Effects of inlet boundary conditions on spiral casing simulation

Abstract: The Hölleforsen hydropower plant is situated on the river Indalsälven in Sweden and consists of three Kaplan turbines units with a total installed capacity of 50 MW at the operational head of 27 m. The main task of the current research work is to simulate the flow through the penstock, spiral casing and distributor (guide vanes and stay vanes) of the Hölleforsen model, well known as the Turbine-99 test case. Numerical simulations are performed on the model with 1:11 scale of the prototype turbine, a runner diameter of 500 mm, a runner speed of 595 rpm and a volume flow rate of 0.522 m3/s at the head of 4.5 m. Flow in the penstock and spiral casings are very complex with several flow phenomena appearing simultaneously: such as separation, turbulence and unsteadiness. This is due to the viscosity, boundary conditions and complexity in the geometry. To gain some information on the flow characteristics, numerical simulation of the complex three-dimensional turbulent viscous flow through the penstock, spiral casing and distributor of the turbine is performed. Conservation mass and momentum equations of the flow are analysed using finite volume method with the software ANSYS CFX10.0. The objectives are to study the effects of the inlet boundary conditions and also to achieve the boundary conditions for subsequent simulations including the runner and ultimately the entire system. The flow in hydropower plant is turbulent and highly unsteady. Therefore analysing the effect of the upstream geometry and different turbulence model are of great interest to analyse how the boundary conditions affect the runner and draft tube flow. Two types of turbulence models, standard k- ε: and SST k-ω: Based turbulence model are applied to study the flow characteristics. Comparisons are made between the numerical simulations with and without the penstock and the available experimental results in the spiral casing. The numerical results are found to have better agreement with the experimental results when the penstock is included in the simulation, as expected. Therefore, detailed inlet boundary conditions are necessary to simulate accurately the spiral casing flows if the penstock is not included in the simulation. On top of that the numerical simulation seems to show little sensitivity to the turbulence model.