Intelligent learning approach for transition control and protection of solar PV integrated electric vehicle charging station

Recently, the control of distributed generations (DGs) with electric vehicles (EVs) parking lot charging systems has received much attention for supporting emergency local loads during grid outages. To achieve optimal transition control of grid-connected (GC) and standalone (SA) modes, an intelligent learning-based algorithm is implemented for optimal handling of the disturbance in the system by injecting a fault condition over certain cycles during a weak/strong grid operation. Here, a fuzzy Q-learning (FQL) based algorithm is implemented for detecting and monitoring the distributed grid condition during different scenarios and a seamless transitioning control in GC and SA modes of operation. In this context, the implementation of designed FQL algorithm enables dynamic adjustments of control signals in response to real-time uncertainties in two major applications (grid synchronization and islanding detection mode). The FQL algorithm focuses on real-time decision-making in response to sudden fault scenarios, addressing the problem under various modes of operation. To validate the proposed FQL controller accuracy, a real-time hardware-in-loop (HIL) simulation and their experimental validation is carried out of a 16kWp grid integrated PV sub-system with different load profiles. The performance evaluation of the FQL controller is investigated under linear and nonlinear local load scenarios.