The effectiveness of the PANS modelling approach (Partially-Averaged Navier-Stokes; proposed principally by Girimaji, 2006 and further developed by Basara et al., 2011, 2018) is demonstrated by the computation of various flow configurations relevant to external vehicle aerodynamics, including single isolated cars, but also the quasi-steady 'wind tunnel' and the unsteady 'on-road' overtaking manoeuvre, with the letter process involving moving vehicles. The PANS modelling strategy represents a seamless hybrid RANS/LES (Reynolds-Averaged Navier-Stokes / Large-Eddy Simulation) computational method. The representative length and time scales of the residual turbulence entering the PANS-related equations of motion are determined through a modified RANS model formulation that characterizes the unresolved sub-scale fraction of turbulence. In the current implementation of the PANS method, the underlying RANS model is the near-wall four-equation eddy-viscosity model known as the k-ε-ζ-f model, as proposed by Hanjalic et al. (2004). This model is sensitized appropriately to account for fluctuating turbulence by introducing a grid-spacing-dependent resolution function in the equation governing the length-scale supplying variable (ε). Accordingly, the grid spacing (Δ) is just one among several model parameters used to determine unresolved turbulent scales. The corresponding model formulation suggests a relationship that involves multiple turbulent quantities with a high level of coherence. Consequently, the grid spacing emerges as a less dominant factor in this hybrid LES/RANS model. As a result, this modeling rationale inherently incorporates more physics, enabling the use of coarser spatial and temporal resolutions. Initially, the PANS methodology is scrutinized by simulating a set of geometrically simpler configurations, but characterised by complex flow straining and associated turbulent interactions, including flow separation and swirling effects, as well as mean compression and tumbling motion in a canonical piston-cylinder arrangement. All simulations were performed with the computational code AVL-FIRE, which is based on the second-order accurate cell-centered finite volume method with the governing equations integrated term-by-term over the ‘polyhedral’ control volumes.