Critical analysis of turbulence models for simulating positive surge waves in open channels in a RANS-VOF setup

The study presents a systematic assessment of 2D RANS-VOF simulation of positive surge wave propagation in open channels using three widely used turbulence models: standard k-epsilon, RNG k-epsilon and the k-omega SST. The numerical simulation results are validated on the basis of rigorous comparison with data obtained from existing physical model studies and in-house experiments. The study finds that the standard k-epsilon model fails to simulate the free surface features of the surge wave precisely. The RNG k-epsilon and the k-omega SST models show better performance in this regard. The standard k-epsilon model is also unable to replicate the physical trends of the measured velocity data. While the overall agreement between the velocity characteristics estimated by the RNG k-epsilon model and experimental data is satisfactory, the model exhibits weakness while capturing the recirculation zone beneath the breaking surge having Froude number 1.6. The k-omega SST model simulates the pertinent velocity characteristics and the relevant flow structures with better accuracy in comparison to the RNG k-epsilon model for both undular and breaking surges. The non-hydrostatic pressure field is well approximated by both the RNG k-epsilon and the k-omega SST models. A comparison of the free surface levels computed by SGN models reveals that the limitation of the depth averaged non-hydrostatic models in capturing the breaking of waves may be overcome by the proposed RANS-VOF approach. Comparison with 2D LES results obtained from existing studies on positive surge modelling indicates that the proposed 2D RANS-VOF model is capable of producing results almost at par with the LES models.