Optimization of a highway infrastructure design through a combined computational fluid dynamics and evolutionary algorithm framework

Design of highway infrastructure significantly influences the performance and fuel consumption of vehicles. However, standard design practice heavily relies on empirical knowledge, based on difficult-to-generalize and cost-prohibitive experiments. Here, we propose a global optimization framework based on computational fluid dynamics (CFD) and evolutionary algorithms to identify the optimal design of a highway infrastructure that minimizes aerodynamic drag on moving vehicles. Although the framework is broadly-applicable, we investigate the design using an actual section of highway with vertical side walls. The highway's geometry was captured by terrestrial laser scanning and is used to generate the computational domain for three-dimensional CFD. The drag on the vehicles is estimated through CFD using a k-$\varepsilon $ turbulence model. Reduction in the aerodynamic drag is attained by enhancing the interaction between the vehicles and the vertical side walls through the addition of solid slabs to the walls. The optimization process is performed through an evolutionary algorithm, which iteratively evolves the size and positioning of the added slabs toward an optimal design. Our results demonstrate that the proposed computational framework can inform future highway infrastructure designs.