Experimental Investigations of Sediment Erosion in Francis Turbine Using Non-Recirculating Sediment Test Rig

Abstract

Water-suspended sediments wear down hydro turbine components by erosion, reducing their lifetime. Nonetheless, even heavily sediment-loaded rivers are a valuable renewable and clean energy resource. Forecasting sediment erosion and optimizing the hydraulic turbine design for extended durability by numerical flow simulations became imperative to operate the hydropower plant safely and economically in sediment-laden environments. This strategy is challenged by the unknown model parameters, that is, the lack of reliable validation data. Therefore, in this study, a 2 kW Francis-type model turbine is tested in a non-recirculating sediment-laden test facility. The Francis turbine runner and the guide vanes are coated with four different colors to visualize locations of surface degradation qualitatively due to erosion during operation in representative sediment concentrations. The turbine is tested at two different operating conditions OP₁ and OP₂ for visualizing the locations of erosion. For operating condition OP₁, with low rotational speed, erosion is primarily observed at the leading and trailing edge of the suction and pressure side, respectively. The Francis turbine runner is particularly eroded in the transition between the hub and the blades on the suction side towards the trailing edge. Meanwhile, for operating condition OP₂, with higher rotational speed, the trailing edge of the pressure side of the blade and the region of the shroud close to the trailing edge of the blades are found to be vulnerable to erosion. Test for the material erosion of the runner is conducted at OP₃ conditions with high sediment concentration over 45 h, reporting the weight loss in intervals of 15 h. It is observed that the cumulative erosion rate of the Francis runner made of brass material is 0.016 mg g⁻¹ h⁻¹ after 45 h of operation. Similarly, the percentage loss of the runner with respect to the hours of operation is calculated to be 0.0065%, 0.04%, and 0.07% for 15, 30, and 45 h of operation, respectively. This data set can also be useful for qualitative and quantitative validation of computational fluid dynamic simulations for erosion prediction.