Hydropower plants in Sweden has a long history dating back to the early 20th century. In the design of these, now old facilities, expected probable flows were based on recorded precipitation data. As a safety valve of the dam various types of spillway can be installed, to either spill water when no power is produced. Or to regulate water levels above and below the dam to prevent or mitigate flooding, and to be able to prevent overtopping of a dam. As the climate is ever changing, and now average temperatures climb, this leads to an increase in both average annual precipitation and likelihood of expected flood events in Sweden. This leads to a case of old infrastructure designed for old conditions, which now may exceed flow discharge capabilities for safe operations. A need for reevaluation of the existing dam fleet is needed to explore if it can face the new conditions brought by increased precipitation. Previous work has been done by Computational Fluid Dynamics (CFD) to compare scale model data of existing dams, recorded when they were designed. Such comparisons show mostly low differences between the CFD simulations and scale model results, in the range of 1-2%. Some show as much as 10%. With CFD now being a standard tool in many industries involving fluid mechanics, there are several guidelines on how to perform CFD with respect to quality. However CFD is an approximate science and the physics needs to be simplified by a number of models with inherent limitations. Hence to gain trust in CFD simulations for dam operators there needs to be validation cases. Validation cases for single outlet spillway setup exist, but dams often do not have geometries as simple as the existing validation cases. Hence, a need for well defined experiments for CFD validation of hydraulic designs.

This thesis aims to provide experimental data suitable for use in evaluating CFD methods for assessing spillway capacity. To this endeavour a purpose built experimental setup was designed to produce experimental data of a quality and complexity not found elsewhere today. The main feature of the experiment is three outlets with the capability of recording the discharge passing through each outlet individually. Other tools used for evaluating flow conditions in the channel include Acoustic Doppler Velocimetry.

The results consists of experimental data gathered for three different variations of the experimental geometry. First a Deep channel flowing past a sharp corner, which produced low velocities in the channel leading up to the outlets. A second variation where the channel floor was raised to induce larger velocities in the flow leading up to the outlets, for increased differences in the flow distribution between the different outlets. As a final variation, the channel width was reduced leading to even higher velocities, and larger differences in both waterlevels at different points in the experimental channel, and in the discharge recorded in the different outlets. The distribution of the flow discharge across the outlets varies with the different geometries, and with the different inflows tested. At low flows differences in distribution between the outlets were negligible, especially with the first channel layout. At higher flows the differences grow, and show clear differences that should be replicable in simulations. Thus showing results that could be used for validation of CFD methods in regards to flow distribution across a spillway with multiple outlets. Additional data to use for validation include ADV data gathered in the channel leading up to the outlets, which documents recirculation zones introduced due to the geometries. As potentially simulations could produce flow distribution results that are correct while not simulating flow behaviour correctly.