Neuro-Adaptive Predictive Control of Flywheel Energy Storage for Hybrid Power Systems

In this paper, a non-linear neuro-adaptive step-ahead predictive control (NASPC) based on neural networks is presented for a low-rated flywheel energy storage (FES) to ensure the frequency regulation of a hybrid multi-area power system. FES energy-level constraints are construed by evolving a control-oriented second-order structure that interprets hardware constraints of flywheel rotor speed into transformed power constraints. The competency of the proposed NASPC framework is validated by the time-domain simulations of a three-area power system with incoming portions of solar energy.