A Spline-Based Modeling Algorithm for Application to Aerodynamic Shape Optimization Based on CFD Analysis

In early phases of conceptual design stages for developing a new car in the modern automobile industry, the lack of systematic methodology to efficiently converge to an agreement between the aesthetics and aerodynamic performance tremendously increases budget and time. During these procedures, one of the most important tasks is to create geometric information which is versatilely morphable upon the demands of both of stylists and engineers. In this perspective, this paper proposes a Spline-based Modeling Algorithm (SMA) to implement into performing aerodynamic design optimization research based on CFD analysis. Once a 3-perspective schematic of a car is given, SMA regresses the backbone boundary lines by using optimum polynomial interpolation methods with the best goodness of fit, eventually reconstructing the 3D shape by linearly interpolating from the extracted boundaries minimizing loss of important geometric features. As a preliminary study, an aerodynamic shape optimization based on CFD analysis is conducted using ANSYS fluid dynamics solutions. To maximize the performance, we employed state-of-art design methodologies such as design-of-experiments (DOE) and surrogate modeling techniques that incorporate a series of procedures from creating geometries to conducting CFD simulations. The SMA is programmatically integrated with the CFD procedures to automatically morph and create corresponding automobile external shapes. For validation and verification, three notional vehicles are explored by investigating aerodynamic performance. Not only does the baseline geometries obtained from SMA show well-matched drag coefficients (C_D) but also the optimization study results in a significant amount of C_D reduction.