Impact of multiple roughness parameters on the turbulent channel flow using direct numerical simulation

This study investigates the impact of surface roughness on channel flow using a compounded method that integrates direct numerical simulation with a ghost cell immersed boundary method. A total of 8,400 rough surfaces, derived from the compressor blade, were analyzed to identify suitable roughness characteristics. The probability density functions of several key parameters led to the selection of roughness height ka, effective slope ES, and skewness Sk as primary criteria for identifying 27 rough surfaces for simulations under conditions of Ma = 0.3 and Reτ ≈ 300. Simulation results revealed irregularities in the effects of ka and ES on the roughness function ΔU+, with ΔU+ generally increasing with ka, although the trend became less pronounced for small differences in ka. The influence of ES on ΔU+ was minimal, as the ES values of the rough surfaces studied fell within the range where the dense regime begins. In contrast, a clear correlation was observed between the mean maximum peak-to-valley height kz and ΔU+, with positive Sk cases consistently yielding greater ΔU+ values. Integrating of results from prior studies, a predictive function for ΔU+ was developed using symbolic regression, demonstrating a mean relative error of 5.0% and an coefficient of determination R2 of 0.98. This research provides a comprehensive analysis of real-world rough surfaces and enhances predictive capabilities for their impact on roughness function.