Numerical Simulations and Wind Tunnel Experiments to Optimize the Parameters of the Second Sand Fence and Prevent Sand Accumulation on the Subgrade of a Desert Railway

Abstract

Wind-blown sand significantly affects the construction and safe operation of railways in desert regions. The performance of a wind-blown-sand prevention system with different structural parameters and sand accumulation around the railway subgrade was analyzed in this study. The optimum porosity and opening type of a second sand fence were assessed via wind tunnel experiments and numerical simulations. The results showed that the subgrade intercepted some sand and reduced sand accumulation on the track surface, and the interception rate was 29.70%. The wind-blown-sand prevention efficiency of the subgrade was 88.55%. Moreover, the lower the porosity of the second sand fence, the lower the sand velocity on the windward side and the higher the sand accumulation. The porosities of the first and second sand fences should be 30% and 20%, respectively, to maximize the sand accumulation between the fences. When the second sand fence had horizontal openings, most of the sand accumulated near the surface (within 20 cm) on the leeward side and on the straw checkerboard barrier, and the maximum wind-blown-sand prevention efficiency was 97.16%. When the second sand fence had vertical openings, the efficiency was 93.60%, and the sand accumulation on the leeward side and the straw checkerboard barrier was reduced. As the fence height increased (above 20 cm), the sand prevention efficiency of both approaches increased. The research results can provide guidance for the formulation and optimization of sand prevention measures for railways and highways in deserts.

1. Introduction

Wind-blown sand has a profound impact on the natural environment and its evolution in arid and semi-arid areas [1,2,3]. Anthropogenic activities have exacerbated desertification, thus adversely affecting agriculture, animal husbandry, and human life and health [4,5,6]. Wind-blown sand also significantly affects the safety of railway operations in desert regions (Figure 1). Many researchers have investigated the movement, transportation, and deposition mechanisms of sand particles using field surveys, wind tunnel experiments, numerical simulations, and theoretical analyses, and have proposed wind-blown-sand prevention measures to mitigate the impact on railway operations [7,8,9,10,11].

Reed sand fences and straw checkerboard barriers are excellent wind-blown-sand prevention measures due to their low material cost, convenient construction, high durability, and environmental protection characteristics [12,13]. They have been widely used to prevent the effects of wind-blown sand on railways and highways [14,15]. A reed sand fence is generally located on the windward side at a distance of 150–200 m from the railway line. Its main purpose is to intercept wind-blown sand and mitigate its effect on the straw checkerboard barrier, the sand-collection belt, the shelter forest belt, and the railway line so that sand does not bury the grass. The sand accumulation on the leeward side and between the two sand fences can be controlled by choosing an appropriate porosity and opening type of the fence to prolong its service life. The straw checkerboard barrier between the two fences collects the sand grains not intercepted by the reed sand fence. The railway subgrade in desert areas is elevated to ensure that it is not buried by sand; it intercepts sand particles in combination with the wind-blown-sand prevention system.

A single reed fence does not sufficiently block wind-blown sand [16,17,18]. Thus, two reed sand fences and a straw checkerboard barrier are typically used for wind-blown sand prevention. An appropriate porosity and opening type of the front and back sand fences are required to intercept wind-blown sand and collect sand in the straw checkerboard barrier. For example, the double-row sand fences along the Tao–Erdos and Golmud–Korla railways (Figure 2) have the same porosity for both fences, and the spacing is about 10H. An incorrect porosity and opening type of the two fences has led to the burial of the front fences in some areas; the back fences did not provide sufficient protection from wind-blown sand, and the straw checkerboard barrier did not collect a sufficient amount of sand. Therefore, the determination of the appropriate porosity and spacing of the fences is crucial to mitigate the effect of wind-blown sand. Current research mainly focuses on the structural parameters of the first sand fence, including its height, porosity, spacing, location, and material [1,7,9,12,15,16,17]. However, there have been few studies on the design parameters of the second sand fence and the sand accumulation around the railway subgrade.

Research methods for the evaluation of the sand prevention efficiency of sand fences include field tests, wind tunnel experiments, and numerical simulations [19,20,21]. Field tests are carried out to obtain measurements reflecting the effectiveness of the sand fence. However, the data are intermittent due to the constraints of the observation instruments, weather changes, the terrain, the geomorphology, extreme wind speeds, and other uncertain factors [22]. Thus, it is difficult to conduct a quantitative analysis. Therefore, indoor wind tunnel experiments are commonly used to study the wind-blown-sand prevention efficiency of sand fences. The inlet wind speed, the amount of wind-transported sand, the flow field distribution, and the amount of sand around the sand fence can be measured by pitot tubes and sand collection instruments. Numerical simulations have been widely used to determine the mechanism of wind-blown sand and the prevention of its effects [23].

In this study, wind tunnel experiments and numerical simulations were carried out to assess the efficiency of a sand prevention system of a railway subgrade. The appropriate porosity and opening type of a second sand fence are investigated, and the sand transport amount and sand prevention efficiency determined from the numerical simulations and experiments are compared. The results provide new insights into the development of sand prevention measures for similar projects.

2. Materials and Methods

2.1. Wind Tunnel Experiment

The wind tunnel experiment was carried out in a blow-down wind tunnel at the National Key Laboratory of Desertification and Sand Disaster Prevention in Gansu Province, China. The wind tunnel is 38.9 m long, the test section is 16 m long, the cross-section is 1.2 m × 1.2 m, and the wind speed is adjustable (4–35 m/s). It consists of a power section, a rectification section, an experimental section, and a diffusion section. The front end of the experimental section is equipped with a wedge with a rough surface to obtain a logarithmic wind speed profile under neutral atmospheric conditions.

The layout diagram of the wind tunnel experiment is shown in Figure 3. The model of the railway subgrade consisted of a double track with a track ballast, and the wind-blown-sand prevention system on the windward side of the subgrade included two sand fences and a straw checkerboard barrier. The first sand fence had vertical openings and 30% porosity. The second sand fence had porosities of 20%, 30%, and 40% and horizontal or vertical openings to determine the optimal porosity and opening type. The height of the subgrade model was 25 cm, the height of the sand fence was 15 cm, and the height of the straw checkerboard barrier was 3 cm. The pitot tube was attached to a frame to monitor the inlet wind speed. Three portable sand collection instruments were, respectively, attached to the front and back of the sand fence and the subgrade to measure the amount of sand.

At the beginning of the experimental process, wind field and sand accumulation experiments were conducted to verify the reliability of the numerical simulation results. Then, a sand accumulation experiment was conducted on the single subgrade to clarify the sand protection capacity of the subgrade. The different porosities and opening types of the second subgrade were tested. The inlet wind speeds were 8, 10, and 12 m/s. Before the experiment, the pitot tube was used to measure the wind speed at different heights. The wind speed profile was obtained by fitting the wind-speed point values, which conformed to the logarithmic law, as shown in Figure 4a. A sand source (length 3 m, thickness 0.5 m, and width 1.2 m) was located behind the rough element. The particle size distribution is presented in Figure 4b. After blowing sand for about 5 min, the amount of transported sand without a wind-blown-sand prevention model was measured with a sand collector. Different wind-blown-sand prevention models were placed 10 m from the sand source. The sand accumulation around the subgrade and the sand fence was measured after blowing sand for 5 min. The efficiency of the sand prevention system of the subgrade and the optimal porosity and opening type of the second sand fence were determined.

2.2. Numerical Simulation

A three-dimensional geometric model was established using AutoCAD (version: 2019), the mesh was created in Design Modeler (version: 16.0), and the numerical simulation was carried out using Fluent (version: 16.0). The computational domain was a cuboid area with a length of 150 m, a height of 15 m, and a width of 15 m. The height of the subgrade was 10 m, the height of the sand barrier was 1.5 m, the porosity of the first sand fence was 30%, the porosities of the second sand barrier were 20%, 30%, and 40%, and the opening types were horizontal and vertical. The dimension of the straw checkerboard barrier was 1 m × 1 m × 0.2 m, and the porosity was 40%.

A hexahedral mesh was used for grid division, and the dimensions were 0.28 m, 0.03 m, and 0.02 m in the horizontal x and y directions and in the vertical z direction, respectively. The grid near the surface area and the fence model were locally encrypted. The minimum mesh size was 3 mm. The mesh was checked to confirm its quality.

The boundary condition was velocity at the inlet, outflow at the outlet, and bottom, upper, left, and right boundaries at the walls. A sliding boundary condition was adopted for the front and rear walls of the sand fence. The type was WALL, the roughness was Ks, and the roughness of the other walls was 0 by default. The transport type of the medium sand particles was continuous, and the type was FLUID. The domains and boundary conditions are shown in Figure 5.

The time step of the numerical simulation was 0.01 s, the number of time steps was 60,000, and the discrete format was QUICK (this format can reduce the false diffusion error and has high accuracy and stability). The standard of computational convergence is physical convergence, i.e., the change trend of each component value reaches stability, and the value distribution size is on the order of 10⁻⁵.

The finite volume method was used to discretize the governing equations, and the semi-implicit method for pressure-linked equations (SIMPLE) was used to correct the pressure–velocity coupling effect. An unsteady wind field was obtained. The governing equations of the gas phase and sand phase were obtained from a previous study [18]. The sand particle size was 0.00015 m, the sand density was 2650 kg·m⁻³, the initial sand phase volume fraction was about 1%, the air density ρ was 1.225 kg·m⁻³, and the viscosity was μ = 1.789 × 10⁻⁵ Pa·s [18].

The inlet wind speed profile was calculated as follows [24]:

$$v(h)=\frac{v_{\ast }}{k}\ln \frac{h}{z_0}$$

(1)

where v_* is the friction wind speed, k is the von Kármán coefficient, which is 0.4, z₀ is the length of the rough section, h is the height, and v(h) is the wind speed at height h. The inlet wind speeds were, respectively, 8, 10, and 12 m/s at the height of 10 m, and the surface roughness z₀ was 0.0013 m.

2.3. Verification of the Numerical Simulation Results

A computational domain with the same size as the wind tunnel was established. The inlet wind speed was 10 m/s, the wind speed profile in Figure 4a was adopted, and the relevant parameters of sand particles were consistent with those in the numerical simulation reported in Section 2.2. Moreover, the data of the wind speed and sand transport at the central point of the computational domain were compared with the wind tunnel testing data in Figure 6a,b. In the wind tunnel experiment, the acquisition time of the wind speed data was 3 min and repeated three times, and the calculation error was about 2.0%. The collection time of sand accumulation data was 5 min, data were collected three times, and the error was about 3.0%.

Figure 6 shows that the error of the wind speed between the numerical simulation and wind tunnel experiment was about 5.5%, and the maximum error of the proportion of the sand volume was about 5.7%. This demonstrates that the numerical simulation results of the sand transport and wind speed on a flat surface are consistent with the wind tunnel results, indicating that the numerical simulation model and parameter settings are reasonable.

2.4. Evaluation Indicators

The evaluation of the efficiency of the wind-blown-sand prevention system considers the sand transport at the entrance and the center point of the subgrade surface [25,26]. The relative sand transport rate is defined as follows:

$$M=\frac{m_{0\left(z,x\right)}-m_{z\left(z,x\right)}}{m_{0\left(z,x\right)}}\times 100,$$

(2)

where m_z(z,x) is the sand transport rate in the x direction at the height of z at the center point of the subgrade surface, m_0(z,x) is the sand transport rate in the x direction at the height of z at the entrance, and M is the relative sand transport rate at different heights.

3. Results

3.1. Efficiency of the Subgrade and Sand Prevention System in Mitigating Sand Accumulation

A wind-blown-sand prevention system is an effective measure to protect the subgrade from being buried and eroded. However, it is impossible to intercept all sand particles on the windward side; thus, some sand is deposited around the subgrade above the height of the sand fence. The subgrade is an elevated trapezoidal structure that causes sand to accumulate at its base. Therefore, it is critical to clarify the role of the subgrade to evaluate the advantages and disadvantages of the wind-blown-sand prevention system.

Figure 7 shows the sand distribution around the single subgrade and the sand prevention system at a wind speed of 12 m/s and a flow direction in the x direction. Less sand was deposited on the windward side of the subgrade without a sand prevention system, and most of the sand was deposited on the leeward side of the subgrade. A large amount of sand was deposited on the subgrade surface and track line. However, little sand accumulated on the windward side and track surface of the subgrade with the sand prevention system. Thus, the wind-blown-sand prevention system intercepted many sand particles, thereby reducing the amount of sand reaching the line and track. The sand accumulation on the leeward side of the subgrade occurred due to the swirling sand particles formed by turbulence.

The sand transport rate and sand prevention efficiency at different heights at the center of the subgrade surface are depicted in Figure 8. Figure 8a shows that the sand transport rate at the center of the subgrade decreased with an increase in height. The sand transport rate was lower when the sand prevention system was installed and was less than that of the single subgrade. This indicates that the windward slope of the subgrade intercepted some of the sand particles, thus reducing sand accumulation on the subgrade. Figure 8b shows that the sand prevention efficiency on the subgrade surface was significantly higher after the sand prevention system was installed (88.55%), indicating that sand accumulation near the subgrade was reduced. The sand prevention system intercepted most of the sand, causing it to accumulate between the two sand fences and the straw checkerboard barrier.

3.2. Optimal Porosity of the Second Sand Fence

The double-row fence is located at the front end of the windward side of the sand prevention system. It intercepts the sand particles and reduces the intensity of the wind-blown sand flow. The sand accumulation between the two fences must be monitored in consideration of the service life of the fence. The height, porosity, and opening type of a sand fence significantly affect its performance. Many studies have investigated the structural parameters of the first sand fence in the sand prevention system; the porosity should be 0.3 for a fence with a height of 1.5 m. However, few studies have analyzed the porosity and opening type of the second sand fence.

The porosity depends on the location of the sand fence and the protection object. A sand fence with low porosity accumulates more sand particles on the windward side. A fence with high porosity accumulates less sand on the windward side and more on the leeward side. The sand accumulation amount is larger, and the service life of the fence is longer. The second sand fence blocks incoming sand particles, protects sand catchment measures on the leeward side, and prevents the subgrade from being buried. It is necessary to intercept the incoming sand particles between the two reed fences (Q1) and between the second fence and the sand catchment measures (Q2) to extend the service life of the fence (Figure 9).

Figure 10 presents the sand accumulation for porosities of 20%, 30%, and 40% of the second sand fence at an incoming wind speed of 12 m/s. The openings were vertical, and the flow direction was x. The least amount of sand was accumulated on the windward side of the sand fence when the porosity was 40%. Most of the sand accumulated on the leeward side of the sand fence. Wind erosion occurred on both sides of the fence when the porosity was 30%, and there was a negligible difference in sand accumulation on the windward and leeward sides. When the porosity was 20%, the sand accumulation on the windward side of the sand fence was higher than for the 30% and 40% porosities. Numerous sand particles were deposited in the Q1 area and near the straw checkerboard barrier on the leeward side, indicating that the sand accumulation at Q1 and Q2 must be monitored. As the sand accumulation continued, the fence was buried.

If the porosity of the first reed fence is too low, the sand will accumulate in front of the sand fence and will bury it. The fence cannot protect the railway line for very long, resulting in a waste of resources and high construction costs. The porosity of the second reed fence should be relatively low so that the sand particles are contained between the two fences. The results indicate that the wind-blown-sand prevention efficiency is optimum when the porosity of the second fence is 20%.

3.3. Optimal Opening Types of the Second Sand Fence

The amount of sand accumulation around the second fence was found to differ for different opening types. Figure 11 shows the sand accumulation at the second sand fence for vertical and horizontal openings and 20% porosity. The incoming wind speed was 12 m/s, and the flow direction was x. When the second sand fence had vertical openings, most of the sand particles were deposited in the Q1 and Q2 areas, and the amount of accumulated sand was very low on the windward slope of the subgrade. The reason for this is that the first and second sand fences had vertical openings. Thus, sequential air flow occurred between the two sand fences and the airflow velocity increased, thus removing some of the sand and depositing it around the subgrade.

However, the sequential air flow was disturbed when the second sand fence had horizontal openings, and the sand accumulated around the second sand fence. The sand accumulation was higher in the straw checkerboard barrier, and the baffle near the surface intercepted the moving sand particles and deposited them around it. The diffused air at the top of the sand fence caused many sand particles to move upward and be deposited on the straw checkerboard barrier.

3.4. Wind Tunnel Experimental Results

The numerical analysis clarifies the wind-blown-sand prevention efficiencies of the single subgrade and the second sand fence due to different structural parameters. Wind tunnel tests were conducted on the single subgrade and the second sand fence with different porosities and opening types to assess the sand prevention efficiencies and verify the numerical results.

Figure 12 exhibits the sand transport and sand prevention efficiency at different heights at the center point of the single subgrade surface and the sand prevention system. Figure 12a shows that the amount of sand at the subgrade surface increased and then decreased with the height. The amount of sand was higher for the single subgrade than for the sand prevention system because the sand prevention system intercepted a large number of sand particles. The maximum difference in the sand accumulation between the two was 10 g·cm⁻²·s⁻¹ (at a height of 30 cm). The sand accumulation was significantly lower when the subgrade surface was higher than 20 cm and was nearly the same for the subgrade and the sand prevention system at a height of 50 cm. The reason for this is that there were fewer saltation sand particles, and the windward-side sand prevention system only intercepted near-surface sand particles. Figure 12b shows that the maximum sand prevention efficiency of the single subgrade was 32% (at a height of 25 cm), and that of the sand prevention system was 87%, which is consistent with the numerical simulation results.

Figure 13 presents the sand accumulation of the single subgrade and the sand prevention system. The single subgrade was found to have more sand accumulation. Most of the sand was deposited on the windward slope, reflecting the mitigation of sand accumulation by the subgrade. However, the sand accumulation on the subgrade surface was substantially reduced after installing the wind-blown-sand prevention system.

Figure 14 shows the sand prevention efficiency for different porosities of the second sand fence. The sand accumulation between the two sand fences decreased with an increase in porosity. The maximum efficiencies were 93.60%, 87.24%, and 79.52% for porosities of 20%, 30%, and 40%, respectively, thus showing a decreasing trend. As the porosity of the second sand fence increased, the number of sand particles crossing the sand fence increased, and the number of sand particles reaching the subgrade and the sand prevention efficiency decreased.

The sand accumulation between the two sand fences was the largest when the porosity of the second fence was 20%, thus reaching the maximum sand prevention efficiency (Figure 15). As the porosity of the second sand fence increased, the sand accumulation decreased, and the sand prevention effect at the subgrade was weakened.

The wind-blown-sand prevention efficiency for different opening types of the second sand fence and a porosity of 20% is exhibited in Figure 16. Blue represents the vertical opening of the second sand fence, and red represents the horizontal opening. When the porosities of both fences were 20%, the maximum sand prevention efficiency of the fence with horizontal openings was 97.16%, and that of the fence with vertical openings was 93.60%, a difference of about 3.56%. The sand prevention efficiency near the surface (below 20 cm) was higher for the fence with horizontal openings than for the fence with vertical openings. The trends in the sand prevention efficiency of the fences with both openings became similar as the fence height increased. This result shows that the opening type has a more pronounced effect on the near-ground sand transport and a less pronounced effect on the saltation of sand particles.

The sand accumulation for different opening types of the second sand fence and a porosity of 20% is shown in Figure 17. The trends in the sand accumulation of the fences with both openings were similar. On the whole, the sand accumulation for the fence with horizontal openings was greater than that for the fence with vertical openings.

4. Conclusions

In this study, the sand prevention ability of a railway roadbed and sand prevention system was analyzed, and the performance of the second sand fence with different porosities and opening types was studied by numerical simulation and wind tunnel experiments. The following conclusions were drawn.

(1)

The subgrade was found to accumulate sand on the windward side, thus reducing the amount of sand accumulation on the subgrade. The interception rate of the subgrade was 29.70%. Most of the sand was intercepted after the sand prevention system was installed, which increased the wind-blown-sand prevention efficiency to 88.55%.

(2)

The lower the porosity of the second sand fence, the lower the sand velocity on the windward side and the larger the sand accumulation. When the porosity exceeded 30%, the sand accumulation on the windward side decreased, and the sand accumulation on the leeward side and around the subgrade, as well as the risk of burying the subgrade, increased.

(3)

The second sand fence disrupted the sequential flow of particles coming from the first sand fence when the second sand fence had horizontal openings. Most sand accumulated on the leeward side and on the straw checkerboard barrier. The maximum wind-blown-sand prevention efficiency was 97.16% for the fence with horizontal openings and was 93.60% for the fence with vertical openings. As the height of the fence increased (above 20 cm), the sand prevention efficiencies of both approaches increased.