Stochastic power management strategy for hybrid energy storage systems to enhance large scale wind energy integration

Abstract The strong variability of renewable energy sources (RES) often hinders their integration in power systems. Hybrid energy storage systems (HESS), based on complementary storage technologies, enable high RES penetration towards modern and sustainable power generation, improving energy systems performances and stability, while reducing CO2 emission. This paper introduces a novel power management strategy for a HESS consisting of a flywheel and a LiFePO4 battery coupled to a 2 MW wind turbine operating in interconnected mode. The power management strategy is based on the simultaneous perturbation stochastic approximation (SPSA) principle and targets a smoother power profile at the point of interface to the grid and, at the same time, a reduced solicitation of the battery. The underlying algorithm falls within the gradient-based optimization category, being able to pursue the envisaged goals without requiring a detailed model of the objective function. The main and novel contribution of this research aims to extend the SPSA recognized advantages, demonstrated in control applications, in the field of real-time HESS power management. Real datasets are employed to size an economic storage section and define representative simulation scenarios in order to validate the suitability of the proposed approach. Simulations are performed over one day timeframe in Matlab/Simulink for the most representative days extracted from the wind turbine yearly generation profile, employing a 1 s timestep. Results obtained prove that the proposed strategy ensures a substantially reduction of the power profile fluctuations at the point of interface to the grid, by more than 80% compared to the wind profile. Moreover, a power ramp mitigation of 65% on average towards the battery if compared to the flywheel.