Efficiency of CCHE2D numerical model in flow and sediment modeling (case study: Karun River)

 Introduction
Flow hydraulics is considered as the most important factor in the morphological changes of the river by influencing the sediment transport and the shape of the waterway (Adamsson et al, 2003). Accordingly, the hydraulic simulation is very critical in all river engineering projects. In order to model the river flow, many and diverse models with different hypotheses and numerical methods are developed and used. Due to the complexity of hydraulic and sedimentary phenomena in rivers, it is not possible to solve the governing equations by analytical methods, so numerical methods are usually used. In mathematical models, due to the fact that the real size of the river is modeled, there is no scale limitation, although in some cases, fine networks cause the duration of model execution to be longer(Liu et al, 2010). The application and accuracy of mathematical models depends on the definition of the equations governing the phenomenon and the numerical methods used. The aim of the current study is to simulate the hydraulic flow and sediment with CCHE2D software in the Karun River in the Molasani to Farsiat hydrometric stations. In this study, the effect of meshing dimensions and different turbulence models on the accuracy of flow hydrodynamic modeling as well as different sediment equations on river bed changes using CCHE2D model has been investigated.
 
Methodology
The case study is about 110 km and the three hydrometric stations are located on this reach of the Karun River. Molasani and Farsiat stations were considered as upstream and downstream boundary conditions, and Ahvaz station was used as a control station to validate the model. The information used in this research includes information on the elevation of the Karun River, the flow rate and the water level in three hydrometric stations were obtained from the Khuzestan Water & Power Authority (KWPA). The CCHE2D model was developed in 1997 at the Center for Computational Science and Engineering under the supervision of the University of Mississippi, USA. This model is an analytical system for two-dimensional, unsteady and turbulent river flows, sediment transport and water quality assessment. The model is designed with the purpose of application in the areas related to the prediction of the river bed and erosion of its banks, meander movement and water quality. Numerical models in the field of two-dimensional hydraulic simulation of flow and sediment usually need extensive and complete input information. The reason for this is the dependence of the turbulent flow process in natural channels on various factors. The accuracy of the results obtained from the execution of the model will be directly related to the quality, quantity and accuracy of the input data. To simulate the flow in CCHE2D, it is necessary to determine the duration of the simulation and time steps in performing the numerical calculation process. In this research, different time steps has been used in different stages, and the time steps in most of the simulation stages were considered equal to 150 seconds, and in some situations, smaller time steps have been used in order to solve the instability of the model.
 
Results and discussion
In order to check and analyze the model in relation to the mesh size and choose its optimal size, modeling was performed for three different sizes of 150,000, 270,000 and 350,000 mesh nodes. The results obtained for the mesh size of 150,000 nodes show a significant difference with the sizes of 270,000 and 350,000 nodes. This difference is occurred due to the fact that at the beginning and end of the hydrograph, the water level in the network is low and the number of calculation nodes that are located in the dry area is more than the nodes that are located in the wetter area. This problem is partially solved by increasing the flow rate in the hydrograph and increasing the water level. Also, in order to choose the appropriate flow turbulence model for three modes of k-ε model, Parabolic Eddy Viscosity and Mixing Length, simulation was done and the observed water level difference was compared with the measured results at Ahvaz hydrometric station. Based on the obtained results, the k-ε turbulence model has the best match with the observed values ​​with an error of 2% and was chosen as the base model for further modeling. It is worth noting that most of the conducted studies also show that the k-ε model is more capable of modeling the flow in rivers and provides more accurate results. In order to implement the sediment model, the suspended sediment rating curve of Ahvaz station was used as total load sediments. Based on the obtained results, the Ackers - White sediment transport model with a coefficient of 0.1 adaptation factor shows a better agreement with the observational results.
 
Conclusions
The CCHE2D simulations show that the eddy viscosity distribution in the meanders is more intense, which can have a direct effect on the erosion process of the outer bends. Based on the investigations, the maximum erosion and sedimentation took place upstream of the Ahvaz hydrometric station and in the meandering areas of the river. The intensity of erosion in the downstream area of ​​Ahvaz station has decreased significantly. One of the most important reasons for this issue is the low slope of the river bed downstream of the river. In general, the process of modeling, distribution and pattern of flow and sediment parameters by CCHE2D model is reasonable and has acceptable accuracy, and this model can be used for other processes such as changes in the river morphology and displacement of meanders, flow analysis and erosion at the structures.