Design and Control of Hybrid Fuel Cell Vehicle Powertrains using Multi-Objective Genetic Algorithms across Diverse Driving Cycles

Abstract

Attainment of environmentally friendly methods of transportation is largely contingent upon the availability and practicality of hydrogen fuel cell vehicles (FCVs), but FCVs currently have a number of limitations in establishing their widespread user base due to their excessive hydrogen consumption, high system costs, and difficulty in regulating energy utilization based on the behavior of system operators under highly variable dynamic loads associated with typical transportation activity of users. Therefore, understanding the linkage between the fuel cell energy system and supplementary energy storage systems in the hybrid FCV powertrain is an important determinant of total performance, longevity, and fuel economy. This thesis addresses the issue of reducing hydrogen consumption while maintaining the performance of the energy system through the development of an optimal energy management strategy. MATLAB/Simulink has been used to develop a complete hybrid FCV simulation model using an ultracapacitor-based energy storage system coupled with a PEMFC stack for the purpose of this research. The development and implementation of a multi-objective genetic algorithm (MOGA) were carried out in the development of the hybrid FCV simulation model to mitigate operational limitations (i.e., power balance, SoC limits) and simultaneously reduce hydrogen consumption and maximize system efficiency. The MOGA optimization process was performed under a population-based, evolutionary framework employing the convergence of more than one objective with predefined operational constraints. The optimization framework takes into account both important sizing variables, like fuel cell configuration and ultracapacitor operating limits, and control parameters. To analyze the performance of the combined approach, we simulated various standardized driving cycles alongside an actual route in Algeria—Ouenza to Annaba—chosen for its accuracy in mimicking the conditions under which the evaluated vehicles would be driven. Simulation outputs showed that there were Pareto-optimal values from our analysis, allowing a reduction of hydrogen usage by 30% vs. baseline and greater improvements in energy efficiency and SoC trajectory stability when compared to the baseline method across every tested driving scenario. As a result, optimized designs used smaller and cheaper fuel cells, without compromising the performance of the vehicle. What makes this research unique is that we utilized this methodology to combine multiple standardized driving cycles with an actual route while optimizing both energy management and component sizing parameters through an MOGA framework. Therefore, we provide a means for developing replicable, regionally adaptable advanced fuel cell vehicle powertrain design solutions.