. The side wall is set as the adiabatic wall.

(5) Windbreak wall
The windbreak wall is set as an adiabatic boundary.
Thus, the output power of this model is where η tur is the energy conversion efficiency, which is set as 0.72, 15 V̇is volume flow at the turbine domain inlet, and v tur is the airflow velocity of the turbine face. The computational domain adopts the hexahedral mesh method, and the mesh distribution is shown in Figure 6. Besides, the quality of the mesh has been improved near the walls. The minimum quality of the mesh is 0.46. To evaluate the sensitivity of mesh number in the simulation, the airflow velocity at the chimney outlet was calculated for the grid numbers 1,380,665, 1,981,657, 2,531,817, and 3,121,856. The airflow velocity is 6.90, 7.16, 7.28, and 7.24 m/s, respectively. It is found that grid number 2,531,817 can meet the calculation requirements.

| Dust deposition simulation
The geometry model and environmental data in the dust deposition simulation are still based on the Spanish prototype, which is the same as that of the erosion model. In this paper, the simulation is carried out under a solar radiation of 850 W/m 2 and an ambient temperature of 298.15 K.
The main boundary conditions are also shown in Table 3. The difference is that radiation heat transfer needs to be considered in this numerical simulation. Hence, DO radiation model and solar ray tracing are used to simulate the radiation heat transfer and solar radiation, respectively. The collector cover is a semitransparent glass wall, with a thickness of 5 mm, and was set as a mixed boundary. The physical properties of the main structural materials and working fluid are shown in Table 5. 32 The simulation method is verified based on experimental data Chaichan et al. 21 The solar radiation is 725 W/m 2 and the airflow velocity at the chimney outlet is 1.18 m/s when the collector cover is clean. The airflow velocity at the chimney outlet changes to 0.76 m/s due to the dust deposition. In the numerical simulation, the transmittance coefficient is set to 0.8 and 0.66, respectively, which represent the cases of clean cover and dirty cover. The chimney outlet airflow velocity, calculated by the simulation, is 1.23 and 0.81 m/s, respectively. This indicated that this numerical method could predict the influence of dust deposition on SCPP performance. The airflow velocity obtained by the numerical simulation is slightly higher than the experimental results. This is mainly because the steady-state numerical simulation is adopted in this paper, that is, under the given solar radiation condition, the collector cover, the airflow, and the ground heat

| Turbine blades erosion
As can be seen from Figure 7, the main erosion area is the front of the blade pressure surface. The erosion of the blade tip is more serious than that of the root. The erosion rate decreases gradually from the top to the root. In addition, the erosion rate of the pressure surface is higher than that of the suction surface. The main erosion area of the suction surface is the leading edge. The erosion rate is mainly related to the impact angle and particle velocity. The greater the impact angle and particle velocity, the greater the erosion rate. Sand particles come from the front and bottom of the blade. Particles directly hit the leading edge of the turbine blade, while some particles hit the pressure surface of the blade. The particle velocity at the tip of the blade is higher than that at the root, so the erosion rate at the tip of the blade is also higher. At the same time, the impact angle of the particles colliding with the leading edge of the blade is large, leading to a large erosion rate of the leading edge of the blade.
For the suction surface of the blade, the particles can only collide with the front edge of the suction surface. After the collision, the particles bounced off the blade and were carried away by the airflow. Therefore, erosion of the suction surface of the blade only occurs on the leading edge.
The trajectory of a single sand particle hitting the blade is helpful to verify the above analysis. As shown in Figure 8, the particle enters the turbine rotational domain and is driven to rotate by the airflow. After hitting the front edge of the suction surface of the blade, the particle bounces and is taken away, without secondary collision with the blade. However, the particle that hits the pressure surface bounces obviously. This particle does not break through the boundary layer again and collides with the wall for the second time, but moves along a plane parallel to the pressure surface. It flows out of the rotating area at the gap between the blades.
(1) Particle mass concentration It can be found from Figures 9 and 10 that the particle mass concentration has little effect on the erosion position, which is still mainly located at the leading edge of the blade, and the erosion area decreases from the tip to the root of the blade. The erosion area of the suction surface is very small, only concentrated at the leading edge. Most of the erosion areas are the blade pressure surface, and the erosion area of the pressure surface increases slightly with the increase of the particle mass concentration.
It can be seen from Figure 11 that the sand particle mass concentration has a great influence on the erosion rate. The average erosion rate increases linearly with the increase of the sand particle mass concentration. When the particle mass concentration is 8.3 × 10 3 μg/m 3 , the average erosion rate of the blades is 2.36 × 10 −10 kg/ (m 2 s). When the particle mass concentration is 8.3 × 10 6 μg/m 3 , the erosion rate of the blades reaches 3.71 × 10 −7 kg/(m 2 s). The main reason is that when the particle diameter is constant, the number of sand particles hitting the blades per unit of time increases  As a result, the erosion rate of the blades also increases linearly.
(2) Particle diameter The blade pressure surface is the main erosion area. Figure 12 shows the effect of particle diameter on the pressure surface erosion rate distribution. It can be found from Figure 12 that the erosion area of the blade pressure surface gradually increases with the increase of the particle diameter. When the particle diameter exceeds 70 μm, the erosion area decreases slightly, and the main erosion area gradually moves from the tip of the blade to the root. When the particle diameter is 110 μm, there is almost no erosion at the blade tip, but the erosion at the blade root is more intensive and continuous.
When the particle diameter is too small, the momentum of the particles is also too small. The sand particles cannot pass through the boundary layer on the blade wall, so it is difficult to hit the blade surface. With the increase in particle size, the mass of the sand particles increases. During the upward movement of particles driven by airflow, the velocity of heavier particles is very small, leading to a small erosion rate on the pressure surface of the blade.
From Figure 13, the average erosion rate of the blade surface increases first with the increase of particle size and basically remains the same after the particle size increases to 70 μm. This is because the increase of the dust particle size would lead to the erosion wear rate of the single particle increasing. When the particle diameter is less than 70 μm, the erosion wear area also increases, so the surface erosion rate increases. However, when the particle size exceeds 70 μm, the erosion wear area decreases, and the balance between the two causes the erosion wear rate of the blade surface to remain unchanged.

| Crosswind
To study the influence of the ambient crosswind and windbreak wall, four calculation conditions are set for simulation. The description of these four cases is shown in Table 6. which is 14.13 m/s according to Equation (23). If v 195 is 0 m/s, it means that there is no crosswind in the numerical simulation. Thus, Case 1 represents the case without considering the crosswind. Case 2 takes the crosswind effect into consideration and does not set the windbreak wall at the collector inlet, while in Case 3, the influence of the windbreak wall on the crosswind is investigated. The aim of Case 4 is to study the influence of crosswind on the SCPP performance at night. In Case 4, the ground heat flux is regarded as zero. As shown in Figure 14A, the airflow velocity in the chimney of Case 1 is the largest. Without crosswind, the airflow in the system still has a high velocity after flowing out of the chimney, forming a free jet in the chimney exit area. The airflow at the chimney exit presents a typical free jet structure.
Compare Figure 14A with 14B, it can be found that the crosswind destroys the jet structure of the airflow at the exit of the chimney, and the airflow at the exit of the chimney is trapped by the crosswind and flows to the rear of the chimney. The airflow velocity decreases obviously due to the crosswind, and the airflow velocity in the collector in the −X area is almost zero, so the velocity distribution in the system is no longer symmetrical. This is shown in Figure 15A.
By comparing Figure 14C with 14D, it can be found that the windbreak wall can restrain the negative effect of the crosswind. The airflow velocity in the chimney is increased, and the airflow velocity distribution in the collector becomes symmetrical, as shown in Figure 15. For Case 4, the air temperature in SCPP is the same as the ambient temperature, and there is no air density difference. In theory, the air velocity in the chimney should be close to 0 m/s at this time. However, there is still air flowing in the chimney. At this time, the system can be regarded as a wind power generation device, and the crosswind plays a positive role.
In Case 1, after the hot air flows out of the chimney outlet, it is jet-like and keeps a high temperature, as shown in Figure 16. After that, the hot airflow exchanges heat with the outside air, and its temperature gradually decreases. Under this condition, the air temperature in the system and air temperature rises can reach 318.9 and 20.8 K, respectively. In Case 2, the crosswind makes the heat loss of the hot air flow in the collector, resulting in a lower air temperature in the chimney. After the hot air flows out of the chimney, it was trapped by the ambient crosswind and flowed to the back of the chimney. The air temperature in the chimney was only 308.3 K as a result of the crosswind. It can be found that the airflow temperature in the +X direction collector area is relatively high. Crosswind causes the airflow temperature in the chimney to decrease significantly, the temperature difference between inside and outside the system, and the airflow density difference, resulting in the airflow pressure drop in the chimney and the output power of the system, which is not conducive to the operation of the system. As shown in Figure 16C, the windbreak wall makes the temperature in the windward area of the collector rise slightly, and the air temperature rise at the chimney inlet is higher than that of Case 2.
The wind reduced the available pressure drop of the chimney by about 40%, and the airflow velocity in the chimney decreased to 4.6 m/s. For Cases 1, 2, and 3, the output power is 38.87, 10.36, and 32.28 kW, respectively. Due to the influence of crosswind, the output power decreased by 73.3% and only decreased by 16.9% after the installation of the wind wall.
In Case 4, the airflow velocity in the chimney is 2.41 m/s, which can provide an available pressure drop of about 24 Pa, and its theoretical available power is 3.65 kW. This indicated that the crosswind at night is beneficial to improve power output.
Considering the effect of sand-laden crosswinds on the system, the sand particles are coupled to the ambient wind simulation. The result shows that the windbreak wall reduces the sediment concentration of the chimney outlet airflow from 0.33% to 0.058%. It shows that the windbreak wall can block some sand from entering the system and protect the key parts of SCPP.
The windbreak wall under no-wind conditions also has a negative effect. The results show that under nowind conditions, the output power of the system is 36.06 kW, decreasing by about 7% compared with that of Case1.

| Sand and dust deposition
Dust deposition will affect the transmittance of the collector cover in SCPP. The accumulation of dust on the collector cover is generally uniform, eventually forming a thin dust layer. The thickness of the dust layer is closely related to the inclination of the collector cover and exposure days. According to Hegazy, 33 when the glass plate is parallel to the ground, the relationship between the glass plate transmittance in the desert climate and the exposure days is as follows: where τ is the glass transmittance and τ clean is the transmittance of clean glass, which is 0.92 in this paper. According to Equation (25), after 30 days of exposure, the transmittance of the glass cover plate decreases from 0.92 to 0.67, decreasing by 27.17%. The decrease in the collector cover transmittance will inevitably weaken the solar radiation entering the collector. This would affect the performance of SCPP.
After 30 days exposure, the airflow temperature rises at the chimney inlet decreased from 22.0 to 16.2 K (Figure 17). Sand deposition leads to a decrease in the ground heat flux. The ground temperature decreases, and the heating of the air is weakened, resulting in a decrease in airflow temperature.
When the collector is clean, the output power of SCPP is 39 kW, which is close to the test data of the Spanish prototype. After 15 days, the output power is reduced to 31 kW. After 30 days, the output power was only 27 kW, decreasing by 12.9%. This indicates that dust deposition has a great influence on the performance of F I G U R E 17 Effect of the sand deposition on temperature rise and power output F I G U R E 18 Effect of the sand deposition on the collector efficiency SCPP. In only 1 month, the power output was seriously reduced due to the dust deposition.
Analysis of the influence of dust deposition on collector efficiency is helpful to explain the above phenomenon. The collector efficiency can be expressed as where c p m , , ṁ, T Δ , R col , and I is the specific heat capacity, mass flow rate, air temperature rise, collector radius, and incident solar radiation, respectively.
The curve of the collector efficiency with the exposure days is shown in Figure 18. When the collector cover plate is clean, the collector efficiency is 34%. The collector efficiency measured by the Spanish prototype SCPP is about 30%. This further demonstrates the rationality of the numerical simulation method. After 30 days, the collector efficiency was only 23%, decreasing by 11 percentage points. The collector is the energyabsorbing unit in SCPP, which has a direct impact on its performance. The sand deposition has a huge negative effect on the collector efficiency, leading to a sharp decline in the system performance. Therefore, it is necessary to consider adding a sand and dust removal device for regular dust removal for SCPP.

| CONCLUSION
In this paper, the effects of the desert environmental parameters, including sand movement, crosswind, and sand deposition, on SCPP are studied by using CFD fluent. The influence of sand particle erosion on the turbine and the influence of sand and dust deposition of the collector cover on the output power and thermal efficiency of SCPP are analyzed. In addition, the effect of the windbreak wall at the entrance of the collector on counteracting the negative effect of crosswinds is explored. The main conclusions are as follows: (1) The main erosion area of the turbine blade is the leading edge of the blade pressure surface. The erosion rate gradually decreases from blade tip to root. The erosion area of the suction surface is much less than that of the pressure surface, mainly concentrated on the leading edge. (2) The sand particle concentration has little effect on the erosion position of turbine blades, which will slightly increase the erosion area. The erosion rate increases linearly with the increase of the sand particle concentration. With the increase of particle diameter, the erosion rate increases first and then tends to stable while the erosion area of the pressure surface gradually increases. (3) The available pressure drop of the chimney is reduced by about 40%, and the output power is reduced by 73.3% because of the crosswind. Windbreak walls can weaken the negative effects of crosswinds. (4) Sand deposition makes the heat-collecting performance of the collector deteriorate sharply. After 30 days of operation in the dusty environment, the output power of SCPP decreased by 12.9%, and the collector efficiency decreased from 34% to 23% due to sand deposition.