Sediment Erosion Prediction for a Francis Turbine Based on Liquid-Solid Flow Simulation Using Modified PANS

Abstract

A challenge in the prediction of the sediment erosion is the proper estimation of the motion and velocity of the solid particles, where Reynolds average Navier-Stokes (RANS) methods show limited resolution to determine the motion of solid phase affected by flow fluctuations. The present study adopts a modified partially averaged-Navier Stokes (PANS) method to analyse the sediment erosion prediction for Francis turbines. Numerical simulations were carried out to obtain liquid-solid two-phase flow information in entire flow passage of a Francis turbine using Eulerian-Lagrangrian approach. The hydraulic performance such as efficiency and discharge of the turbine achieved experimentally, are used to validate the present simulation method. The results show that the modified PANS model can improve the prediction accuracy and the smallest unresolved-to-total ratio of turbulence kinetic energy, fk, decided with the consideration of the difference between local average grid size and smallest grid size shows a slight accuracy improvement. Based on the two-phase flow field, sediment erosion was predicted in stay vane, guide vane and runner using a semi-empirical equation obtained from an erosion experiment of liquid-solid flow. It is noted that higher physical resolution captured by the turbulence model causes a diminution of the sediment erosion predicted. Further, the numerical simulation reveals that sediment erosion in stay vane is lower than guide vane and runner, whereas the highest values of the erosion intensity occurs in the runner. The sediment erosion due to fine solid particles in the turbine is mainly resulted from cutting. However, high sediment erosion due to deformation is also produced at the leading edges of stay vane and guide vane.

A Case Study: Sediment Erosion in Francis Turbines Operated at the San Francisco Hydropower Plant in Ecuador

“…A setup used previously in [32] was used for the solid phase, where the injection was applied at the domain inlet and fully elastic collision was assumed at the walls. An analysis of the adequate number of particle injections was also carried out to generate a statistically meaningful sampling.…”