Abstract
Surface roughness strongly influences turbulent boundary-layer development and frictional drag, particularly when distributed heterogeneously across a surface. This study examines spanwise heterogeneous roughness—alternating smooth and rough strips—using Reynolds-Averaged Navier–Stokes (RANS) simulations with a roughness-function wall model. The numerical setup replicates a well-established experimental arrangement to evaluate how spanwise width variations affect mean flow structure and frictional resistance. Verification and validation show good agreement with reference data, with deviations of about 2–4 % for boundary-layer thickness and global friction coefficients. The results indicate that small spanwise widths intensify interactions between adjacent smooth and rough regions, producing a thicker boundary layer and higher wall shear in the rough segments. As the width increases, this interaction weakens and the heterogeneous surface behaviour approaches that predicted by a simple mixed-average of smooth and rough conditions. Although global friction values for heterogeneous cases lie close to this mixed-average estimate, the local friction distribution remains highly sensitive to spanwise width, especially for the narrowest configurations. Mean velocity fields also reveal transitional behaviour at smooth–rough interfaces and spanwise momentum redistribution. Overall, the findings demonstrate that steady RANS captures the essential mean-flow effects of spanwise heterogeneous roughness and offers practical insight into drag behaviour under mixed surface conditions.
| Original language | English |
|---|---|
| Article number | 124208 |
| Journal | Ocean Engineering |
| Volume | 349 |
| DOIs | |
| Publication status | Published - 15 Mar 2026 |
| Externally published | Yes |
Keywords
- Frictional drag
- RANS simulation
- Spanwise heterogeneous roughness
- Turbulent boundary layer
- Wall-function roughness model
Fingerprint
Dive into the research topics of 'Study of the spanwise heterogeneous roughness impact using RANS'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver