Roughness Effect on Velocity Distribution in Selected Reach of Shatt al-Arab River

ABSTRACT  Shatt al-Arab is the only navigational artery in Iraq, extending from the city of Qurna to its mouth in the Arabian Gulf at the city of Al-Fao within the governorate of Basrah for a length of approximately 204 km. Its width ranges from 400 m to 2000 m, and its depth ranges from 8 m to 20 m. The southern part of it, 93 km long from Umm al-Rassas Island to Ras al-Bisha, represents the international border between Iraq and Iran, where the Thalweg line represents the border between the two countries, which is the deepest point in the riverbed (according to the 1975 Algiers Agreement). The western bank (the Iraqi side) within the common border of Shatt al-Arab is subject to continuous erosion, which leads to the shifting of the Thalweg line towards Iraqi territory and thus leads to loss of Iraqi land to Iran. Reducing flow velocity along the Iraqi side can lead to reducing or preventing erosion in the river. Increasing the riverbed roughness will reduce the velocity of flow and then reducing the erosion. This principle was adopted in this study to investigate the effect of increasing roughness in a strip along a reach of the riverbed on the distribution of longitudinal velocity in cross-sections at the rest of the selected reach. A reach of Shatt al-Arab with a length of 2500 m, located 34 km north of Fao City, was selected to represent the study area. This reach was simulated by using numerical modeling CFD solver (fluent) with three different roughnesses for an upstream part of the river bed and the velocities compared with the natural (original) roughness of Shatt al-Arab. The results showed an appreciable effect of the increased bed roughness on the velocity distribution and the maximum velocity location by shifting it to the other side.


INTRODUCTION
Shatt al-Arab is located in southeastern Iraq within the governorate of Basrah and is approximately 204 km long. The final 93 km of the river forms the international border between Iraq and Iran (Al-Fartusi, 2013). The deepest points along the international border between the two countries configure an imaginary line called (Thalweg Line) according to (Algiers Convention 1975). For along years ago, the Shatt al-Arab River stayed suffers from erosion at the western side (the Iraqi side), which caused gradual encroaching of AL-Thalweg line toward the Iraqi territory. That led to losing the Iraqi territory (Ibrahim 2017). One of the methods that can be applied in order to prevent erosion in rivers is reducing the water flow velocity (Ibrahim-Mageed 2014), and in this research, this way can be achieved by increasing the river bed roughness. Because of the difficulties of physical simulation, long time of testing, accuracy results weakness, and scaling problems, a numerical simulation method was used. Shatt al-Arab river problem was numerically simulated by using a commercial CFD solver (FLUENT).

2.DESCRIPTION OF THE STUDY AREA
The study area is a reach of Shatt-al-Arab which is located at 34 km north Al-Fao City which started from the mouth of the Shatt al-Arab to the Arabian Gulf, E=245006, N=3349551, of 2.5 km length between km 60+00 to km 62+500 and 860m width and 11m depth-averaged as shown in Fig.1. An area 500m length and 220m width at the reach upstream was allocated from the riverbed to simulate different roughness.

NUMERICAL SIMULATION
In the current study, different softwares were used to generate a digital model. These softwares were the Civil 3D, Gambit, Space Clam, Hec-Ras, and Ansys Fluent (CFD). The CFD solver FLUENT used in the present study solves the three-dimensional Reynolds-averaged Navier-Stokes equations for incompressible flow. FLUENT solves the governing equations sequentially using the control volume method.
Shatt al-Arab River cross-sections data were provided from the General State of Survey (Ministry of Water Resources). A certain area at upstream of the reach on the Iraqi side was selected to increase the roughness in each case, as shown in Fig.2. Four cross-sections (S1, S2, S3, and S4) were selected at 500 m spacing along the reach in which the first section started at a distance of 250 m upstream of the reach to investigate the lateral behavior of the vertical distribution of the longitudinal velocity. At each section, the velocity profiles at three vertical lines (A, B, and C) located at (1/4, 1/2, and 3/4) of the river width started from the Journal of Engineering Volume 26 August 2020 Number 8 Iraqi side, respectively, as in Fig.3. These lines were used to compare the vertical velocity distribution in each case. Four (4) cases were run in the software. The first case (natural conditions) all riverbed has the same roughness height of (0.05 m equivalent to Manning's n of 0.033, according to Stickler's equation (1). In other cases, the natural conditions were maintained, but the roughness height in the designated area was changed to 0.25, 0.75 and 1.25 m (equivalent to Manning's n of 0.033, 0.04, and 0.043) in the (case 2, case 3, and case 4) respectively. The boundary conditions that were used are the inlet condition (mass flow inlet), the water surface condition (atmospheric pressure), the bed and allocated area condition (stationary and non-slip walls), and outlet conditions (both the pressure at the outlet and the constant water surface elevation) were imposed. Furthermore, the Volume of Fluid (VOF) method and Shear Stress Transport (SST) k-ω turbulent model were used in this simulation.

RESULTS AND DISCUSSION
Four runs were simulated to investigate the effect of changing bed roughness at the upstream part of the river reach on the longitudinal velocity (v) distribution along the rest reach. Each run represents Manning's n of (0.025, 0.033, 0.04 and 0.043). The results arranged as Figures (4 to  15). In each figure, the vertical distribution of the longitudinal velocity (v) for the four runs was represented. Also, for all lines (A, B, and C) in all sections. Fig.4 shows that when Manning's n was increased in allocated area to (0.033, 0.04 and 0.043) the average value of the longitudinal velocity at section S1 vertical line A (S1 A) decreased by (5%, 12%, and 25%) for (n=0.033, 0.04 and 0.043) respectively, compared with the natural case (Manning's n =0.025).    Fig.5 shows that when Manning's n was increased to (0.033, 0.04 and 0.043) the average value of the longitudinal velocity at (S1 B) increased by (5%, 25% and 34%) for (n=0.033, 0.04 and 0.043) respectively.

Figure 5.
Longitudinal Velocity Distribution at (S1 B). Fig.6 shows that when Manning's n was increased to (0.033, 0.04 and 0.043) the average value of the longitudinal velocity at (S1 C) increased by (5%, 18% and 20%) for (n=0.033, 0.04 and 0.043) respectively.       9 shows that when Manning's n was increased to (0.033, 0.04, and 0.043) the average value of the longitudinal velocity at (S2 C) increased by (21%, 49%, and 71%) for (n=0.033, 0.04 and 0.043) respectively, compared with the original case.     11 shows that when Manning's n was increased to (0.033, 0.04 and 0.043) the average value of the longitudinal velocity at (S3 B) increased by (60%, 162%, and 210%) for (n=0.033, 0.04 and 0.043) respectively, compared with the original case.     13 shows that when Manning's n was increased in the allocated area to (0.033, 0.04, and 0.043). The average value of the longitudinal velocity at (S4 A) decreased by (9%, 90%, and 35%) for (n=0.033, 0.04, and 0.043) respectively, compared with the original case (Manning's n =0.025). Still, in the case of n = 0.043, the velocity direction was in the opposite direction of flow.       15 shows that when Manning's n was increased to (0.033, 0.04, and 0.043) the average value of the longitudinal velocity at (S4 C) increased by (15%, 409%, and 645%) for (n=0.033, 0.04 and 0.043) respectively, compared with the original case.   Figure 15. Longitudinal Velocity Distribution at (S4 C).

CONCLUSIONS
The simulation results show that when increasing the roughness at the allocated area, the vertical distribution of the longitudinal velocity decreased at the western side and gradually shifted laterally direction to the eastern side. It was observed that the maximum velocity relocated to the middle of the river width. Furthermore, the vertical distribution, the highest velocity, shifted away as increasing bed roughness in the allocated area. These performing is considered a good indication of altering the bed erosion and shoulder erosion at the western side.