Hydropower operation and lifetime analysis

Major questions in the hydropower industry are how to safely operate the turbines during the transients that are needed due to the continuously changing request for electric power, which effect this has on the hydropower plants, which costs it is associated with, how to plan for efficient additional maintainance due to the new circumstances, and which lifetime that is expected for different components. In this project two different approaches will be used to try and provide answers to these questions.

Hydropower operation needs to adapt to changes, both in production patterns induced by changes in the services hydropower provide to the energy system and potential impacts caused by changes in climate, while keeping its availability and safety at a reasonably low cost. With increased knowledge it is possible to adapt the operating sequences of hydropower units to maintain availability and safety, to predict actual costs of providing these auxiliary services, and to plan maintainance and estimate remaining lifetime. To achieve this, this project will apply two approaches (named ML and CFD), both with specific aims:

• ML (Machine Learning in hydropower stations): To study operation of existing hydropower plants using machine learning techniques, aiming to develop methods for planned operation with maximum liftetime of the facilities, consequence predictions of different scenarios, identification of anomalies, planned maintainance, and cost of operation.
• CFD (Computational Fluid Dynamics of off-design and transient forces): To numercially study how the flow in water turbines during different operation sequences influences the lifetime of the machines, aiming at guidance for avoiding operation that reduces the lifetime.