Combined effects of passing clouds and wind gusts on an interactive wind-PV system with battery storage using neural networks

This article analyzes the combined effects of random clouds and wind gusts on a grid-linked rooftop wind-photovoltaic system (WPS) with battery storage. For short period intervals, the wind speed and the solar radiation are considered as the sole sources of power whose variations influence the system variables. The stochastic nature of gust), winds creates a quasi-permanent transitory environment for the wind power subsystem, where the wind generator voltage and power undergo steep changes in short periods of time. Likewise, passing clouds affect the variables of the photovoltaics. Since both subsystems have different dynamics, their respective responses are expected to impact differently the overall system behavior. The system responses are predicted with artificial neural networks (ANN). Unlike conventional modeling, the ANN simulation does not require the knowledge of System parameters. This allows circumventing two major problems: parameter uncertainty and system non-linearity, which could be high for large systems with multiple components placed in unpredictable environment characterized by turbulent movement of wind and random cloud passage. The dynamics of the system are captured and stored in the connection weights between two consecutive layers. The dispatchability of the battery-supported WPS as well as its stability and reliability during gusts and cloud passage are also discussed.