Numerical investigation of the flow in a swirl generator, using OpenFOAM

Transcription

1 Applied Mechanics Numerical investigation of the flow in a swirl generator, using OpenFOAM Oscar Bergman 07 June /35

2 Applied Mechanics Outline Outline Introduction Purpose and goal Experimental rig and measurements Numerical setup Results Conclusions Future work 2 /35

3 Introduction Introduction 3 /35

4 Purpose and goal Purpose and goal Conduct steady-state and unsteady 3D simulations on a swirl generator. Compare with measurements and theoretical design data from previous studies. Provide results for helping future studies in solving the problem with precessing vortex ropes in water turbines. Provide a tutorial of the case to the OpenFOAM community. 4 /35

5 The swirl generator Experimental rig 5 /35

6 The swirl generator Experimental rig Strout - Holds up the nozzle 6 /35

7 The swirl generator Experimental rig Strout Guide vanes - creates a swirling profile 7/35

8 The swirl generator Experimental rig Strout Guide vanes Free runner - redistributes the total pressure - rotates freely 8 /35

9 The swirl generator Experimental rig Strout Guide vanes Free runner Draft tube - plexiglass walls - windows for LDV measurements 9 /35

10 The swirl generator Experimental rig Strout Guide vanes Free runner Draft tube Theoretical design profile Cross-section 1 & 2 10 /35

11 The swirl generator Measurements Total volume flow: 30 l/s 11 /35

12 The swirl generator Measurements Total volume flow: 30 l/s Runner rotating at 870 rpm 12 /35

13 The swirl generator Measurements Total volume flow: 30 l/s Runner rotating at 870 rpm Laser Doppler Velocimetry - measuring the meridional and tangential velocities 13 /35

14 The swirl generator Measurements Total volume flow: 30 l/s Runner rotating at 870 rpm Laser Doppler Velocimetry Pressure transducers - measuring the static pressure at the wall 14 /35

15 Numerical setup Numerical setup The grid million hexahedral cells - Coupled parts by General Grid interfaces (GGI) 15 /35

16 Numerical setup Numerical setup The grid Solvers - For steady-state:» SIMPLE pressure corrector» Rotation through different frames of reference 16 /35

17 Numerical setup Numerical setup The grid Solvers - For steady-state:» SIMPLE pressure corrector» Rotation through different frames of reference - For unsteady:» PISO pressure corrector» Real rotation with a sliding grid at the interface 17 /35

18 Numerical setup Numerical setup The grid Solvers - For steady-state:» SIMPLE pressure corrector» Rotation through different frames of reference - For unsteady:» PISO pressure corrector» Real rotation with a sliding grid at the interface Turbulence model: standard k-ε model with wall-functions 18 /35

19 Numerical setup Numerical setup The grid Solvers Boundary conditions - Velocities and turbulence: Homogenous Neumann at outlet - Pressure: Zero mean pressure at outlet and homogenous Neumann at all other boundaries 19 /35

20 Numerical setup Numerical setup The grid Solvers Boundary conditions Convection scheme 1 st order upwind at startup 2 nd order linear upwind when stable 20 /35

21 Unsteady results Unsteady results Rotational speeds rpm (runner rotates freely) rpm (runner rotates freely according to fluent simulations) rpm (linearly interpolated) 21 /35

22 Unsteady results Unsteady results Rotational speeds Initial boundary condition Results from the steady-state simulations 22 /35

23 Design profiles Meridional velocity Tangential velocity Unsteady results 23 /35

24 Comparison with LDV at 870 rpm Unsteady results Survey axis 0 Survey axis 1 Survey axis 2 24 /35

25 Comparison with LDV at 920 rpm Unsteady results Survey axis 0 Survey axis 1 Survey axis 2 25 /35

26 Comparison with LDV at 890 rpm Unsteady results Survey axis 0 Survey axis 1 Survey axis 2 26 /35

27 Moment acting on the runner 870 rpm 920 rpm Unsteady results 890 rpm 27 /35

28 SIMPLE based solver vs. PISO based solver Unsteady results Survey axis 0 Survey axis 1 Survey axis 2 28 /35

29 SIMPLE based solver compared to measurements Survey axis 0 Survey axis 1 Unsteady results Survey axis 2 29 /35

30 Pressure at MG0 (at the throat of the draft tube) 870 rpm 890 rpm 920 rpm 30 /35

31 Fourier analysis of results from 920 rpm 31 /35

32 870 rpm 920 rpm Unsteady results 32 /35

33 Conclusions Conclusions Unsteady simulations accurately predicts the flow 920 rpm was corresponding most with the measurements 870 rpm was corresponding most with the theoretical design profiles The inclusion of all parts of the swirl generator have added more frequencies to the flow. Moment on the runner: Rotational speed Moment on runner 920 rpm Nm 870 rpm 0.23 Nm 890 rpm 0.08 Nm 33 /35

34 Future work Future work Further investigation of the SIMPLE based solver Other turbulence model such as LES or DES Customize a solver for adjusting the rotational speed in accordance to the moment acting on the runner 34 /35

35 Acknowledgement Acknowledgements I would like to give my thanks to: Department of Applied Mechanics Supervisors Håkan Nilsson and Olivier Petit 35 /35