This study investigates differential-equation-based formulations for computing wall-distance functions in Computational Fluid Dynamics (CFD). The wall distance directly influences turbulence modeling by controlling near-wall damping functions and blending behavior, and is particularly critical for industrial applications, for example in maritime contexts involving complex ship geometries and flow configurations. Several approaches are compared, including linear and nonlinear p-Poisson, Screened-Poisson, Eikonal, regularized Eikonal/Hamilton–Jacobi, and Laplace methods. Each formulation is discretized and assessed for numerical stability, efficiency, and accuracy against an exact geometric benchmark.
The validated models are applied to hydrodynamic and aerodynamic ship flows. For a model-scale bulk carrier (, ), Reynolds-Averaged Navier–Stokes (RANS) simulations with Shear Stress Transport (SST) turbulence show that different wall-distance formulations alter resistance, trim, and sinkage by less than 0.1%. A temporally constant wall-distance field proves sufficient for accurate propulsion predictions. In contrast, a full-scale feeder ship () analyzed with a hybrid RANS/LES (IDDES) model exhibits greater sensitivity to wall-distance definitions.
Among the tested methods, the convective Eikonal or Hamilton–Jacobi formulations with deferred correction achieve the best compromise between robustness, computational cost, and accuracy, whereas p-Poisson and Screened-Poisson variants are more parameter-sensitive and computationally expensive.
扫码关注我们
求助内容:
应助结果提醒方式:
