Numerical Simulation Of Hydraulic Characteristics Of High Discharge Spillway Tunnel

The Jinping-I dam is one of the China's biggest hydroelectric power plant projects with a discharge tunnel.As high discharge is present,discharge tunnel operates with the risk of occurrence of cavitation damage.This degree project uses Computational Fluid Modelling(CFD)as a tool in ANSYS 14.0 for modelling and uses the velocity distribution,the static pressure distribution,and the modern cavitation theory to find the cavitation risk.The main content of the dissertation and achievements are as follows:1.CFD modelling of the whole tunnel spillway is done to find the possibility of cavitation damage.As the remedy to cavitation damage,aeration is recommended as it is cheapest and most effective solution for flows.At critical locations with low pressure where cavitation may occur,aerators are used to supply air.In total,4 different aerators are recommended to prevent from cavitation damage.All the aerators consist of deflector and offset and are of same type as used in experimental study.2.The numerical modeling is done using Reynolds-average Navier-Stokes equations.The realizable k-? model is selected for viscous model.VOF technique is used for air water phase intersection for the CFD modelling.The modelling is done as unsteady flow for three different reservoir water levels at 1880 m,1865m and 1850 respectively with two different gate opening conditions of full and half open.GAMBIT software is used for meshing.Velocity distribution,static pressure distribution,water cavity length and cavitation number required to find the cavitation risks are obtained.3.The results obtained from the CFD modeling using the VOF technique is verified using the experimental data.For full flow condition at water level 1880 m,the error of 3.10%,6.16%,5.05% and 4.80% were found for velocity distribution,pressure distribution,water surface profile and aeration cavity length respectively.For other cases,all the pressures,velocities,water surface data and aeration length data simulated from numerical simulation were found to be within permissible error.