Experimental investigation of the effect of permeability and angle of gabion submerged vane on bed topography

Submerged vanes are used as an eco-friendly low-cost measure for the management of river bed sediment. The performance of the vane is more sensitive to its angle of installation. A high angle increases the vane’s nose scour which results in its failure, and a lower angle reduces the vane’s performance. The design angles for the impenetrable vane are known, whilst for the impermeable vane are not clear. Therefore, it is our goal to experimentally examine the performance of permeable submerged gabion vane (SGV) with different permeability and different installed angles (α) in two conditions of clear water (CW) and live bed (LB). Experiments were performed in an 11-m-long and 0.66-m-wide bed flume. A SGV with a size 0.16-m wide, 0.80-m high, and 0.20-m thick was installed in the center of the flume with the submergence ratio (T/do = 0.5). The flow depth was 0.16 m, and the flume bed was covered with uniform sand bed (d50 = 0.48 mm). The SGV was filled with four different sizes of very fine, fine, medium, and coarse gravels and installed at different angles: α\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\alpha$$\end{document}=10°, 15°, 20°, 25°, 30°, and 35°. Comparing the maximum scour depth around the SGV at the end of each experiment, it shows that the vane’s nose scour is minimum at α = 20°, whilst for the LW condition, the minimum scour depth occurs at α = 25°. Moreover, under all hydraulic conditions, the results reveal that the permeability of the gabion can reduce the vane’s scour depth and the minimum scour depth was for gabion filled with gravel (D50 = 4 mm).


Introduction
The equilibrium conditions of rivers change for various reasons. As a result, the instability and destruction of river banks destroy agricultural lands and infrastructures and supply a large amount of sediments into the rivers. Hence, researchers have developed different measures to provide necessary protection against possible hazards. These techniques are classified into coverage and flow pattern modification categories. Recently, there has been a great interest in eco-friendly flow pattern modification measures for river bed sediment management (Pagliara and Kurdistani et al. 2017, Odgaard 2009, Yeo et al. 2005, and Rosgen 2001. Submerged vanes are flow pattern modification structures installed in-series on the bed with a specific angle relative to the flow direction, a certain transverse distance from the river bank, and specific longitudinal spacing. A wake vortex is created by installing these vanes, changing the transverse distribution of the bed shear stress which results in improving the lateral distribution of the bed sediment. The first efforts were made by Odgaard and Spoljaric (1986) to develop the basis of the theory for designing impermeable submerged vanes at the University of Iowa. Accordingly, these vanes are also known as Iowa Vanes. The first experiments by Odgaard and Wang (1990) proved that by designing a proper arrangement of vanes, the flow and sediment motion at river bends could be regulated as if it moves in a straight river path. Field experiments by Odgaard and Mosconi (1987) and Fukuoka and Watanabe (1989) indicated the application of submerged vanes in stabilizing the outer bank. Odgaard and Wang (1991b) experimentally Responsible Editor: Amjad Kallel proved that the flow patterns could be substantially changed by a proper arrangement of vanes without a significant change in the flow energy gradient which causes redistributed bed shear stress and sediment transport around the vanes. Vanes with the proper shape, dimensions, and alignment placement can effectively redistribute the bed shear stress such that, the sedimentation occurs on one side of the channel cross-section and create bed scouring on the other. The initial studies and most of the subsequent studies regarding the shape of submerged vanes were rectangular plates. Other shapes such as airfoil-like cross-sectional profiles (Odgaard and Spoljaric 1986), a trapezoidal tapered plate, and a plate in the shape of a parallelogram with its top swept forward or backward to the approaching flow (Ouyang 2009), curved and angled vanes (Behbahan 2011) and rectangular with leading and curved edge (Azizi et al. 2012) have also been experimentally studied. The results by Azizi et al. (2012) show that cutting the leading edge of the vanes leads to a decrease in the local scour around the vanes. On the other hand, Azizi and Shafai Bajestan (2020) found that the moment of the momentum of the tapered vane is reduced, and thus, the longitude distance between the vanes should be selected in the lower range as recommended by Odgaard (2009). The effect of submerged sheet pile vane also has been extensively investigated experimentally by Boniford et al. (2015). Their results showed that the uneven surface of the sheet pile reduces the erosive strength of the vortex and has better performance. Ouyang and Cheng (2016) developed a numerical model to determine the effect of three different vane shapes on the induced transverse bed profiles. Their innovation approach showed that the tapered vanes are the most effective compared to the forward-swept and rectangular shape vanes if installed at a lateral space of 1.6 water depth. The effects of the field application of submerged vanes on river sediment management and river bank protection have received much attention since it was innovated by Odgaard and Spolarjic (1986). The first field study was conducted to control the scour at the outer bank of the Nishnabotna River in Iowa (Odgaard 1987). Sheet piles vanes were installed against outer bank scour at the meandering river bend in Taiwan (2009). Shafai Bejestan and Forughi (1993) designed and installed permeable wovenstraw vanes with an installed angle of 25° in the Karkheh River, Iran. Their results showed because of the permeability of the vanes, the pressure difference on the two sides of the vane decreases minimizing the drag force which can reduce the local scour around the vanes. Benjamin et al. (2021) studied the effect of vanes on bluff stability in the USA. Rodriguez et al. (2020) investigated the application of submerged vanes in 5 projects in Colombia.
The results of Tan et al. (2005) also showed that a vane higher than two to three times the height of the bed form would create too much blockage to the flow and reduce the mobility of the sediment significantly. Conversely, if the vane is lower than the optimum height, bed load particles can easily climb over the vane, thereby reducing its effectiveness. All mentioned studies have been conducted on impermeable submerged vanes made of concrete, steel, and wood. However, due to strict environmental regulations in many countries, permeable eco-friendly river sediment management has attracted researchers in recent years. Cunningham and Lyn (2010) experimentally studied the effect of submerged gabion weirs in a 90° river bend on the bed topography. The effect of permeable groins made by sixlegged elements was experimentally studied by Kalamizadeh et al. (2021) and Najjaran et al. (2020). Ferro et al. (2019), Hajibehzad et al. (2020), and Shokrian- Hajibehzad et al. (2022) studied scouring around a permeable groin structure combined with a triangular vane in a 180° flume bend. Teraguchi et al. (2010) and Teraguchi et al. (2011) investigated the characteristics of the field flow and scouringsedimentation mechanism of the live bed around permeable and impermeable spurs and bandal-like structures. Their results showed that permeable spurs reduce the velocity near the river bank, and reverse eddies behind permeable groins prevent over-siltation. The analysis of the scouringsedimentation pattern showed that small scour holes are uniformly formed around piers (piles) of permeable groins. Zhang et al. (2010) argued that the basic characteristic of bandal-likes is the upper permeable section of this structure which divert high-velocity flow in the vicinity of the water surface to the main channel, acting as a groin. To improve the performance of the bandal-like vanes, Sardasteh et al. (2021) studied a new structure which is a combination of bandal-like and triangular. The study conducted by Zhang et al. (2013) shows that permeable spurs decrease their nose scour by 50% compared to the use of impermeable spur in river management.
Eco-friendly techniques are used in river rehabilitation projects to control the river bed, stabilize the river direction in river bank projects in meandering bends, and reconstruct natural habitats. Strict environmental regulations have been imposed not to use rigid structures such as concrete or masonry materials in rivers, emphasizing the use of natural permeable measures. Gabion vane which is made of natural stone can be a good choice for river restoration. However, the criteria for selecting the vane's angle and its permeability is unknown and no specific study, at least to the best of our knowledge, has been found in the literature to recommend these parameters for permeable vanes. Therefore, it is the main goal of this study to conduct an extensive experimental study on the effect of the vane's angle and its permeability on the performance of a submerged gabion vane on bed topography. For this reason, gabion vanes were filled with four different gravel sizes, installed at five different angles of attack to the flow direction experimentally investigated under two different flow conditions of clear water and live bed.

Dimensional analysis
The effective variables in bed scouring due to the installing gabion submerge vanes in this experiment are as follows: Parameters in this equation are flow density (ρ), density of bed sediment (ρ s ), the dynamic viscosity of water (μ), the acceleration of gravity (g), average flow velocity (V), discharge (Q), submerge vane length (L), submerge vane height (H), submerge vane thickness (t), vane angle of attack (α), the flow depth (y), flume width (B), average particle diameter (d 50 ), maximum depth of the scour hole (d z ), and critical velocity of sediment particles (V c ).
If three variables of flow depth (y), flow velocity (V), and flow density (ρ) are selected as repeated variables, and applying Buckingham's π theory, the following general nondimensional relation will be obtained: Because the values of V c , B, H, L, t, ρ s , and y are constant during the experiments, Eq. 2 can be summarized as follows: In this equation, Fr = Froude number, and since the minimum Reynolds number in this study was 0.16 which shows fully turbulent flow and thus the effect of Reynolds number is negligible and can be deleted from analysis, the final general dimensionless relation is as follows:

Experimental setup and equipment
Laboratory equipment used in this study includes a glass fiber-reinforced plastic (GRP) flume with a fully horizontal semi-circular cross-section with a diameter of 1200 mm and a length of 13 m. To reduce the inflow turbulence, a (1) f , s , , g, V, Q, L, H, t, , y, B, d 50 device made with straw in the form of a honeycomb was placed at the inlet. The desired flow discharge was supplied from a reservoir by a 100-mm-diameter polyethylene pipe. A valve was installed on the inlet pipeline to control the flow discharge entering the flume. A small sharpedged metal rectangular weir with a width of 600 mm was installed at the end of the flume to control the water surface in the flume. A magnetic flow meter with a precision of ± 0.1 l/s was installed on the inlet pipeline to measure the flow discharge ( Figs. 1 and 2a). The flume floor 100-mm thick was filled with sedimentary materials of silica sand of a median diameter (d 50 ) of 0.48 mm. The finished flume width at the bed is 660 mm. Figure 2b shows the particle size distribution of sediments. The sediments have a standard deviation of √ (d 84 ⁄d 16 ) = 1.75 and a uniformity coefficient of C u = d 60 ⁄d 10 = 3.05, indicating almost uniform materials. The critical shear velocity of the bed sediments is calculated from the Melville Eq. (1997): The critical flow velocity is obtained from the logarithmic velocity distribution relation (Eq. 7): which was found to be equal to V c = 0.302 m/s. or Fr = 0.16.
Submerged vanes in this study were made using resistant wire mesh with a mesh size of less than 2 mm by cutting, bending, and welding. Rectangular submerged vanes, due to ease of construction and installation, are the simplest and most practical ones (Azizi et al. 2012). The scour depth around the rectangular shape also is higher than other vane shapes (Gupta et al. 2010); therefore, rectangular vanes shape were used in this study (Fig. 2C). The dimensions of each vane were calculated based on the design recommendation of previous studies (Odgaard 2009) which is presented in Table 1. The total height of each gabion vane is 160 mm of which 80 mm of its height is sunk into the bed so that the vane is stable.
Four different size distributions (Fig. 3) of white calcium carbonate were selected to fill the gabion with the following specifications:

Experiments method
In each experiment, the surface of sedimentary bed materials was first leveled and controlled by the laser meter. Submerged vanes filled with materials of different particle size distributions were installed 6 m from the beginning of the flume main channel on the bed with different angles of 10°, 15°, 20°, 25°, 30°, and 35°. The vane was placed at the flume center. Water entered the flume by closing the tailgate and gradually opening the inlet valve to prevent scouring of the leveled bed. Once the water level in the flume reached the top of SGV (160 mm), the inlet valve and the tailgate were gradually opened so the flow discharge and the submergence ratio reach the desired values simultaneously. The submergence ratio recommended by Odgaard 2009 to be in the range of (0.2-0.4) y and since the height of the gabion vane above the bed in all experiments is constantly equal to 80 mm, the flow depth in all experiments was kept constant equal to 160 mm. In all tests, the water depth is considered to be a constant value of d 0 =160 mm. Because the cross-section of the flume is semi-circular, the cross-section of the flow is almost similar to the trapezoidal cross-section with the side slope Z as defined in Fig. 4c. The origin of the reference system is positioned at the center of the vane and on the undisturbed bottom of the channel (Fig. 4a, b). For the main experimental program, two series of experiments were carried out in this study. In the 1st series, the clear water condition was established using flow discharge equal to 24.44 l/s, the average flow velocity is 0.184, or Fr = 0.16, and the ratio V/V c = 0.66. For the 2nd series, the flow discharge was 50 l/s, the average flow velocity = 0.377 m/s, Fr = 0.33, and the ratio V/V c = 1.24 which the live bed condition exists. In each series,  (16 h) showed that 86% of the maximum scouring depth occurs in 3 h. At the end of each experiment, water in the flume was slowly completely drained and then the bed level was measured by a laser meter at different points at intervals of 20 mm. By applying these data and the use of Tecplot software, the bed topography was plotted to have a better view of the location and dimensions of scour hole and point bars. Table 2 shows the characteristics of each experiment and the maximum scour depth around SGV.

Observation (clear water condition; Fr = 0.16)
The submerged vane is installed in series. To observe the sediment patterns around the vanes, four gabion vanes of the same permeability and angle are installed in the bed flume (Fig. 5).
Observations during the experiment showed that scouring initiates from the nose of the vane; sediments are transported behind the vane and gradually deposited forming a longitudinal point bar behind the vane. The main reason for initiating scouring from the toe (nose) flows collision to the vane nose and the formation of a downward flow velocity component, forming a spiral vortex around the vane nose. This vortex strongly throws separated materials upwards and transports them downward by the longitudinal component of flow velocity. According to previous studies (Odgaard 2009 andAzizi et al. 2020), a transverse vortex behind the vane is developed which transports the eroded sediment behind the vane. Topographic maps at the end of the experiments show that the volume of scour depth and the volume of deposited sediments are almost equal. This is because, in live bed conditions, sediments are not transported from upstream. The rate of scour is very fast at the beginning of the experiment and it gradually reduced and remain relatively unchanged after almost 260 min. The dimensions and volume of the deposited sediment vary with changes in the vane angle and permeability of the gabion materials. In the case of greater scour depth, the dimensions of the sediment deposited are higher. Figure 6 shows the bed topographic map for four of the experiments.

Observation (live bed conditions; Fr = 0.33)
In the live bed condition, a single vane is installed in the flume bed (Fig. 7) due to the bed form formation which exists. Little information is available on scour around the vane at live bed conditions since almost all of the previous studies have been carried out under clear water conditions, whilst rivers in the flood seasons experience live bed conditions. Therefore, in this study, to investigate the effect of bed form movement on scour hole dimensions and geometry of the point bars, all the experiments which were performed in clear water conditions (V/V c = 0.66) were performed for the live bed conditions (V/V c = 1.24) too. Observations during the experiment showed that once the experiment starts, the bed forms, ripples, or small dune is developed from the beginning of the flume and moves downstream. The scour also rapidly initiates from the nose of the vane. The scour hole rapidly grew and extended around its tip. The main difference observed in live bed condition was that the ripples when passing the vane, the scour depth decreases, and then the scour depth increases. The maximum scour depth is gradually decreased over time. Transported sediments are deposited in the side of the vane downstream of the toe, forming a longitudinal point bar where its height and dimensions are increased over time. The point bar dimensions vary with the vane angle and permeability of the gabion materials. Figure 8 shows a view of the bed form during the live bed conditions for three experiments as an example. The important point that was observed in live bed experiments was that by passing the bed form through the vane, the scour depth decreases at first (due to pouring out the sediment into the scour hole), but immediately returns to the previous conditions due to transport of sediment out of scour hole under the influence of three-dimensional flow vortices patterns. Of course, in this case, the height and length of the point bar behind the vane are more than in the case of clear water conditions. This is because sediments are not only supplied from the scour hole but also carried from upstream by the flow. Therefore, the scour depth at the end of the experiment depends on whether the bed form enters the hole or not. Because of this at the end of the duration test (3 h), we tried to make sure that at the time of the shot down of the pump, the bed form is not very close to the scour hole, and if it is so, we wait for a few minutes (10-15 min) until it passes the vane. We took this precaution so that the comparison of scour depth in different experiments is not partially a function of the filling of the hole by the bed form.
During experiments, the transported sediments are transported toward one side of the vane and accumulated and form a longitudinal point bar. The same observation also was reported by Odgaard (2009). Odgaard (2009) has shown that once the vane is installed at the proper angle, a secondary rotational flow will be formed behind the vane which redistributed the bed shear stress causing accumulation of sediment at one side of the vane. As a result, the bed rises on one side and goes down on the other side of the vane. A spiral vortex is also formed at the left corner of the nose by colliding the flow to the vane head thickness placed in an angular position relative to the flow and diversion of downward streams behind the vanes, separating bed materials at the toe. The separated materials are transported by the flow toward the middle part and behind the vane, forming a point bar. Figure 9 shows a view of the bed topography in LB experiments. For the case of live bed conditions, due to high flow velocity and thus higher strength of the vortices, the maximum scour depth is usually higher than the CW condition; in some experiments, the scour depth was so deep it caused to unstable the vane.

Observation vortex behind the vane
The pressure difference on both sides of the submerged vane will generate a horizontal rotational current downstream. This is because the velocity component on the upstream face of the vane (high-pressure zone) is upward, whilst it is downward at the downstream face (low-pressure zone). Thus a larger spiral vortex from the upper edge corners of the vane is developed. The vortex is transferred downstream by the flow, and a longitudinal spiral vortex (LSV) in the channel is developed. This vortex is the main reason for redistributing the bed shear stress distribution and consequently the bed sediment. The strength of LSV depends on the approaching flow velocity as well the vane's submergence and the vane's angle. To observe how long is the vortex under two different flow conditions, the dye was injected at the upstream face of the vanes and the results are shown in Fig. 10. These effects are also observed in the LB condition, experiment # VF25, up to 20H from the vane (Fig. 10a). For the CW conditions because of low flow velocity (Fr = 0.16), the length of vortex behind the vane is 4H (Fig. 10b). The study of Azizi et al. (2020) also found the same vortices.

Bed cross-section profile
By installing a submerged vane, a secondary current due to existing of an upward velocity component on the upstream or highpressure side and a downward velocity component on the lowpressure or downstream side is developed (see Fig. 11 (Odgaard 2009)). The resulting vortices roll up to form a large vortex (tip vortex) springing from a position that is somewhat below the top elevation of the vane. This vortex causes transverse sediment transport which results in degrading on one side and aggregating on the other side of the channel (see Fig. 12 (Odgaard 2009)).
Experiment observations showed that these bed topography changes are mostly formed around the vane and a little downstream of the vane. And the dimensions of scour depths and point bar formed in this area are larger; therefore, to measure the scour depth and point bar heights, two cross-sections, one   Fig. 13. The vertical axis represents the variation of bed level (dz/d 0 ), scour depth or the point bar height and the horizontal axis is the relative distance (dy/d 0 ) from the flume center. Figure 13a and b show that, in general, the scour depth and point bar height is high for the case of the vane installed at α = 30°, whilst for α = 20, the scour depth and point bar height is low. Figure 13a shows that the largest scour holes are formed on the two sides near the vane, and the depth and width of the scour hole in front of the vane are greater than behind the vane. In this case, the excavation zone extends more in the zone close to the vane. Figure 13b (at a distance of 40 cm downstream the center of the vane) shows that the largest and most point bar is formed

Effect of the vane's angle
In addition to the shape and vane's dimensions, the vane alignment or the angle of the installed vane to the approaching flow is important for the best performance of the vane on redistributing the bed sediment. Installing vanes at a higher angle of attack leads to greater the strength of vortex behind the vane. This allows the vanes to be installed with greater longitudinal and transverse distances than each other. However, this will increase the scouring depth around the vane which causes the failure of the structure. Therefore, in the present study, the effect of SGV at a different angle of attack is tested under both CW and LB conditions to find out the best angle for less scour around the vane's tip.

On-scour depth
To see the effect of α on the depth of scour at the nose of the vane, the ratio of scour depth to the flow depth was plotted versus the vane's angle for different gabion's permeability as shown in Fig. 14a, for the clear water condition and Fig. 14b for the live bed condition. For the CW condition (Fig. 14a), as can be seen, the scour depth decreases once α increases from 10 to 20°. Then, the scour depth starts to increase sharply until α = 25°, then the rate of increase of scour depth slows down. The scour around the vane's tip is similar to the scour around the bridge abutment and mainly depends on the strength of the helicoidally vortex (HV) around the vane's tip which depends on the percentage of blockage of the flow surface due to the vane or (A v cosα /A) *100 in which A v is the vane's area and A is the flow cross-section area. For a small angle, the blockage is small and the strength of the helicoidally vortex is small. As the α increases, the blockage area  increases, and thus, the strength of the helicoidally vortex increases which leads to more scouring at the vane's tip. For the case of the impermeable submerged vane, the angle has been recommended to be between 10 and 20° by Odgaard (2009). Gupta et al. (2010) recommended the angle to be less than 20° to have less scour at the vane's tip. On the other hand, the experimental results of Tan et al. (2005) on the use of a submerged vane to divert more sediment without considering its effect on scour showed that it should be about 30°. The submerged vanes have been and are being installed with angles 10-20° to the flow direction in several rivers throughout the world which can be referred to the field installations such as Maording bend of Feng-Shan Creek, Taiwan (2009);Missouri River, USA (2009);and Kosi river, Nepal (2009). Gupta et al. (2010) stated that the use of an angle of attack higher than 20° is not practical in the field. Our results also show that for the case of permeable gabion vane, the best performance of the vane to have less scour around the vane's tip is that the attack angle should be in the order of 20°. For a higher angle of attack, more scour depth can be expected. Our results also show that the permeability of the vane can affect its performance on scour depth around its tip. As shown in Fig. 14b, although the best angle of attack for all permeability tested near is 20°, the scour depth around its tip has differed as the vane's permeability changes. As the size of the material inside the gabion is larger or the vane is more permeable, the scour depth decreases. This may be discussed by the fact as the vane's porosity increases, more portion of flow passes through the vane, and thus, the strength of the down flow velocity components and consequently the strength of helicoidally vortex decreases which leads to less scour depth. The percentage reduction of scour depth gabion vane filled with C material compared to the gabion vane filled with VF material, which can be calculated using the data in Table 2, is in the order of 67%, 85%, and 46% if installed at α = 10°, 20°,and 35°, respectively. These results show that the best performance for use of the gabion vane, with respect to having less scour depth, is to be filled with larger gravel and installed at α = 10-25°. For the LB conditions, as shown in Fig. 14a, the scour depth starts increasing as α increases for all vane's permeability. However, the rate of increase in dimensionless scour depth is low at first until it reaches α = 20° Then, the rate of increase is increasing sharply. The percentage reduction of scour depth gabion vane filled with C material compared to the gabion vane filled with VF material, which can be calculated using the data in Table 2, is in the order of 23%, 15%, and 20% if installed at α = 10°, 20°, and 30°, respectively. In the case of LB, the flow intensity is higher and so the strength of the vortices is higher which results in higher scour depth. In the LB condition, the scour depth for the gabion filled with coarse material and installed at α = 10°, α = 20°, and α = 35° increased by as much as 204%, 840%, and 170% compared to the CW condition respectively. Comparison of our results for the case of LB condition at an angle of attack of 20° with the results of Boniford et al. (2015) (α = 20° and Fr = 0.29) show that the scour depth ratio (y/d 0 ) for the SGV filled with VF in our study (α = 20° and Fr = 0.33) is equal to 0.69 which is approximately equal to the scour ratio for the plane vane in the study of Boniford et al. (2015). This ratio is obtained for the SGV filled with coarse material in our study to be 0.59, whilst in their study is 0.42 for the sheet pile vane. This is due to the different surface roughness of the vanes used in these studies.

On-point bar height
Our experiences in the test of LB conditions also showed that after installation of the vanes in series, sedimentary point bars are formed behind the vane at one side and bed degradation at the other side. Installation of the SGV in the field will result that the eroded bank will be repaired in 2 to 3 years. Since this important function of the SGV has not been studied so far, we examined the effect of the installation of different types of SGV installed at a different angle on the height of the point bar which is formed behind the vanes. The results are presented in this section. Figure 15a and b shows the variation of the dimensionless height of the point bar versus the angle of attack for different types of SGV and the two conditions of clear water and live bed, respectively. Figure 15a (CW conditions) shows the variation of dz max /d 0 of the point bar behind the vanes for different α and permeability. The plotted curves in this figure are similar to Fig. 14a. This is because the same amount of sediment which has been deposited is supplied from the scour hole. The minimum point bar height occurs for α = 20° which has the minimum scour depth. As the permeability of the gabions increases the scour, the point bar height also decreases. Figure 15b (LB condition) shows the variation of dz max /d 0 of the point bar behind the vanes for different α and permeability. As shown, the height of the point bar increases gradually by increasing the vane angle from 10 to 25° and the rate of increase of dz max /d 0 increases sharply when the vane's angle varies from 25 to 30°.

Effect of permeability
The experimental results under both hydraulic conditions showed that the scour hole dimension and depth and point bar dimensions and height increase when fine-grained materials are used in the body of submerged vanes. According to the results in Fig. 14a and b, the scour depth and, Fig. 15a and b, the point bar height decrease when the permeability of the gabion vane is changed from fine-grained to coarse-grained materials. As shown in Fig. 14a and b, the maximum relative scour depth decreases by increasing the permeability of the gabion. Referring to the results presented in Table 1, the percent reduction of scour depth of the gabion filled with coarse material compared to the gabion filled with very fine material for the CW condition and α = 20o is 57%. Zhang et al. (2013) also found that the scour depth at permeable spur is 50% less than the scour depth of impermeable spur. The reduction of scour depth at LW conditions, however, is less than the case of CW conditions. Referring the data in Table 1 for α = 20o, the scour depth at gabion filled with coarse material is 15% less than the gabion filled with very fine material. Teraguchi et al. (2010) and Teraguchi et al. (2011) by investigating the characteristics of the flow and scouring-sedimentation mechanism of the live bed around permeable and impermeable spurs found that permeable spurs modify the flow patterns in such that less scour depth occurs around the permeable spurs. According to Fig. 15a and b, the relative height of the point bar decreases with increasing the permeability of the body materials at a constant vane's angle.

Conclusions
Permeable concrete or sheet pile submerged vane has been used for bed sediment management in different projects around the world. However, due to environmental policy restrictions in many countries, the use of concrete or sheet piles is prohibited. In this regard, the gabion submerged vane can be a proper alternative. Gabion vane is permeable and flow can pass through the void which may affect the strength of the vortices behind the vane. Therefore, the design criteria may be different than the impermeable vane which already have been developed. Therefore, it was the main goal of this study to experimentally carry out experiments for gabion submerged vane to determine the effect of the vane's angle and gabion permeability on the scour and sedimentation around the vane. The previous study was carried out in clear water conditions, and in the present study, both clear water and live bed conditions were performed. The gabions are filled with four different gravel sizes to investigate the effect of their permeability and installed at different angles. The following results were obtained from this study: 1 Gabion submerged vane can perform well for river bed sediment management. 2 The permeability of the gabion has a great effect on the scour depth. In LB condition, for the case of gabion filled with coarse material installed at 20°, the scour depth can be reduced by 15% than the gabion filled with very fine material, whilst the percentage reduction for the same vane's angle for the CW is 57%. 3 In LB condition, the maximum scours depth for the gabion filled with coarse material and installed at α = 20° increased by as much as 840% compared to the CW condition, respectively. 4 The best performance of the gabion submerged vane in LB and CW condition was found when the vane was installed at 20°. 5 In CW condition, a lower or higher vane's angle than 20° can result in higher scour depth which may cause the failure of the structure and is not recommended. 6 In LB condition, the vane can be installed at an angle of 10-25°. For α = 20°, the maximum scour is less and is recommended. 7 In LB condition, ripples are formed on the flume bed which moves downstream and can affect the scour depth. 8 The results presented in this study are based on flume experiments. It is advised that more studies should be conducted in larger flumes and under different flow conditions before they can be used in the field. The criteria presented here is for the rectangular shape of the gabion submerged vane and for other vane shapes further study is needed.