Improved frequency control techniques for PV and wind integrated power system including EV charging station

Abstract

Renewable Energy Sources are volatile and stochastic in nature; newlinetherefore, modern power networks face many practical difficulties as newlinerenewable energy sources are popularly used. The frequency stability of newlinemodern power systems has been considerably harmed by fast and irregular newlinepower fluctuations produced by non-conventional power production sources newlineand flexible loads. The foremost goal of power system frequency control is to newlinemake sure that generation and demand are balanced at all times. In reality, newlineobtaining accurate estimates of the imbalance of power in both transmission newlineand distribution systems is challenging, especially when renewable energy newlinepenetration is high. Electric vehicles have become a feasible tool to reduce the newlineoccasional impact of renewable energy sources engaged in frequency newlineregulation, primarily because of vehicle-to-grid technologies and the quick newlineoutput power management of EV batteries. newlineThe rapid response of EVs increases the effectiveness of the LFC newlinesystem significantly. This proposed research uses a Long Short Term Memory - Recurrent Neural Network (LSTM-RNN) based deep learning strategy used newlineto identify active power fluctuations in real time. The most significant feature newlineof memorizing past historical load fluctuation and pattern variations is stored newlinein the memory structure. The system performance is enhanced by automatic newlinedynamic variation of system parameters concerning non-linear load newlinevariations. The proposed method has more advantages than the conventional newlinemethod of PID tuning and fuzzy controller since these methods have the newlinedrawbacks of more complex system design, and computational time, and less newlineadaptive to the dynamic non-linear system. newline