Study on the long-term sediment abrasion characteristics of turbines considering geometric gradients

Abstract

To mitigate the risk of prolonged operational downtime caused by runner abrasion in high-altitude hydropower stations subject to sediment-laden flows, accurate long-term abrasion prediction is essential. However, existing models often overlook the progressive deformation of flow passage boundaries, which can significantly influence local flow structures and particle-wall interactions. In this study, a sediment abrasion prediction method is developed by integrating experimental data from rotating-disk slurry abrasion tests with an abrasion simulation framework that incorporates dynamic mesh deformation and an abrasion rate acceleration factor. The approach couples a CFD-based Euler-Lagrange model with geometry-updating algorithms to reflect evolving surface profiles and feedback effects on particle impact behavior. The method was validated against in-situ measurements from a Francis turbine runner operating continuously for 5723 h. Results show that the proposed model, which accounts for boundary evolution, improves long-term abrasion prediction accuracy by 28.71 % compared to traditional fixed-geometry models. The model also captures the shift of maximum abrasion zones induced by localized vortices formed around abrasion pits which is consistent with experimental observations. This research provides a robust and transferable methodology for abrasion assessment and can serve as a technical reference for the optimization of turbine design and sediment management strategies in sediment-prone hydropower systems.