Measurements of fluctuating wall shear stress beneath a turbulent boundary layer generated in a tow tank facility

Date of Completion

January 1997


Engineering, Mechanical




This research is an experimental study aimed at learning more about the makeup of the wall shear stress field beneath a turbulent boundary layer. Conventional anemometry equipment and classical data evaluation techniques are employed. A knowledge of the mean and fluctuating wall shear stress is necessary to more fully understand the drag on submerged bodies, the details of flow separation and cavitation, and the mechanisms governing heat transfer rates in wall-bounded flows.^ The tow tank facility at the U.S. Coast Guard Academy was used to carry out this study. A towed model capable of generating a fully developed turbulent boundary layer at relatively low tow speeds was designed, built and tested. A procedure for calibrating a flush-mounted hot film sensor to measure fluctuating wall shear stress is described in detail.^ The intensity of the wall shear stress fluctuations, $\tau\sb{\rm rms}/\tau\sb{\rm avg}$, was measured to be 0.25 and 0.36 for R$\sb{\theta}$ = 3050 and 2160 respectively. The probability density of wall shear stress fluctuations exhibits positive skewness, and lack of flow reversals at the wall. Correlations between velocity and wall shear stress verify the existence of large coherent structures. The spectral features of the fluctuations were examined in both normalized and dimensional form using conventional scaling laws. All scaling laws worked equally well in collapsing the spectra to a single curve.^ A multiple sensor array was designed and implemented to examine the convection and decay of wall shear stress fluctuations and make comparisons to available wall pressure data. Streamwise cross-spectra revealed that wall shear stress fluctuations decay more rapidly and scale differently than those for wall pressure. Low frequency fluctuations are longer lived, and convect downstream at higher speeds $\rm (U\sb{c}/U\sb{\infty}=0.80)$, than those which contribute at higher frequencies $\rm (U\sb{c}/U\sb{\infty}=0.52)$. Estimates of sensor spatial resolution errors using measured cross spectra indicate that spanwise averaging of the streamwise fluctuations is primarily responsible for attenuation of the autospectra. ^