Deep Learning RBFNN MPPT Developmemt for Hybrid Energy Microgrid

The role of a significant Maximum Power Point Technique (MPPT) in augmenting the energy extraction and conversion efficiency of a variable power source such as Photovoltaic (PV) system is indispensable. So, an effective Radial Basis Function Neural Network (RBFNN) MPPT approach, which is based on Gaussian activation function is adopted in this work owing to its exceptional attributes of quick learning and universal approximation. The proposed approach focusses on the working of a microgrid that entails Wind-PV-Battery along with a capacitor bank. The fluctuation in power output caused by varying weather conditions inherent in wind and photovoltaic (PV) primary energy sources can be effectively mitigated by integrating a Battery Energy Storage System (BESS).The inclusion of a capacitor bank provides the essential voltage stability to the microgrid by minimizing the effects of voltage fluctuations in addition to storing of surplus electrical energy. Achieving a consistent and steady power supply from photovoltaics (PV) is accomplished by employing a DC-DC Boost converter alongside the RBFNN MPPT technique. Utilizing a Proportional Integral (PI) controller manages the Pulse Width Modulation (PWM) rectifier, facilitating the establishment of a reliable DC power source from the Doubly Fed Induction Generator (DFIG) within the Wind Energy Conversion System (WECS). A three-phase voltage source inverter ( \(3\varphi\) VSI) is used to convert the stable DC link voltage to AC, and an LC filter is used to reduce harmonics. The grid voltage synchronization is accomplished based on dq theory with the assistance of PI controller. The performance of RBFNN MPPT in heightening the efficiency of PV system is appraised using MATLAB simulation and the RBFNN MPPT technique is estimated to operate with an excellent efficiency of 97%.