Flywheel energy storage system controlled using tube-based deep Koopman model predictive control for wind power smoothing

This paper introduces an approach for wind power smoothing using a flywheel energy storage system (FESS) controlled by a novel tube-based deep Koopman model predictive control (MPC) method. Wind power, despite its potential to reduce carbon emissions, faces significant challenges due to power fluctuations caused by variable wind speeds. To address these fluctuations, FESS is proposed as a result of its rapid charge–discharge response capability. To control the FESS, a deep neural network (DNN) is used to approximate the Koopman operator for system linearization, allowing the application of a linear MPC controller. To enhance robustness, a tube-based MPC approach comprising a nominal MPC and an ancillary MPC is introduced. The stability of the nominal system and the input-to-state stability (ISS) of the controlled actual system are rigorously established. The effectiveness and robustness of the proposed method are demonstrated through simulations and comparisons with PID and conventional MPC method.