Development of an Intelligent Passive Device Generator for Road Vehicle Applications

Abstract

ABSTRACT Flow control has a tremendous technological and economic impact, such as aerodynamic drag reduction on road vehicles which translates directly into fuel savings, with a consequent reduction in greenhouse gas emissions and operating costs. In recent years, machine learning has also been used to develop new approaches to flow control in place of more laborious methods, such as parametric studies, to find optimal parameters with few exceptions. This paper proposes an intelligent passive device generator (IPDG) that combines computational fluid dynamics (CFD) and genetic algorithm, more specifically, the Non-dominated Sorting Genetic Algorithm II (NSGA II). The IPDG is not application specific and can be applied to generate various devices in the given design space. In particular, it creates three-dimensional passive flow control devices with unique shapes that are aerodynamically efficient in terms of the cost function (i.e., aerodynamic drag and lift). In this paper, the IPDG is demonstrated using a rear flap and an underbody diffuser as passive devices. The three-dimensional Reynolds-averaged Navier-stokes (RANS) equations were used to solve the problem. Relative to the baseline, the IPDG generated flap-only, and diffuser-only provide drag reductions of 6.3% and 5.4%, respectively, whereas the flap-diffuser combination provides a drag reduction of 7.4%. Furthermore, the increase in the downforce is significant from 624.4% in flap-only to 4930% and 4595% in the diffuser and flap-diffuser combination. The proposed method has the potential to evolve into a universal passive device generator with the integration of machine learning.

Keywords

• machine learning
• flow control
• shape optimization
• genetic algorithm
• drag reduction