Beyond uniform roughness: spatial variability of hull roughness and implications for single-parameter representation

Hull surface roughness plays a critical role in determining ship resistance. In current practice, ship resistance prediction procedures of the International Towing Tank Conference (ITTC) commonly assume uniform roughness, represented by a globally applied equivalent roughness parameter. This study examines the implications of this assumption when spatially non-uniform roughness is present.

A review and structured synthesis of ITTC documents (1933–2024) indicate that hull surface roughness has been consistently represented by a globally applied parameter in resistance prediction procedures. This suggests that the uniform roughness assumption functions as a procedural abstraction within the established framework.

To investigate its physical implications, model-scale Computational Fluid Dynamics (CFD) simulations are performed under controlled roughness conditions using two benchmark hull forms: the KRISO Container Ship (KCS) and the KRISO Very Large Crude Carrier 2 (KVLCC2). Spatial roughness distributions—including linear (steep, mild, and reverse), random, and wall shear stress-based configurations—are considered under a constant area-averaged equivalent roughness height.

The results show that total resistance is sensitive to the spatial distribution of roughness even when the area-averaged equivalent roughness height is identical. This trend is consistently observed across both hull forms and roughness distribution types. Under the most severe roughness condition, deviations in the total resistance coefficient (C_T) reach approximately 12% for the KCS and 16% for the KVLCC2 relative to the uniform roughness baseline. These differences are primarily associated with variations in frictional response, together with changes in boundary layer development and wake recovery behaviour.

The findings indicate that, under spatially non-uniform roughness, resistance cannot be fully characterised by area-averaged roughness alone. Together with the ITTC document review, the results support interpreting the equivalent roughness parameter as a procedural abstraction based on a uniform roughness assumption. This representation captures the overall effect of roughness magnitude, but it does not fully describe resistance behaviour when roughness varies spatially. The results do not question the validity of existing ITTC methods within their intended scope; rather, they clarify the physical meaning and limitations of the uniform roughness assumption. They also provide physical context for the use and limitations of single-parameter roughness representation within the existing resistance prediction framework. Extension to full-scale conditions and more realistic roughness representations remains a subject for future work.