Ki Zhng , Peiwen Zho , Jinchng Zho , Xingxin Zhng

a College of Civil Engineering, Lanzhou Jiaotong University, Lanzhou, 730000, China

b School of Economics and Management, Lanzhou Jiaotong University, Lanzhou, 730000, China

Keywords:HDPE board sand fences Windproof efficiency Sand control efficiency Reasonable spacing Porosity

ABSTRACT Recently,sand fences,made of high density polyethylene(HDPE) board,which is resistant against harsh plateau environmental conditions, could possibly be used to control the wind-blown sand along the Qinghai section of the Golmud-Korla Railway. In this paper, wind tunnel simulation was performed on the protective effect of multiple rows of HDPE board sand fences on wind-blown sand. Two parameters have been used to reflect the protective effect of sand fences, that is, windproof efficiency and sand control efficiency,referring to the reduction rates of wind speed at a height of 0.04 m and sand flux in a run of test with the fence compared those in the control run of test without it, respectively. The test results showed that when the porosity was 30%, the single-row sand fence had the best wind proof and sand control efficiency with respective values of 46%and 65.3%.The windproof efficiencies of double-row and triple-row sand fences at 20H were 59% and 62%, respectively. Compared with the windproof efficiency of the single-row sand fence, those of the two-row and three-row fences increased by 13% and 16%, respectively; The maximum sand control efficiency of the double-row and triple-row sand fences were 92.1% and 96.5%, increased by 26.8% and 31.2%, respectively, compared with that of the single-row sand fence. This study indicates that the sand control efficiency after adding one or two rows was not double or triple that of the single-row fence, while the first-row sand fence played a major protective role. Compared with the double-row sand fences, the effect of the third row sand fence in triple-row sand fences is limited, and thus double-row HDPE board sand fences with the porosity of 30% are recommended to control wind-blown sand along the Qinghai section of the Golmud-Korla Railway.

1. Introduction

Sand fences play an important role in the mechanical sand control system, and are usually installed on the windward side of railways, highways, and other ground transportation links to protect the railways from disruption of operation and from obstruction by wind-blown sand by reducing the wind speed and blocking the drifting sand(Wang,Qu,Ling,Xie,&Xiao,2017).Porosity is the key structural feature of sand fences(Cornelis&Gabriels,2005;Heisler& Dewalle, 1988). Sand fences reduce the wind speed within the area in a certain distance downwind of them, and therefore, wind speed reduction is closely related to porosity. The lower the porosity, the greater the reduction in wind speed (Dong et al.,2007). However, a sand fence with low porosity may generate a large eddy current, which enables the wind speed to recover quickly,thereby weakening the shielding effect of fences.Thus,for a sand fence with a given geometric design, height, length, thickness, opening size, and geometrical shape, there is an optimal porosity that will enable it to reduce the wind speed at most and provide the best shielding effect so that the objectives of wind speed reduction and shielding distance are balanced. Since the middle of the last century, scholars have conducted extensive research on a range of sand fence porosities, mainly using field monitoring (Bofah & Al-Hinai,1986; Boldes, Colman, & Leo, 2001;Hagen & Skidmore,1971a, Hagen & Skidmore,1971b; Miller et al.,1975; Schwartz, Fryrear, Harris, Bilbro, & Juo,1995; Wilson,1997),indoor wind tunnel simulation (Guan et al., 2003; Iversen, 1981;Judd et al.,1996;Lee et al.,2002;Park&Lee,2003;Wu et al,2013,Wu, Zou, Zhou, Zhang, & Shi, 2015; Yaragal et al., 1997), and numerical simulation by computer (Alhajraf, 2004; Fang & Wang,1997; Packwood, 2000; Pattone, Shaw, Judd, & Raupach, 1998;Vigiak, Sterk, Warren, & Hagen, 2003). The results show that the optimal range for sand fence porosity is 30-50%.As discovered by Hagen&Skidmore,1971a,the minimum wind speed was generated within the shielding distance downwind of the fence when the fence porosity was 40%. In field tests carried out by Lin, Jin, Zou,Cong, & Wen, 1984 on the southeastern edge of the Tengger Desert,sand fences with a porosity of 30-40%achieved the greatest effectiveness in preventing the accumulation of sand grains on the windward side; Lee and Kim (1998,1999) found that sand fences with a porosity of 30-40%were the most effective in reducing the speed and the surface pressure of the airflow downwind of the fence.Dong et al.(2007)believed that the optimal porosity was 20%or 30% when there was a significant difference in the dynamic characteristics of the airflow,based on the results of their study of the average velocity field of the fence using particle image velocimetry in an indoor wind tunnel test.

In addition, in these studies, most scholars focused mainly on single-row sand fences, because the study area was usually in the sandy desert where wind power is relatively low, which actually cannot meet the demand for sand control in areas with strong wind. Relatively less research has been conducted on the airflow over multiple barriers (Bitog et al., 2009; Gillies et al., 2017). Wu et al. (2013) compared the protective effect of wind barriers with different row numbers and spacing.Results showed that multi-row wind barriers provided better shelter than one-row barrier, and 5-7 H (H is the height of the fence) was suggested as the optimal row spacing for multi-row wind barriers. Besides, three-row wind barriers provided better shelter than two-row barriers,while tworow wind barriers could work well where disasters caused by blown sand did not occur frequently or intensely. Woodruff et al.(1963) concluded that two- and three-row wind barriers were most effective in reducing wind velocity, which was confirmed by later studies(e.g.,Cornelis&Gabriels,2005).Thus,it may be fruitful to investigate multi-row sand fences as a protective measure for wind reduction and sand resistance along railways,especially in the plateau region where wind power is high.

Sand fences made of various materials have been applied to control wind-blown sand depending on the geographical characteristics. For example, for sand control along the Golmud-Lhasa section of the Qinghai-Tibet Railway, upright polyethylene (PE)netting is used (Cheng & Xue, 2014; Xiao, Qu, Yao, Pang, & Zhang,2010; Zhang, Qu, Liao, Niu, & Han, 2010); along the Lanzhou-Xinjiang Secondary Railway, a concrete-insert plate sand-blocking wall and wire netting are applied (Cheng, Lei, Li, &Wang, 2016; Wang, Qu, Ling, Liu, & Xiao, 2018), and for the Taklimakan Desert Highway (currently the longest highway in the world’s mobile deserts), the reed barrier sand control system is used (Dong et al., 2000, Dong, Chen, He, Han, & Wang, 2004).However,in the plateau region,PE netting,given the poor stability and durability (Wang, Qu, Ling, Liu, & Xiao, 2018) and its weak resistance against ultraviolet(UV)light and aging(Li,Zhe,Hu,Wu,& Hua, 2017; Zhou, Zhang, Yu, & Zhang, 2013), and wire meshes,given their poor corrosion resistance, are not so suitable for the local environmental characteristics of strong ultraviolet light, high altitude, and intense salinity-related desertification (Zhang, Yu,Wang, Shi, & Feng, 2019). Recently, the high density polyethylene(HDPE) board is a potential candidate material for sand fences controlling wind-blown sand in this plateau region along the Qinghai section of the Golmud-Korla Railway, which crosses the southern edge of Qaidam Basin where the ultraviolet light is strong,the climate is dry (average annual rainfall of 28.4-44 mm) and windy,and wind-blown sand is severe(Fig.1).The HDPE board has resistance against heat and cold (e.g., Dai, Yan, Guo, Hu, & Yang,2017; Rowe & Shoaib, 2018), UV light (e.g., Becerra and d’Almeida, 2017), chemicals (e.g., Chen et al., 2018), and aging(e.g., Fairbrother et al., 2019). Meanwhile, it cannot so easily be destroyed by strong wind as the HDPE net,and it is also convenient for construction.However,the specific shelter parameters of HDPE board sand fences, such as the optimal porosity and number of rows,should be determined before their wide application along the railway.

In this study, indoor wind tunnel testing was carried out to study the flow field and sand control efficiency of single-row HDPE board fences with different porosities, with the objective of ascertaining the optimal porosity and the best protection distance for the single-row HDPE upright sand fence. In addition, wind tunnel simulations on the protective effect of multi-row HDPE board sand fences were performed using the most effective protection distance and the optimal porosity. The study results are of considerable significance for railway sand control engineering and can provide strong theoretical support for its design and improvement.

2. Wind tunnel testing

2.1. Test design

The wind tunnel testing was performed at the wind tunnel laboratory of the Cold and Arid Regions Environmental and Engineering Research Institute,Chinese Academy of Sciences,Lanzhou,China. The open-circulation wind tunnel consists of four sections:the power section, the stabilization section, the working section,and the diffusion section. The total length of the wind tunnel is 37.78 m,the length of the working section is 16.23 m,and the crosssectional area is 0.6 m high and 1.0 m wide(inner wall dimensions).The wind speed in the tunnel can be adjusted within the range of 2-40 m/s (Dong et al., 2007).

The wind tunnel simulation test consisted of the flow field and windblown-sand tests, in which control tests without sand fences were conducted first. Four HDPE board models with respective porosities of 20%, 30%, 40%, and 50% (Fig. 2a, b, c, d, respectively)were tested, all having a length, thickness, and height of 1.0 m,2 mm, and 15 cm, respectively. The height was scaled by 1:10 according to the height of the fence that currently existed along the railway.

Experimental wind velocities(v)were 12,16,20,and 24 m/s for both the flow field and windblown-sand tests,measured by a pitotstatic probe placed in the center of the wind tunnel at a height of 30 cm above the ground. For the flow field test, the wind speed profiles(Fig.3a)near the HDPE board sand fence were measured on the windward side at positions of 1H,2H,6H,and 8Hfrom the fence and on the leeward side at positions of 1H, 2H, 4H, 6H,10H,15H,20Hand 30Hfrom the fence, whereHis the height of the test model,using 10 pitot tubes at measurement heights(z)of 0.4,0.8,1.2, 2.0, 4.0, 8.0, 12.0, 16.0, 20.0, and 24.0 cm. As each run was finished, the pitot tubes were moved to the other location.

The sand transport rate was measured using a 30 cm high sand trap placed 10Hdownwind of the fence (Fig. 4a). The sand trap had 15 collection boxes with a size of 2 cm × 2 cm (Fig. 3b).Sand bed was covered by sand with a length, thickness of 4 m and 5 cm, respectively, and the particle size mainly ranged from 0.08 to 0.315 mm (mean value, 0.168 mm). Before the test, the sand transport rate for the condition without sand fences was performed, and the test duration at each wind speed in the test was determined according to the capacity of the sand trap (Li,Zhe, Hu, Wu, & Hua, 2017). Specifically, the duration was 3 min when the wind speed was 12 m/s, 2 min when the wind speed was 16 m/s,1.5 min when the wind speed was 20 m/s, and 1 min when the wind speed was 24 m/s. The tests were repeated for three times.

Fig.1. The geographic location for the Qinghai section of the Golmud-Korla Railway.

The optimal porosity and the effective protection distance for double- and triple-row HDPE board sand fences were determined by the test results of the single-row HDPE board sand fence. 30%porosity and 20H fence spacing were selected. The wind speed at each point was also measured with the pitot tube, and the sand transport rate was also measured using the same sand trap in the four experimental wind speeds of 12,16, 20, and 24 m/s. For the double-row HDPE board sand fence a row was added at the back of the single-row HDPE board sand fence at a distance of 20H(Fig.4b).The measurement points on the windward side were 1H, 2H, 4H,and 8Hfrom the first row,those between the two rows were 1H,3H,10H,17H,and 19Hfrom the first row,and those on the lee side were 1H, 2H, 4H, 6H, 10H, 15H, and 20H(Due to the length of the test section,30H cannot be measured)from the second row(Fig.3c).For the triple-row HDPE board sand fences,two rows were added at the back of the single-row HDPE board sand fence at a distance of 20H(Fig. 4c), and for the measurement points of the triple-row HDPE board sand fence, those on the windward side of the first row and those on the leeward side of the third row were the same as those for the case of the second row fence.Both the measurement points between the first and the second rows and those between the second and the third rows were 1H,3H,10H,17H,and 19H(Fig.3d).For sand transport rate measurements of double- and triple-row HDPE board sand fences, the sand trap was placed 10H downwind of the second fence for double-row test and 10H downwind of the third row for triple-row test,and the test duration in different wind speeds was the same as those in the single-row test.

2.2. Data processing

The wind speed data were collected every 0.5 s for a run of 20 s,and the average of 40 speed measurements was taken as the final value.An isoline map of the wind speed around the sand fence was obtained based on the data at different distances and heights using Origin 9.0 (http://www.originlab.com) (e.g., Wang, Qu, & Niu,2020). The effect of the fence for reducing wind speed was indicated by the isoline map of the wind speed, which facilitated the evaluation of the effective protection distance of the sand fence.Therefore, the windproof efficiency of the fence can be expressed by the following formula(Wang,Qu,Ling,Xie,&Xiao,2017,Wang,Qu, Ling, Liu, & Xiao, 2018):

whereφxzis the windproof efficiency of the fence, with values ranging from 0 to 1,xbeing the distance from the fence(expressed as a multiple of the fence height) andzbeing the height from the ground in cm;vxzis the wind speed at point (x,z) in m/s with the fence protection; andv′xzis the wind speed at point (x,z) in m/s under the condition of no fence protection.

The total sand fluxQis the sum of the sand fluxes,which were obtained by weighing the sand collected in the sand trap during each test.The sand control efficiency of the fence can be expressed by the following formula:

where η is the sand control efficiency of the fence,Qis the total sand transport rate at the location 10Hfrom the sand fence in g·m-1min-1,andQ′is the total sand transport rate in g·m-1min-1when there is no fence protection.

For double- and triple-row HDPE board sand fences, the sand control efficiency is computed at 10H after the last fence.

2.3. Wind tunnel similarity

In terms of geometric similarity,it was necessary to consider the ratio of the wind tunnel turbulent boundary layer to the earthattached boundary layer, generally not exceeding 6% blocking degree to avoid the requirements of wall interference (Wu et al.,2013). The experiment scaled the model according to the scale of 1:10. Under the experimental conditions, the thickness of the boundary layer could reach about 20 cm. The maximum height of the model produced in this experiment was 15 cm,so it was in the boundary layer.

The motion similarity mainly considered the similarity of the flow state and wind speed profile,and the wind speed proflie was verified by pre-experiment and conformed to the logarithmic distribution.The simulation scale was tried to be in line with Reynolds number independence,the general requirements recommended by White (1996). Four wind velocities of 12,16, 20, and 24 m/s were selected in the wind tunnel experiments, and the flow Reynolds numbers (Re = ρvL/μ, where Re is the flow Reynolds number, ρ denotes the air density and equals 1.205 kg m-3,v is the free-stream wind velocity, L is the characteristic length at 0.98 m, and μ is the kinematic viscosity of air at 1.4 × 10-5m2s-1) ranged from 1.01 × 106to 2.02 × 106. When the Reynolds number was greater than 1 × 105, the fluid motion state did not change with the increase of the Reynolds number.

Fig. 2. The four HDPE boards with different porosities. (a) 20% porosity; (b) 30% porosity; (c) 40% porosity; (d) 50% porosity.

3. Results

3.1. Protective effect of single-row HDPE board sand fences with different porosities on wind-blown sand

3.1.1. Airflowfield over HDPE board sand fences with different porosities

Studying the trends in airflow variation around the HDPE board sand fence near the ground surface can provide important and useful guidance in understanding the formation mechanism of sand resistance because the wind-blown sand flux generally occurs in the near-surface layer of the desert, and the near-surface wind speed has a significant impact on the sand movement(Shi,Jiang,&Han, 2014).

Fig.3. Photos of measurements.(a)Measurement of the flow field over the single-row sand fence by the pitot tube;(b)measurement of wind-blown sand flux over the single-row sand fence;(c)measurement of the flow field over the double-row sand fence by the pitot tube;(d)measurement of the flow field over the triple-row sand fence by the pitot tube.

The velocityv= 24 m/s was selected to show the airflow field over the HDPE board sand fence;the results are shown in Fig.5.The airflow field over the sand fence could be divided into three zones:the deceleration zone before the fence,the acceleration zone up the fence,and the vortex zone downwind of the fence;the field varies in each zone according to the porosity of the fence. When the porosity was 20%, the airflow deceleration zone was formed downwind of the fence as the speed of the air flowing through the fence with a relatively low porosity changed significantly when encountering the obstacle (Fig. 5a). As most of the airflow moved upward after encountering the obstacle, the acceleration zone of the airflow was formed above the fence owing to the concentrated and continuously accelerated airflow during the upward movement. The vortex zone was formed downwind of the fence by the small amount of air passing directly through the holes; the vortex was generated because the airflow was shunted owing to the difference in velocity between upper and lower airflows, which produced an adverse pressure gradient impeding the forward movement of air and resulting in the backflow. The cloud map for the 30% porosity fence (Fig. 5b) was similar to that for the 20%porosity fence and was not analyzed herein. Like that for the condition of the fence with a porosity of 20%, the airflow under the condition of the fence with a porosity of 40% also had a clear deceleration zone,acceleration zone,and vortex zone(Fig.5c).The differences between the two airflow fields were that when air passed through the fence with a 40% porosity, the intensity of the airflow passing through the holes was greater than that through the fence with a porosity of 20%,the range of the deceleration zone and the acceleration zone of the airflow were reduced, and the difference in velocity between upper and lower airflow fields downwind of the fence was reduced,resulting in a decrease in the range of the vortex zone. When the porosity was increased to 50%(Fig. 5d), most of the airflow passed through the fence directly,resulting in a reduction in the ranges of both the deceleration zone and the acceleration zone and the disappearance of the vortex zone.

We defined the effective protection distance of a fence against wind speeds as the distance at which 60% of the reference wind speed is recovered below the height of the fence (Wu., 2003; Qu,Liu, Lei, Li, & Yu, 2001; Wang et al., 2005). The effective protection distances in this study were obtained by averaging the four distances in which 40%reduction in wind speeds at the respective four heights of 2,4,8,and 12 cm.If the experiment wind speed of 24 m/s was taken as the reference,the fences with porosities of 20%,30%, 40%, and 50% yielded effective protection distances of 19.5H(±1.7H), 19.8H(±1.2H), 15.8H(±2.3H), and 11.6H(±2.7H),respectively. The effective protection distance first increased and then decreased with the increase in the fence porosity. There was an optimal porosity providing the longest effective protection distance downwind of the fence because the more air that passed through the pores, the faster the initial velocity downwind of the fence was restored. The HDPE board sand fence with a porosity of 30% provided the largest effective protection distance.

3.1.2. Wind speed profile and windproof efficiency

A comparison of wind speed profiles at different locations on the leeward side of the HDPE board sand fences with different porosities(20%,30%,40%,and 50%)and that for the no-fence condition is shown in Fig.6.It can be seen from the figure that the wind speed profiles downwind of the fences with the four different porosities were different from that (the logarithmic profile) without fence,but they became increasingly close to the logarithmic profile as the distance downwind of the fence increased.Taking the experimental wind speed of 20 m/s as an example, the wind speeds in the leeward direction of the fences with the four porosities were lower than those without fence,whereas the wind speeds within 4Hand above the fences with porosities of 20%-40%,and those within 1Hdownwind of the fence with a porosity of 50%, were greater than the free wind speed. Negative wind velocities (in the opposite direction of the free wind)were generated within 2H-6Hdownwind of the fences with porosities of 20% and 30%, and about 8Hdownwind of the fence with a porosity of 40%,and there were not negative wind velocities generating downwind of the fence with a porosity of 50%, indicating that negative values downwind of the fence decrease with an increase in the fence porosity.The recovery of wind speeds within 20Hdownwind of the fence was the fastest for the fence with a porosity of 50%, followed by that with a porosity of 40%,and was the slowest for the fences with porosities of 20% and 30%, which were essentially identical.

Fig. 4. Layout of wind tunnel test for wind-blown sand flux over (a) the single-row HDPE board sand fence; (b) the double-row HDPE board sand fence; (c) the triple-row HDPE board sand fence.

Fig.5. Airflow fields over HDPE board sand fences with different porosities(velocity v=24 m/s).(a)20%;(b)30%;(c)40%;(d)50%.The sand fence was at the distance 0H(H,height of fence model).

Fig. 6. Wind speed profiles at different location on the leeward side of the sand fences with different porosities compared with that without fence (v = 20 m/s).

Fig. 7. Windproof efficiency of single-row sand fences with different porosities at different wind speeds on the leeward side (height z above wind tunnel floor is 4 cm).

Fig. 7 shows the windproof efficiencies at different locations calculated according to wind speed at a height of 4 cm downwind of the fence when the free wind speeds were 12,16,20,and 24 m/s the speed of the backflow downwind of the fence was treated as 0 when calculating wind proof efficiency. The windproof efficiencies of the fence with a porosity of 20%at different wind speeds were selected for illustration. The windproof efficiency values increased from 1Hto 2Hfor most experimental wind speeds. The windproof efficiency gradually decreased as the distance increased when the distance from the fence was beyond 8H, diminishing to about 20-30% when the distance was 30H. The windproof efficiencies at 30Hon the leeward side of the fence were 22%,26%,28%,and 29% for wind speeds of 12, 16, 20, and 24 m/s, respectively,indicating a slight decline in windproof efficiency as wind speed decreased.If the experimental wind speed of 24 m/s is taken as an example, it can be seen that the windproof efficiencies were 79%,87%, and 29% at 1H,10H, and 30H, respectively, downwind of the sand fence when the porosity was 20%; the windproof efficiencies were 76%, 98%, and 28% at 1H,10H, and 30H, respectively, downwind of the fence when the porosity was 30%; the windproof efficiencies were 71%, 88%, and 18% at 1H,10H, and 30H, respectively,downwind of the fence when the porosity was 40%; and the windproof efficiencies were 60%,45%,and 17%at 1H,10H,and 30H,respectively, downwind of the fence when the porosity was 50%.The equations for the wind speed recovery curves at 4 cm above the floor on the leeside of the sand fences with different porosities(20%, 30%, 40%, and 50%) followed a parabolic function:

WhereWris the proportion of wind recovery on the leeside of the sand fences with different porosities compared with that of without fence at the height of 4 cma,b, andcare regression coefficients,and their values andR2are shown in Table 1.In addition,it indicated that the windproof efficiency gradually decreased with an increase in porosity for the same wind speed, but the values were essentially identical for the fence porosities of 20%and 30%.

3.1.3. Sediment transport have changed to Sandflux densuty profiles

Sand flux density profiles for the control group without sand fences and those with HDPE board fences under four groups of porosity in the experimental wind speed of 24 m/s are shown in Fig.8.It revealed that sand flux density profiles under the effect of sand fences no longer showed an exponential decrease of sand flux density with increasing height as that without a sand fence.In the layer under the height of 1H,due to the barrier effect of the fence,the total flux densities were smaller than that of the control group,while in the upper layer over 1H, due to the impact of the acceleration of the fence convergence, the total flux densities became larger than that of the control group.When the porosities of HDPE board fence were 20%, 30%, 40%, and 50%, the quantities of sand collected by the trap were 549.2, 514.6, 570.4, and 615.1 g,respectively,while the quantity of sand collected in the control run of test was 915.6 g. Thus, as the porosity was 30%, sand transport rate was more reduced than the other porosities, that is, has the best sand control efficiency.

Fig. 9 shows the sand control efficiencies of the HDPE board sand fences with different porosities at wind speeds of 12,16, 20,and 24 m/s. The variability of these results was small (Standard error:0.6%～4.8%). If the location at 1.5 m (10H) downwind of the fence with a porosity of 30% is taken as an example, it shows that the sand control efficiencies of the HDPE board sand fence were 65.3%, 61.1%, 50.8%, and 43.8% when the wind speeds were 12,16,be concluded that the HDPE board sand fence had the best sand control efficiency when its porosity is 30%.20, and 24 m/s, respectively, indicating that the sand control efficiency of the fence gradually weakened as the wind speed increased. If the wind speed of 20 m/s is taken as an example, the sand control efficiencies of the fences were 45.5%,50.8%,41.3%,and 37.1% when the porosities were 20%, 30%, 40%, and 50%, respectively, indicating that the sand control efficiency of the fence increased and then weakened with the increase in porosity. It can

Table 1 Values of regression coefficients and coefficient of determination R2 in Eq. (3).

Fig. 8. Sand flux density profiles at 10H on the leeward side of the fences with four porosities (20%, 30%, 40%, and 50%) and that of the control run of test without sand fence under 24 m/s.

Fig. 9. Sand control efficiency of single-row sand fences with different porosities under different wind speeds. The sand trap was placed 10H behind the fence.

3.2. Protective effect of double-and triple-row HDPE board sand fences with 30% porosity on wind-blown sand

For the experiments whose results are given in this section,the porosity and the spacing for the double- and the triple-row HDPE board sand fences were set to 30% and 20H, which were just the optimal porosity and the effective protection distance found from the test results for the single-row HDPE board sand fence (Section 3.1).

3.2.1. Airflowfield over double- and triple-row HDPE board sand fences

Fig.10 shows the airflow fields over the double-and triple-row HDPE board sand fences atv= 16 m/s. For the double-row sand fence,the deceleration zone,the acceleration zone,and the vortex zone were generated around the first row. Within 10Hdownwind of the fence,the wind speed below the fence height dropped below the range of -7.3 to 8.5 m/s. Being blocked by the first row, the wind speeds below the height of the fence dropped below the range of-7.7 to 4.7 m/s within 10Hdownwind of the second row.The second row witnessed a greater drop in the airflow speed,and the vortex area was small downwind of this row, but the range where the airflow speed was lower than 10.0 m/s was larger than that for the single-row fence. For the triple-row sand fence, the wind speeds below the height of the fence within 10Hdownwind of the first,second,and third rows dropped below the ranges of-7.0 to 8.5 m/s, -6.6 to 4.7 m/s, and -1.4 to 4.5 m/s, respectively. After the airflow passed through the three fence rows, the vortex was only generated near the location 4Hdownwind of the fence, and the wind speed declined more noticeably.

Fig.10. Airflow fields over the double-row (a) and triple-row (b) HDPE sand fences with 30% porosity (v = 16 m/s). (a) The double-row HDPE sand fence with 30% porosity; the second row was taken as the origin,and the speed was measured at 0 for the first row of the sand fence.(b)The triple-row HDPE sand fence with 30%porosity;the third row was taken as the origin, and the speed was measured at 0 for the first row of the fence, 20H for the second row of the fence, and 40H for the third row of the fence.

3.2.2. Windproof efficiency and sand control efficiency

Fig. 11 shows the windproof efficiencies downwind of singlerow, double-row, and triple-row HDPE board sand fences with a porosity of 30%.If the experimental wind speed of 20 m/s is taken as an example,it can be seen that the windproof efficiencies were 74% and 46% at locations 10Hand 20H, respectively, downwind of the single-row fence;the windproof efficiencies were 93%and 59%at locations 10Hand 20H, respectively, downwind of the doublerow fence; and the windproof efficiencies were 95% and 62% at locations 10Hand 20H, respectively, downwind of the triple-row fence. The windproof efficiency downwind of the fence increased markedly as the number of rows increased.

Fig. 11. Windproof efficiency of sand fences with different numbers of rows at different locations on the leeward side (z = 4 cm).

The sand control efficiencies of the double-and triple-row HDPE board sand fences are shown in Fig.12.The sand control efficiencies of the double-row and the triple-row fences were 92.1%, 90.3%,86.2%, and 79.5%, and 96.5%, 92.3%, 88.6%, and 82.8%, respectively,when the wind speeds were 12, 16, 20, and 24 m/s, respectively,indicating that the sand control efficiency of the fence weakened with the increase in wind speed.Compared with that of the singlerow HDPE board sand fence, the sand control efficiencies of the double-row and the triple-row sand fences went up by 26.8%,29.2%, 35.4%, 35.7%, and 31.2%, 31.2%, 37.9%, 39%, respectively,indicating that the greater the wind speed, the wider the range of the increased sand control efficiency. It can be concluded that the sand control efficiency after adding one or two rows was not double or triple that of the single-row fence,indicating that the first row of the HDPE board sand fence plays the major protective role but still performs poorly at high wind speeds, and the protective effect downwind of the fence can be improved by increasing the number of rows. Compared with the double-row sand fences, the effect of adding the third-row sand fence is limited. Therefore, double-row HDPE board sand fences are recommended to control windblown sand in the plateau region suffering from severe natural conditions.

Fig.12. Sand control efficiency of sand fences with different numbers of rows under different wind speeds. The sand trap was placed 10H downwind of the last row.

4. Discussion

HDPE board sand fence was experimentally applied to control drifting sand for the construction of the Golmud-Korla Railway.When the porosities of single-row HDPE board sand fence are low(20%and 30%),the deceleration zone,acceleration zone,and vortex zone of the airflow are generated around the fence,and some zones around the fence shrink or disappear when the porosity increases(to 40%or 50%),which is consistent with the results of the study by Lee and Kim (1999). The test results also show that the windproof efficiencies of fences with porosities of 20% and 30% are considerably greater than those with other porosities; the sand control effect of sand fence with a porosity of 30%is superior to that of others.It can be concluded from the wind speed recovery, the windproof efficiency,and the sand control efficiency that the optimal porosity value and suitable spacing for the HDPE board sand fence are 30%.Tsukahara, Sakamoto, Aoshima, Yamamoto, & Kawaguchi, 2012 used laser sheet visualization technology to measure the flow fields around sand dunes and the erosion of sand dunes, and reported that a porosity of 30% most satisfactorily inhibited wind erosion. Obviously, results of our research were consistent with those of previous studies. These results are applicable for doubleand triple-row HDPE board sand fences.

Sand fences in strong wind areas adopt mainly concrete-insert plate sand-blocking wall, punched steel plate, and wire mesh fences(Wang,2018).This study showed that the windproof efficiency and sand control efficiency at 10Hin the leeside of the double-row and triple-row HDPE board fences are 93.0% and 95.0%, 86.2% and 88.6%, respectively. Cheng, Xin, Zhi, & Jiang, 2017 revealed that when experimental wind speeds were 15 and 30 m s-1, the windproof efficiency of the concrete-insert plate sand-blocking wall with a porosity of 30% exceeded 80%; however, although the windproof efficiency was good, the sand control efficiency was poor due to the large hole size of the board(27.5×3 cm).Wang,Qu,Ling, Liu, & Xiao, 2018 conducted wind tunnel tests on the shelter effect of double-row punched steel plate and wire sand fences.Results showed that under the experimental wind speed of 22 m/s,the maximum windproof and sand control efficiency of the doublerow punched steel plate fences were 87.0%and 84.2%,respectively,while the maximum windproof and sand control efficiency of the double-row wire mesh fences were 60.4% and 83.5%, respectively.Thus, the shelter efficiencies of both double-row punched steel plate and wire mesh fences were smaller than that of double-row HDPE board fences in this study. It can be concluded that in windy areas, the shelter efficiency of the HDPE board fence is higher than that made of other materials.Consequently,due to the excellent durability and significant shelter efficiency, the HDPE board fence is suitable for wind-blown sand control in the Gobi area with strong wind along the Golmud-Korla Railway. Gillies et al. (2017) measured changes in aeolian sand flux over multirow sand fences at the Oceano Dunes in California, USA. Results showed that sand flux decreased exponentially with an increase in distance.The sand flux reached equilibrium at ～93 fence heights for the 10Hspacing and ～27 fence heights for the 7Hspacing, where the sand flux decreased by 59.0%and 91.0%,respectively,indicating that sand fences with a 7Hspacing had a better sand control effect.In that study,the sand flux over the triple-row sand fence with 20Hspacing at 50 fence heights decreased by 89.6%compared with the sand flux without the presence of sand fences. They showed that sediment loss rates at a downwind distance of 50Hin 10Hspacing and in 7Hspacing were 50.0%and 91.0%,respectively,and the latter value for sand flux control index was in accordance with the result of our wind tunnel tests. A comparison indicates that the sand control effect achieved by triple-row sand fences with 30%porosity and larger spacing (20H) is almost as great as that of sand fences with 50% porosity and a smaller spacing (7H), and thus there may be an optimal combination of fence porosity and row spacing for achieving the best sand control effect.

With an increase in the free wind speed, the windproof efficiency changes little, but the sand control efficiency gradually weakens, especially the sand control efficiency of the single-row HDPE board sand fence under strong or extremely strong wind.The wind tunnel tests show that the sand control effect along the railway could be improved by increasing the number of rows of HDPE board sand fences.According to the results for the windproof and sand control efficiencies given in this paper,double-row HDPE board fences with a porosity of 30%are suitable for the plateau sites along the Golmud-Korla Railway with extremely strong wind and radiation. However, as only those double- and triple-row HDPE board sand fences with a porosity of 30%were covered in this paper,it would be well worthwhile to more deeply investigate the protective effect of multi-row sand fences with different arrangements(such as different distances) and porosities in future studies. In addition,as only double-row HDPE board fences with the porosity of 50% were established in the in-situ test section of the sand control system, the comparison between wind tunnel and field observation results cannot be systematically performed, and thus further comparison will be needed in future studies to better guide the application of wind tunnel results to actual practice.

5. Conclusions

With the aim of facilitating the control of wind-blown sand disasters along the Qinghai section of the Golmud-Korla Railway,wind tunnel testing was carried out to determine the windproof and sand control efficiency of HDPE board sand fences with different porosities and rows, and the optimal porosity and the effective protection distance of the single-row fence were determined. When air flows through a fence with a porosity of 20% or 30%, different velocity zones are formed, that is, the deceleration,acceleration, and vortex zones. When the porosity is increased to 50%, the vortex zone disappears. From the recovery of the wind speed at 20Hfrom the single-row HDPE board sand fence, it indicates that the speed recovery is the highest when the porosity is 50%, followed by the porosity of 40%, and the lowest is when the porosity is 20% or 30%. For a given wind speed, the windproof efficiency gradually decreases with an increase in porosity, but with fence porosities of 20% and 30%, the windproof efficiency is essentially the same.The sand control efficiency first increases and then decreases with an increase in porosity, and the sand control efficiency is the highest when the porosity is 30%.

The windproof and sand control efficiencies of the triple-row sand fence are greater than those of the double-row sand fence and clearly better than the wind reduction and the sand resistance effects of the single-row sand fence. The result of the final test shows that the sand control efficiency of the single-row HDPE board sand fence is considerably lower than those for multi-row fences, while there is little difference in the sand control efficiency between double-row and triple-row HDPE board sand fences. Consequently, double-row HDPE board sand fences with the porosity of 30% are recommended to control wind-blown sand along the Golmud-Korla Railway. The research results for multirow HDPE board sand fences can provide direct support for the newly built Golmud-Korla Railway sand control system and can also serve as a reference for systems protecting railways and highways from wind-blown sand in other areas.

Acknowledgments

This research was financially supported by the Young Scholars Science Foundation of Lanzhou Jiaotong University, China(2018016) and the Chang Jiang Scholar and Innovation Team Development Plan of China (IRT_15R29). The authors express appreciation for the anonymous reviewers and the editor, who helped improve the quality of this paper.