Wind Tunnel Experiments of Wind-Sand Environment for Different Width Subgrades

Abstract

Sand disasters significantly restrict ecological restoration and the development of sustainable transport infrastructure in desert areas, and the impact of varying subgrade width and roughness caused by different types and uses of routes on the wind-sand environment is still unclear. To address this, four typical subgrade widths were studied, and wind tunnel experiments were carried out using models. Near the ground surface (at heights < 8.3 cm), a 3.5 cm wide subgrade had a greater effect on the windward wind speed compared with three other widths. The distance required for wind speed recovery on the leeward of the 3.5 cm wide subgrade was greater than that for the three other widths. The 3.5 cm wide subgrade had a larger effect range and extent on the leeward wind flow field compared with the three other widths. The distance needed for the leeward wind flow field to recover at the 3.5 cm wide subgrade was also greater than that for the three other widths. The sand transport rates for the 14, 26, and 41 cm wide subgrades were similar and showed a consistent trend. However, the sand transport rate for the 3.5 cm wide subgrade was more variable and was lower than that for the three other widths at near-ground surface heights but higher at intermediate heights. Width has a minor effect on the wind-sand environment around the subgrades compared to roughness. The research findings provide insights into the relationship between the subgrade width, roughness, and wind–sand environment, offering guidance for mitigating sand disasters along transportation routes. It provides theoretical support for optimizing transportation infrastructure design, promoting green and low-carbon construction, and promoting ecological restoration around the routes.

1. Introduction

Windblown sand is a primary natural disaster threatening the smoothness of roads and the safety of traffic in sandy areas [1,2,3]. It also complicates the selection, survey, design, construction, operation, and maintenance of transportation routes in these areas [4,5,6]. As important symbols of modern transportation, railways and highways have been increasingly constructed in sandy regions due to the rapid development of transportation in these areas. For example, the cumulative length of railways in sandy areas is about 1.5 × 10⁴ km in China, making it the country with the longest railway distribution in sandy regions globally [7]. In terms of roads, Inner Mongolia alone has more than 10 roads crossing sandy regions, with a total length exceeding 3000 km. Research on wind-sand issues along the route is increasingly receiving attention [8]. The subgrade acts as an obstacle to the wind-sand two-phase flow, reducing the sand-carrying capacity of the airflow. This leads to wind erosion or sand accumulation on the subgrade slopes and surface. Many scholars have conducted extensive research on the wind-sand hazards of transportation routes using methods such as numerical simulation and wind tunnel experiments. The main research areas include the wind-sand environmental characteristics along the subgrade [9,10], improvements in subgrade guardrail types [11,12], evaluation of the sand prevention effectiveness of protective measures [13,14,15], and the formation mechanism of subgrade sand damage and its influencing factors [16,17,18]. Among these, research on the causes of subgrade sand damage mainly focuses on the interaction between the subgrade’s cross-sectional and longitudinal structure and the wind-sand flow. Detailed studies have been conducted on subgrade plane alignment parameters [19], characteristics of sand accumulation on curved subgrades [20], subgrade design forms [21,22,23], subgrade height [24], slope gradient [25], wind-road angle [26], and environmental characteristics along the route [27], with improvement proposals provided.

However, the width of the subgrade varies due to different types and grades of routes. For example, the width of a single-track railway subgrade is 3.5 m, while the width of an ordinary national/provincial highway with a design speed not exceeding 80 km·h⁻¹ is 7–7.5 m. The minimum width of a highway subgrade is 26 m. Despite this, no detailed research has been conducted on subgrade width, and its impact on the wind-sand flow field remains unclear. Additionally, the influence of the roughness of different subgrade types on the wind-sand flow field has not been explored. The lack of research in this area may lead to unreasonable designs, thereby increasing construction costs and wasting resources, which not only affects the economic benefits of the project but may also hinder the development of sustainable transportation infrastructure.

Therefore, four typical subgrade widths commonly found in China were selected: single-track railways, widened national/provincial roads, standard highways, and widened highways. Wind tunnel experiments were carried out using models to study the wind speed, flow field, and sand transportation rates of subgrades with different widths. The findings aim to provide a better understanding of the relationship between subgrade width and wind-sand environment (it refers to a geographical environment dominated by wind forces, with loose sandy materials on the surface, leading to processes like wind erosion, transportation, and deposition).

2. Wind Tunnel Experiment Setup

2.1. Models and Dimensions

According to the “Code for Design of Railway Earth Structure” (TB10001–2016, China) and the “Technical Standard of Highway Engineering” (JTG B01–2014, China), the top widths of the subgrades for various types roads are as follows: 3.5 m for single-track railways, 14 m for widened national/provincial roads (with two additional lanes compared with the standard width), 26 m for standard highways, and 41 m for widened highways (with four additional lanes compared with the standard width). The slope ratio for each is 1:1.75 (i.e., a slope of 30°). For the wind tunnel experiment, the subgrade model width was scaled at 1:100. The model top widths were 3.5, 14, 26, and 41 cm, with a 30° slope. The experiment took place at the Dunhuang Gobi and Desert Research Station of the Chinese Academy of Sciences. The wind tunnel used was a closed-circuit blow-type wind tunnel with a cross-section of 63 cm × 63 cm and an experimental section length of 600 cm. The boundary layer thickness was between 12 and 15 cm. Each subgrade model had a length of 62 cm and a height of 8 cm, resulting in a blockage ratio of 12.5%. To simulate different surface roughness, the 3.5 cm width subgrade model (for single-track railways) was covered with 36-grit sandpaper. The 14 cm, 26 cm, and 41 cm models (for widened notional/provincial roads, standard highways, and widened highways) were covered with 80-grit sandpaper (Figure 1).

2.2. Layout of Wind Speed Experiment

Eight measurement points were set up on each subgrade in the windward direction: −20H, −15H, −10H, −5H, −3H, −1H, −0.5H, and −0H. Here, H represents the model height. The “−” indicates the windward direction, while the “+” indicates the leeward direction. The point −0H represents the foot of the windward of the subgrade. On the leeward side, ten measurement points were established: 0H (the foot of the leeward slope), 0.5H, 1H, 3H, 5H, 10H, 15H, 20H, 25H, and 30H. For the subgrade models with widths of 3.5, 14, and 26 cm, five measurement points were set on the surface: the middle of the windward slope, the shoulder of the windward slope, the center of the subgrade top, the shoulder of the leeward slope, and the middle of the leeward slope. For the 41 cm width subgrade model, seven surface measurement points were used. An additional point was placed between the center of the subgrade top and the shoulders of the windward and leeward slopes (Figure 2). At each point, the wind speed was measured at 8 different heights using a Pitot tube. These heights were 0.6, 0.8, 2.1, 8.3, 12.2, 16.4, 20.2, and 24.2 cm. The Pitot tube was positioned at the center of the bottom plate of the wind tunnel. After the experimental wind speed stabilized, measurements were taken every 2 s, with the average values calculated from 30 consecutive measurements. The wind speeds setting for the experiment were 6, 9, 12, 15, and 18 m·s⁻¹. These settings were chosen to reflect typical wind conditions and sand movement in desert areas. The range exceeds the sand-moving wind speed and also covers the common sand transport wind speeds, considering the influence of strong winds. The wind velocity profiles of the desert surface were recreated in the wind tunnel to ensure the results accurately reflect the sand moving, transportation, and deposition processes under different wind conditions.

2.3. Layout of Sand Transport Rate Experiment

A sand bed was laid with a length of 370 cm, a width of 63 cm, and a thickness of 5 cm. It was placed at a distance of 10H in the windward direction of the subgrade. A sand collection device was positioned at a distance of 10H in the leeward direction. The device had sand collection ports at every 1 cm of height. Each port measured 2 cm in width and 1 cm in height, with a total of 50 heights. The experimental wind speeds were 6, 9, 12, 15, and 18 m·s⁻¹. The sand transport rate for each width of the subgrade was measured (Figure 3).

3. Results and Analysis

3.1. Wind Speed

Based on the layout of the wind tunnel experiment, the wind speeds at each measurement point on the subgrade surface and in the windward and leeward directions for each width were determined (Figure 4). A total of 23 observation points were set up upwind and downwind of the embankment and along the subgrade in the experiment (25 observation points for the 41 cm wide subgrade). Wind speeds were measured at eight different heights at each observation point. The distances were expressed in terms of embankment height as the unit of measurement. For example, a distance of 40 cm upwind from the embankment is denoted as −5H (H = 8 cm).

When the subgrade width was 3.5 cm, the wind speed was divided by a height of 8.3 cm. For wind speeds at heights ≤ 8.3 cm, the wind speed started to decrease from the windward direction at a −10H distance and reached the lowest value near the foot of the windward slope at −0H. Then, it substantially increased as the windward slope of the subgrade rose and reached the highest value at the shoulder of the windward slope. Afterward, it rapidly decreased, reached the lowest value (0 or close to 0) near the foot of the leeward slope, and then increased considerably again. By the time it reached 20H in the leeward direction, the wind speed had generally recovered. For wind speeds at heights > 8.3 cm, the wind speed started to decrease at the −5H distance from the windward direction and gradually increased. Then, it reached the highest value between the top of the subgrade and the leeward 1H distance and maintained a high value. Afterward, it began to decrease, and by the time it reached the leeward 15H distance, the wind speed had generally recovered. The variation in wind speed at heights > 8.3 cm was smaller than that at corresponding heights ≤ 8.3 cm.

For the subgrade widths of 14, 26, and 41 cm, the wind speed differences were small, and the variation trends were generally consistent. A subgrade height of 8 cm was taken as a boundary. When the wind speed measurement point was below the subgrade height, the near-ground wind speed (<8 cm) started to decrease from the windward −5H distance and reached the lowest value near the foot of the windward slope (−0H). Then, it considerably increased as the windward slope of the subgrade rose and reached the highest value at the shoulder of the windward slope. Afterward, it rapidly decreased, reached the lowest value near the foot of the leeward slope (0 or close to 0), and then increased again. By the time it reached the leeward 15H distance, the wind speed had generally recovered. When the wind speed measurement point was higher than the subgrade height, the wind speed (>8 cm) started to decrease at the −5H distance from the windward direction and gradually increased. Then, it reached the highest value between the top of the subgrade and the leeward slope and maintained a high value. Afterward, it started to decrease, and by the time it reached the leeward 15H distance, the wind speed had generally recovered. The variation in wind speed at heights above the subgrade (>8 cm) was smaller than that of the near-ground wind speed (<8 cm).

At heights above 8.3 cm, the wind speed differences for various subgrade widths were small, and the variation trends were generally consistent. At a height of 8.3 cm, the wind speed variation for the 3.5 cm wide subgrade was considerably larger than that for the three other subgrade widths. At near-ground heights (<8.3 cm), the 3.5 cm wide subgrade had a greater influence on wind speed in the windward direction compared with the three other subgrade widths. Additionally, the distance required for wind speed to recover in the leeward direction was greater for the 3.5 cm wide subgrade than that for the three other widths.

3.2. Wind Flow Field

Based on the results of the wind speed experiment, contour maps of wind speed on the subgrade surface at various widths and in the windward and leeward directions were obtained using the Kriging interpolation method (Figure 5).

The impact of different subgrade widths on the wind flow field shows significant differences. The 3.5 cm wide subgrade, with its larger surface roughness, exhibits unique flow-field characteristics. This subgrade creates a distinct wind speed acceleration zone at the top and generates a wind speed reduction zone within the upwind −5H range. The wind speed reduction range (refers to the spatial area where the subgrade significantly disturbs the surrounding wind-sand flow field) and intensity on the leeward slope to the downwind 20H range are significantly greater than the 15H range of the other three subgrade widths (14, 26, 41 cm). In contrast, the 14, 26, and 41 cm subgrades, due to their similar surface roughness, exhibit highly consistent wind flow-field changes. They all form a wind speed acceleration zone at the top, with wind speed reduction zones in the upwind −5H and leeward slope 15H ranges, and the reduction in the leeward direction is more pronounced. It is noteworthy that the impact of all subgrades in the leeward direction is more significant than in the upwind direction. However, the 3.5 cm subgrade, due to its larger roughness, results in a wider disturbance range and a longer recovery distance (it is defined as the minimum downwind horizontal distance required for the cross-sectional mean wind speed to approach the free-flow wind speed before the disturbance after the airflow is disturbed by the subgrade structure) in the leeward flow field. This phenomenon indicates that, during the variation in subgrade width, surface roughness is a key factor affecting flow-field characteristics. When roughness is the same, the width variation has a relatively limited impact on the wind flow field. The 3.5 cm subgrade, due to its special roughness characteristics, has a more significant effect on airflow separation and flow-field recovery distance.

3.3. Sand Transport Rate

Based on the layout of the wind tunnel experiment, the sand transport rate experiment results for various subgrade widths were documented (Figure 6). The figures show the unit-width sand transport rate at different heights. Taking a 3.5 cm width subgrade as an example, at a wind speed of 6 m·s⁻¹, the highest unit-width sand transport rate is 0.3325 g·cm⁻²·min⁻¹ at a height of 19 cm.

When the experimental wind speed was 6 m·s⁻¹, the total sand transport rate for the 3.5 cm wide subgrade was the lowest. The sand transport rate showed an unstable state of first decreasing, then increasing, and later decreasing again with height. For the three other subgrade widths, the total sand transport rates were similar and had a generally decreasing trend as the height increased. The sand transport rate for the 3.5 cm wide subgrade was substantially lower than that for the three other subgrade widths in the near-ground height range of 0–15 cm.

When the experimental wind speed was 9 m·s⁻¹, the sand transport rate for all subgrade widths followed a similar trend of first decreasing, then increasing, and later decreasing with height. However, the sand transport rate for the 3.5 cm wide subgrade width was considerably lower than that for the three other subgrade widths at heights of 0–20 cm near the ground and substantially higher at heights of 21–35 cm.

When the experimental wind speed was 12 m·s⁻¹, the sand transport rate for all subgrade widths followed a similar trend of first decreasing, then increasing, and later decreasing with height. The sand transport rate for the 3.5 cm wide subgrade width was substantially lower than that for the three other subgrade widths at heights of 0–22 cm and considerably higher at heights of 24–41 cm.

When the experimental wind speed was 15 m·s⁻¹, the sand transport rate for all subgrade widths followed a similar trend of generally increasing and then decreasing with height. The sand transport rate for the 3.5 cm wide subgrade width was higher than that for the three other subgrade widths at heights of 20–42 cm.

When the experimental wind speed was 18 m·s⁻¹, the sand transport rate for all subgrade widths followed a similar trend of generally increasing and then decreasing with height. The sand transport rate for the 3.5 cm wide subgrade width was substantially higher than that for the three other subgrade widths at heights of 24–46 cm.

4. Discussion

In the wind tunnel experiment, a 1:100 scale was used (model height 8 cm corresponds to prototype height 8 m, and length 0.6 m corresponds to 60 m). The main reason for this scale is to ensure dynamic similarity between the model and prototype in key aerodynamic parameters, such as Reynolds number and friction velocity. For sediment transport, the sand moving wind speed ratio and dimensionless sediment transport rate must match between the model and prototype. However, since sand particles cannot be scaled down exactly, the near-surface motion is distorted. The wind tunnel sidewall constraints and idealized terrain do not match the complexity of real-world environments. The finite wind tunnel dimensions (6 × 0.6 × 0.6 m) may cause sidewall interference, affecting the three-dimensional characteristics of the flow field. Additionally, scale effects on near-surface turbulence may lead to deviations in the sand particle jumping trajectory.

The boundary layer thickness (12–15 cm), which is close to the model height (8 cm), can distort the velocity profile, pressure distribution, and turbulence structure. Due to experimental constraints, no corrective measures were implemented. This limitation may cause differences in flow characteristics at the top of the model compared to the prototype. These differences especially affect the development of the lee-side separation vortex and the sediment accumulation distribution. Quantitative data should be treated as relative comparisons rather than absolute values. Multiple sets of repeated experiments have verified the robustness of the conclusions under different embankment width conditions.

Based on the experiments and analyses, no remarkable differences were observed in the wind speed, flow field, or sand transport rates between the subgrades with widths of 14, 26, and 41 cm. Their trends were also similar, which indicated that width did not have a substantial effect on the wind-sand environment. However, the wind-sand environment of the 3.5 cm wide subgrade showed noticeable differences compared with the three other widths, mainly due to the dissimilarity in the surface roughness of the subgrades. According to the experiment, the 14, 26, and 41 cm widths of subgrades represent highways, with 80-grit sandpaper attached to the subgrade surface, whereas the 3.5 cm width subgrade represents a railway, with 36-grit sandpaper attached to simulate the different surface roughness of highway and railway subgrades. Using a portable surface roughness meter, the roughness of the 80-grit sandpaper was approximately 87.95 μm, whereas that of the 36-grit sandpaper was approximately 195.4 μm (before using the portable surface roughness meter, a standard reference block was used for calibration, and the measurement results were within 10% of the calibration value, which meant no recalibration was required. The measurement parameters used were Ra, with a sampling length of 2.5 mm. The sensor was moved within the measurement area for the same brand of 1000 and 2000-grit sandpapers, and 10 measurements were taken for each sample to obtain the average value. The average surface roughness values were 7.11 and 3.48 μm. Grit number refers to the number of holes per square inch, and a higher grit number indicates finer material. Based on the relationship between grit number and roughness, the roughness values of 80-grit sandpaper were 88.9 and 87.0 μm, and those of 36-grit sandpaper were 197.5 and 193.3 μm. The difference between the two sets of roughness values for the same grit number is small. Therefore, the average value was taken to improve the reliability of the results. The roughness of 80-grit sandpaper was approximately 87.95 μm, and that of 36-grit sandpaper was approximately 195.4 μm. Due to the greater surface roughness of the 3.5 cm wide subgrade compared with that of the three other widths, the frictional effect of the wind near the surface (≤8.3 cm height) was enhanced, and the resistance coefficient increased. This outcome led to a greater effect on the wind speed near the ground on both sides of the subgrade. The distance over which the wind speed decreased over the subgrade was greater than that for the three other widths, and the distance required for wind speed recovery leeward was also greater. The increase in roughness reorganizes the wind flow field on both sides of the subgrade, while changing the near-surface friction and wind speed. As a result, the 3.5 cm wide subgrade affects the leeward flow field to a greater extent and degree than the other three widths, and the distance required to restore the leeward flow field is greater. Additionally, the change in the sand transport rate for the 3.5 cm wide subgrade was more dramatic, and the maximum sand transport rate occurred at a certain distance above the ground at a certain height, rather than at the surface. The wind-sand flow structure did not follow an exponential decrease with height, which is similar to the structure of wind-sand flow observed in the Gobi Desert in the field [26,27]. Moreover, increased roughness caused sand particles to bounce higher, fully utilize the air flow energy [28,29], and resulted in more drastic changes in the sand transport rates for the 3.5 cm wide subgrade. At lower heights, the sand transport rates were lower than the other three widths, whereas at intermediate heights, they were higher. This result is also consistent with related studies in the Gobi Desert, where sand particle jump heights reached 2 m [30], and wind-sand flows jumped up to 9 m. The maximum rate of sand transport from the Gobi Desert surface occurs at a height of 2–6 cm [31,32,33,34].

In wind tunnel experiments, 80-grit sandpaper simulates highway subgrades. Its roughness is consistent with the wear state of actual asphalt pavements, allowing it to reflect the aerodynamic characteristics of the road surfaces. In contrast, 36-grit sandpaper is used to simulate railway subgrades. Although its microscopic roughness is lower than the actual ballast, it approximates the aerodynamic roughness of railway ballast under a 1:100 scale model. This ensures the dynamic similarity of the wind-sand flow field. The role of roughness in wind-sand transport is significant in several ways. At the fluid separation scale, the roughness elements produced by the 36-grit sandpaper cause continuous flow separation, while the width variation only affects localized edge separation. Roughness also directly controls the sand transport rates by affecting the friction velocity. As a result, increasing the subgrade width has little effect on the overall flow field. However, subgrade roughness directly affects the near-surface flow. By increasing surface friction, it promotes particle movement and alters the turbulent structure, influencing sand transport.

It provides clear guidance for optimizing highway wind-sand control engineering. Rather than adjusting the subgrade width, the priority should be optimizing surface roughness. Engineering applications should focus on the design of the roughness element parameters to disrupt near-surface wind-sand flow effectively. To reduce wind speed fluctuations and sand transport rate variations near the surface, the optimization of surface roughness around the subgrade(such as using gravel cover, grid sand fixation, etc.) should be prioritized. For narrow subgrades (such as 3.5 cm), the protective range should be extended (e.g., adding front sand-blocking barriers or lengthening the windward slope of the subgrade) to alleviate the problem of wind flow recovery in the downwind direction.

5. Conclusions

Based on the experimental conditions, the following preliminary conclusions can be drawn:

At heights above 8.3 cm, the wind speed differences between subgrades of different widths were small, and the variation trends were the same. At the height of 8.3 cm, the wind speed variation for the 3.5 cm wide subgrade was substantially larger than that of the three other widths. Near the ground (<8.3 cm), the 3.5 cm wide subgrade affected the windward wind speed over a greater distance than the three other widths, and the recovery distance for leeward wind speed was greater than that for the three other widths. The 3.5 cm wide subgrade had a greater influence range and degree on the leeward flow field, and the recovery distance for the leeward flow field was also greater.

The sand transport rates for subgrades with widths of 14, 26, and 41 cm showed minor differences, and the variation trends were the same, whereas the sand transport rate for the 3.5 cm wide subgrade varied more dramatically. The sand transport rate was lower at lower heights and higher at intermediate heights compared with the three other widths.

Width has a minor effect on the wind-sand environment around the subgrades compared to roughness.