Townsend’s attached-eddy hypothesis explains the asymptotic statistics of wall turbulence through a superposition of self-similar attached eddies. At a given observation location, active portions contribute to momentum transfer, whereas inactive portions retain only wall-parallel fluctuations. Since this distinction depends on the observation height relative to the eddy height , we extract single-eddy intensity functions from DNS of turbulent pipe flow at –. Spectral linear stochastic estimation and differencing isolate wall-coherent contributions at prescribed , yielding the streamwise and wall-normal intensity functions and . When expressed in , both collapse across and share a peak at , close to the von K’arm’an constant. Near , both components remain significant, whereas for , diminishes while remains finite, identifying active and inactive portions. Over the self-similar range, both peak intensities scale as , and retains this scaling in both portions. Integrating the active portion yields an approximately constant contribution, while the inactive portion recovers the classical near-wall logarithmic variation. An outer logarithmic tendency with a slope close to the Townsend–Perry constant arises from integrating active contributions up to their outer cutoff. At the near-wall peak of , the scaling steepens to with , yielding a finite asymptote with a defect. The additional exponent is attributed to viscous attenuation of inactive footprints associated with a -dependent dissipative scale. Wall-coherent motions also produce a broad outer shoulder, while a localized wall-incoherent residual may contribute to a secondary peak.

生成讲解博客的中文版本

生成AI智能校验的中文版本