CFD-Validated Kriging Surrogate for Multi Objective Aerodynamic Optimization of a Go-Kart Drag–Lift Pareto Front under Pitching Moment Constraint

Abstract

Go-kart aerodynamics involves a trade-off between drag reduction, lift control, and longitudinal stability, yet most previous studies do not explicitly constrain pitching moment during design optimization. This study proposes a CFD-validated Gaussian Process Regression (Kriging) surrogate for multi objective aerodynamic optimization of go-kart bodywork using ground clearance, nose angle, and operating speed as design variables. A 25-point maximin Latin hypercube design was evaluated using steady RANS CFD, and the surrogate was used to construct a drag-lift Pareto front under an uncertainty-aware pitching moment constraint at 65 km/h. The recommended knee point design (h = 40 mm, α = 24.79°) achieved Cd = 0.771, Cl = 0.157, and Cm.pitch = 0.171. Compared with the baseline, it reduced lift and pitching moment by about 35% each, with a 4.6% drag penalty. Additional CFD checks at 42 and 89 km/h confirmed that the selected design remained feasible and stable across the operating range. These results demonstrate that surrogate-assisted, stability-constrained optimization can identify practically viable go-kart configurations that improve aerodynamic stability without a large drag penalty, providing a useful framework for bodywork tuning within the studied design range.