Artificial neural network controller based cleaner battery-less fuel cell vehicle with EF2 resonant DC-DC converter

Abstract

The evolution of an eco-friendly, sustainable and efficient energy conversion system is becoming inevitable for automotive manufactures and consumers worldwide in the current scenario, especially in highly polluted areas. The use of Fuel Cells (FC) in automotive applications is a lucrative option in response to the critical need for cleaner energy technology. The net voltage output for a single FC is 0.6 to 0.8 V and these cells are arranged in series to form an FC stack. To compensate for FC stack voltage reduction during acceleration of vehicle requires additional battery or ultra-capacitor, which increases the overall cost of the vehicle. An 8 KW FC stack along with Elongation Factor 2 (EF2) resonant converter fed 10 hp Polymer Electrolyte Membrane Fuel Cell (PMSM) drive for powering two-wheelers is proposed which improves the dynamic performance and vehicle reliability. The proposed system reduces the overall size and cost of the FC stack compared to the existing system and regulates the voltage fed to the PMSM drive under different working conditions without additional battery. The proposed drive’s closed-loop speed control is modeled with a State Space Averaged (SSA) dynamic EF2 converter model and a PI controller, followed by a Levenberg–Marquardt algorithm-based Artificial Neural Network (LM-ANN) controller. The performance of the drive is evaluated for various conditions such as starting, climbing, normal running, and braking condition. Comparison of simulation results indicates that EF2 resonant converter operating at greater switching frequency provides high voltage gain at lower duty ratios and LM-ANN based controller seems to be better to provide smooth operation with reduced torque ripple.