Multiobjective Optimized Smart Charge Controller for Electric Vehicle Applications

Abstract

The continuous deployment of distributed energy sources and the increase in the adoption of electric vehicles (EVs) require smart charging algorithms. The existing EV chargers offer limited flexibility and controllability and do not fully consider factors (such as EV user waiting time and the length of next trip) as well as the potential opportunities and financial benefits from using EVs to support the grid, charge from renewable energy, and deal with the negative impacts of intermittent renewable generation. The lack of adequate smart EV charging may result in high battery degradation, violation of grid control statutory limits, high greenhouse emissions, and high charging cost. In this article, a neuro-fuzzy particle swarm optimization (PSO)-based novel and advanced smart charge controller is proposed, which considers user requirements, energy tariff, grid condition (e.g., voltage or frequency), renewable (photovoltaic) output, and battery state of health. A rule-based fuzzy controller becomes complex as the number of inputs to the controller increases. In addition, it becomes difficult to achieve an optimum operation due to the conflicting nature of control requirements. To optimize the controller response, the PSO technique is proposed to provide a global optimum solution based on a predefined cost function, and to address the implementation complexity, PSO is combined with a neural network. The proposed neuro-fuzzy PSO control algorithm meets EV user requirements, works within technical constraints, and is simple to implement in real time (and requires less processing time). Simulation using MATLAB and experimental results using a dSPACE digital real-time emulator are presented to demonstrate the effectiveness of the proposed controller.