Research on Thermal-Fluid Optimization of Racing Vehicle Powertrain Based on CFD and Data-Driven Approaches

Liao Lingfeng(Oxford Brookes University)

Abstract

The main causes of racing power are high power density and compact structure, which can easily face the risk of thermal runaway under extreme working conditions. Traditional methods that rely solely on CFD (Computational Fluid Dynamics) simulation or empirical design have limitations such as high computational costs and insufficient dynamic response. This study proposes an optimization framework that combines CFD and data-driven methods to obtain multiple physical field details of the flow temperature field through CFD. By combining machine learning, a high-precision surrogate model is constructed to achieve rapid prediction and global optimization of thermal flow parameters. The innovation lies in building a layered optimization strategy that integrates data and physics, significantly reducing computational resource consumption while improving dynamic adaptability. On a theoretical level, it reveals the mechanism by which interdisciplinary integration enhances thermal management efficiency; At the engineering level, it provides an efficient and robust optimization method for the design of racing thermal systems, which has important application value.

Keywords

CFD；Data driven; Racing powertrain; Thermal flow optimization

References

Lu Yuqi, Wang Zhiqiang Numerical simulation of dynamic process of high-speed axial compressor surge based on CFD method [J]. Propulsion Technology, 2025, 46 (8): 38-48. DOI: 10.3724/1001-4055.2409022

Wang Jie, Ma Yongmei, Liu Xintian, etc Simulation and experimental verification of powertrain for Formula E racing cars [J]. Automotive Engineer, 2021 (6): 21-24,42

Niu Bo Optimization and Mechanics Research of Aerodynamic Package for College Formula Racing [D]. Shanxi: Taiyuan University of Science and Technology, 2024

Liu Pan, Hu Ji, Jiang Yunce Design and Calculation of FSC Formula Racing Brake System Assembly [J]. Technological Innovation and Application, 2015 (26): 71-71

DOI: http://dx.doi.org/10.26549/met.v9i1.36161