MODELLING OF DEPOSITION AND EROSION PROCESSES ALONG A 180° OPEN CANAL BEND BY NAYS2DH IN iRIC

This paper presents experiments and computer simulation for the erosion and deposition processes along a 180° open canal bend. The computer simulation is conducted by Nays2DH model in iRIC software. The models are conducted with the curvature's radius ratio (Ø/L) varies within 3.0 to 8.5. Experiments produced that the erosion and deposition actions decrease as Ø/L increases. The minimal erosion and deposition are detected at (Ø/L =8.5). The optimum place of a bridge circular support along the bend is defined. The results of Nays2DH Model are compared with tests. RSQ for the modelled statuses is 88.329% and the correlation factor between simulations and the gauged depths is 93.98%.


INTRODUCTION
Curved open canal are subjected to complicated erosion and deposition processes. Many researches made an effort to solve erosion problem in open canals [1][2][3][4][5]. A study presented CFD model to predict erosion process in three different S-bends [6]. The tests were conducted for S-bends of the curvature's radius ratio (Ø/L=1.5). The results showed areas of the maximum erosion along the S-bend. Based on a recent study specified that narrow rivers were subjected to stabilized meanders [7]. A research pointed to that the 135° bend can be treated as a typical regular meander [8]. Studied the lateral movement of alluvial in the Rio Grande [9]. Previous study recommended an empirical formula for the prediction of scouring in bends, see Eq. (1) [10].
(Smax/W) =207-0.19 Ln (Ø/L) (1) Studied the consequence of flow discharge on the capacity of the sediment transport [11]. A study examined the different constraints affect the computations of CCHE2D model [12]. The roughness coefficient of the bed is the mainly factor affecting the flow in open canals. Others provided results of investigations of the bends' scour in Warta River [13]. The scour at bends was calculate by a numerical model. A maximum scour equation along the bend was introduced. Recent research built a 2-D numerical model to detect the transporting of the sediment through open canals [14]. Then other study used two numerical models to detect the scouring problems at Nile River upstream of the new intake of Esmaeilia canal [15]. Two-dimensional horizontal numerical model to detect the morphological changes in rivers of bends [16]. Two experiments were used to calibrate the model. P r e v i o u s s t u d y applied one of the available solvers in iRIC software to simulate velocity field around piers under contraction effect [17]. Results showed, piles of 4.812 comparative extension yielded minimal velocity. In fact, no available researches studied erosion processes around the supports along curved reachs. This paper presents experiments for the erosion and deposition processes along 180° bend through open canal. Experiments are managed with the curvature's radius ratio (Ø/L) varies between 3.0 to 8.5. The optimum location of bridge supports along the bend are defined. Nays2DH Model is applied to simulate the flow and erosion processes. The numerical results are compared with the tests for open canal.

Description of the model and soil
Investigations examined experimentally through open canal flume of a width 22 = 40cm, length = 400cm and depth = 20cm. The flume is existed in the hydraulic lab of engineering collage in Zagazig University. The photos express the parameters of the phenomenon are given as shown in figure (1). Tests are conducted to analysis the erosion and deposition processes along 180° bend through open canal of solid boundaries, see figure (1a). The model boundaries are built from steel sheets fixed to the bed. The natural soil is used along the model. Some gravel boulders are putted at the inlet to stabilize the soil. The soil examination was done. The medium diameter of sand soil P50 = 1900 micron. The sand soil can be treated as a uniform specimen.

Time effect on the testing process
Eight tests are conducted to examine effects of the time on the erosion and deposition processes. The time of the test are (10, 20, 30, 60, 90, 120, 180 and 300 minutes). The relation between the rates of the maximum erosion depth related to the erosion of 300 minutes (Smax/ Smax 5 hours) is plotted against the time as illustrated in figure (2). It can be realized that about 83% of the erosion can be attained by the ending of the first hour. The applied time for the experimental tests is fixed as one hour.

The location effect of the support (ω)
The location effect of the support on the erosion process was detected by 32-tests. It was conducted to examine effects of the locating a support of diameter 3.2cm in the centerline of the open canal at different positions. The details of the model are presented in figure (1). The tests are conducted for the centerline's radius Ø = 50cm, Froude (i.e., Froude number) ≅ 0.3:0.65 and ω (i.e., the angle of location) = 41°, 53°, 70°, 90°, 110°, 127°, 139° and 180°, respectively. erosion processes. The support's diameter = 3.2cm was fixed in the centerline of the open canal (i.e. ω=90°). The tests are conducted for centerline's radius Ø = 50cm, 66cm, 100cm, and 160cm.

Optimum location of the Support
The results of the deposition and erosion phenomenon through open canal with bends at different ω and Ø/L =3.0 are analyzed in the following section. Figure (3a and 3b) presents the examined erosion and deposition at the support itself. It was looked out that, SPmax/Wdown (i.e., the relative erosion depth at support) is the minimum for the status of ω=41°. In contrast, DEPmax/Wdown (i.e., the relative deposition depth at support) is the maximum for the status of ω=41°. Figure (4a and 4b) presents the examined erosion and deposition at the outer bend. It was looked out that, the rate of the depth of erosion Smax/Wdown is the minimum for the statuses of ω=90° and 180°. The rate of the maximum deposition DEmax/Wdown is the minimum for the statuses of ω=90° and 180°. Figure (4c and 4d) presents the examined phenomenon at the inner bend. It was looked out that, Smax/Wdown is the minimum for the statuses of ω=90° and 180°. Moreover, DEmax/Wdown is the minimum for the statuses of ω=70° and 180°.

The Optimum Curvature
The results of the deposition and scour phenomenon in bends of different curvatures are analyzed in the following section. The status of Ø/L=8.5 gives minimum values of Smax/Wdown and DEmax/Wdown at the outer boundary, see figure (5a, and 5b), respectively. Moreover, Ø/L=8.5 gives the minimum values of DEmax/Wdown at the inner boundary, see figure (5d). It was looked out that, Smax/Wdown is the minimum for the status of Ø/L=3 and 8.5, see figure (5c). Figure (6a and 6b) presents the examined erosion and deposition at the support itself for different Ø/L and ω=90°. It was looked out that, SPmax/Wdown and DEPmax/Wdown are the minimum for the status of Ø/L = 8.5.

OVERVIEW OF NAYS2DH MODEL
The numerical two dimensional in plan Nays2DH model was built by Dr. Yasuyuki Shimizu [18]. It is a powerful model used to detect the flow and sediments behavior through open canals. Nays2DH model is a solver imbedded in iRIC software [19]. A research also presented basic equations of the sediment transport and flow used in Nays2DH model [18]. The main formulas of the flow include the continuity and the momentum equations.
Few papers were presented for applications of Nays2DH model. Based on a study discussed 6-solvers embedded in iRIC software [20]. Environmentally assessed water conditions and the sediment action in the reservoir of Ogaki Dam for different operation statuses using Nays2D model [21]. A recent study applied Nays2D model for open canal subjected to water surface fluctuations [22]. The model was compared with observatories of gauged water surface levels.

DESCRIPTION OF THE GENERATED NUMERICAL MODELS
The generation of the Nays2DH Model passes through few steps. At the first, the mesh is generated as illustrated in figure (7a). The mesh is schematized into 10-cells in lateral direction and 70-cells in longitudinal direction. The dimensions of a single cell are 0.02 × 0.02m. The total number of cells is 700. The topographic bed of the model is given as figure (7b).

The case of locating a support
Nays2DH Model was built for the statuses of proposed open canal with a support existed in different locations. The flow properties include followings: the discharge = 0.0024 m 3 /sec, the flow surface is uniform slope of 0.0045. The supports' locations include 4-positions (i.e. ω=41°, ω=53°, ω=70°, and ω= 90°).

CALIBRATION OF THE NUMERICAL MODEL
The results of Nays2DH Model are compared with tests. Figure (8) presents the maximum Pmax/Wdown through the examined reach and for both the measurements and the calculated ones by Nays2DH model against Froude. It was looked out that, there is an acceptance between the measurements and Nays2DH for Froude <0.5. On the other hand, the gap is noticeable between measurements and the calculations for the range Froude > 0.5. General speaking, Nays2DH model gives more scoured depths than the detected in the lab. RSQ for the modeled cases are 88.329% and the correlation factor between the numerical modeling and the gauged = 93.98%.

NUMERICAL RESULTS
The results of Nays2DH Model are presented for the statuses of proposed open canal without supports, see figure (9). Figures (9a, 9b, 9c, 9d, 9e and 9f) display the numerical outcomes of the scour and deposition progression along the studied reaches for the several flow conditions. It is obvious that, the scour depth increases as Froude increases.
The results of Nays2DH Model are presented for the statuses of different positions of support, see figure (10). The figure illustrates that, scouring areas around the support is minimized for the status of ω=41°. In the opposite way, the status of ω=70° gives minimum scouring processes in the studied reach. Figures (11a, 11b, 11c and 11d) display the numerical generation maps for the streamline's distribution for different positions of the support. It is easily seen that, the flow lines become in an equal distribution across the section in the cases 18 ω=70°, and 90°. Figures (12a, 12b, 12c and 12d) display the numerical generation maps for the velocity vectors distribution for different positions of the support. The figure illustrates that, the velocity vectors become in an equal distribution across the section in the cases ω=70° and 90°. (2) Results of Nays2DH Model are comparing with measurements for the status of open canal without any supports. RSQ for the modeled cases are 88.329% and the correlation factor between the numerical modeling and the gauged =93.98%.
(3) The numerical outcomes of Nays2DH Model show that scouring areas around the support is minimized for the status of ω=41°. The status of ω=70° gives minimum scouring processes in the studied reach.