Optimized coil design and advanced neural network control for enhanced wireless power transfer in electric vehicles using Taylor-based Firefly and Dove Swarm Optimization

This paper introduces a new approach to boost the efficiency of wireless power transfer (WPT) for electric vehicles (EVs), through advanced coil design optimization and control techniques. Current EV charging methods struggle with slow charging times and limited adaptability. The present approach tackles these issues by optimizing electromagnetic coupling and resonance conditions, leading to improvements in both efficiency and system reliability. The proposed method includes a novel optimization algorithm, the Taylor-based Firefly and Dove Swarm Optimization (F-DSO), to fine-tune coil parameters, alongside a Bessel Filter-enhanced Fuzzy Recurrent Neural Network (FRNN) controller. The F-DSO algorithm adjusts dynamically to enhance coupling and mutual inductance, over an air gap of 135 mm. The FRNN controller maintains stable, accurate battery charging with minimal signal interference. Experimental validation shows the system achieves a peak efficiency of 95%, a state-of-charge (SOC) of 45%, and a steady battery voltage of 25 V. This approach surpasses previous configurations, such as the H-Shape DST, QDQ, and square coil setups. Additionally, the integrated LCC and series-series compensation circuits ensure reliable voltage control and stability under varying loads. These findings set a new standard for WPT, showcasing the potential of the F-DSO algorithm for broader use in the EV sector, offering more efficient, adaptable, and user-friendly charging solutions.