Among the renewable energy sources, hydropower plays an important role by providing approximately 60% of the renewable electricity. Globally, there is a growing installed capacity of renewable energy sources. This, along with the energy policies to reduce greenhouse gas emissions promotes the development of alternative renewable energy sources such as solar and wind power. The penetration of intermittent energy sources seriously impacts the energy balance as well as the stability of the electrical grid. Therefore, it is required to guarantee a smooth integration of this share into the existing power grids. Hydraulic power plants are one of the key components to stabilize the electric grid. As a result, the extended operations and flexibility of hydraulic turbines increase, and hydraulic turbines are subject to unstable flow conditions and unfavorable load fluctuations at off-design operations. A better understanding of off-design and transient effects, particularly in full-scale hydraulic turbines, has the potential to provide new methodologies to predict the sources of load fluctuations on the runner and to mitigate issues associated with them. Such knowledge can increase turbine refurbishment time intervals and avoid structural failures in extreme cases.

This thesis aims to develop methodologies (i.e., experimental and numerical) to assess Kaplan turbines flow conditions and flow effects on the structure under different operational conditions. The work is divided into two parts; an experimental measurement campaign performed on a full-scale Kaplan turbine, Porjus U9, and a numerical investigation of fluid-structure interaction in the corresponding model turbine. In the measurement campaign, several operational conditions ranging from start-up, speed-no-load, steady-state, load variations, emergency shutdown, runaway, and stop were examined. Steady-state and load variation measurements were carried out under on-cam and off-cam conditions. The main objective was to investigate the effect of the operation conditions on the pressure and stain fluctuations on the runner as well as the strain variations on the shaft. This would lead to propose a measurement methodology in which the blade loading can be predicted by strain measurements on the shaft. The pressure and strain measurements on the runner showed that different sources of fluctuations corresponding to a specific operating condition, e.g. part load and start-up, resulted in load fluctuations on the runner blade. The region in the proximity of the runner blade hub was observed as the most critical in terms of high strain value. During a start-up sequence, the strain measurement on the shaft revealed that both guide vane opening, and runner blade’s angle have a great effect on the strain value on the shaft. A correlation between the blade and shaft measurements seems to exist.

The numerical simulations performed on the Porjus U9 model demonstrated that the added inertia and damping were important, whereas the stiffness was negligible. The dimensionless added polar inertia was 23%–27% of the reference value. Added damping significantly contributed to the moment at low excitation frequencies, whereas the inertia became dominant at higher frequencies. Considering the presence of multiple perturbations in the simulations, the added polar inertia could be assumed independent. Whereas, the interaction of the harmonics modified the added damping value, particularly at high perturbation frequencies.