The European Green Deal is prescribing a very ambitious rise in the annual production of renewable energy sources, and grid regulation services are increasingly important in the course of a massive integration of intermittent renewable energy sources. Although hydropower effectively regulates power, the growing demand for flexibility in the coming decades means that hydroelectric technologies must be adaptable, often operating in conditions for which they were not originally designed to ensure availability and system reliability. This also includes a significant increase of the number of start-ups per year. To address these challenges, this paper suggests a two-fold forecasting method to predict damage due to fatigue in the runner blades of Francis turbines. Firstly, stresses during transient operations are predicted using a quasi-steady approach that concatenates measurements from steady-state operations through Voronoi tessellation. Following that, a neural network is trained on an existing dataset of recorded blade stresses during transient operations. The effectiveness of both approaches is validated against measurements in a reduced scale model of a Francis turbine. The comparative analysis highlights the strengths and weaknesses of each method, offering a detailed understanding of dynamic damage accumulation during start-up sequences compared to steady-state operations.