Drag and velocity measurement on a two-dimensional bluff body has been performed in a pulsatile flow. Circular and rectangular cylinders were selected in the present study. The projected width and length of the cylinders were d =30 mm and L =300 mm, respectively. The cycle-averaged velocity was 〈V_θ〉 ≅15 m/s and the amplitude was in the range of V′_θm =0.45 to 3.6 m/s. The pulsation period of the flow was set in the range of T =1.5 to 10.0 s. The Reynolds number based on the phaseaveraged mean velocity and the projected width was in the range of R_eθ =20000 to 40000. The cycle-averaged drag coefficient increases with the root mean value of the deviation from the cycle-averaged velocity of approaching flow to the cylinder. The cycle-averaged drag coefficient for T <6.0 s ( T〈V_θ〉/d <3.0× 10³) increases compared with that estimated from the drag coefficient in the steady flow. For T ≥6.0 s ( T〈V_θ〉/d ≥3.0× 10³), the drag coefficient of the steady flow can be used to calculate the cycle-averaged drag. The effect of the pulsation period on the phase-averaged drag occurs during temporal deceleration of the pulsatile flow and can be related with the increase of the velocity deficit and width in the wake behind the cylinder. The phase-averaged drag will be represented as the sum of the momentum deficit due to the phase-averaged flow, which contributes to the cycle-averaged drag, and the pressure gradient by the pulsatile flow. This can be formulated semi-empirically with the cycle-averaged value of the phase-averaged drag coefficient, phase-averaged velocity and the derivative of the pulsatile flow with respect to time.

Effect of favorable pressure gradient on the wall similarity in turbulent boundary layer has been investigated experimentally. An appropriately adjusted pressure gradient produces an equilibrium boundary layer of which momentum thickness Reynolds number is constant. The wall shear stress was measured by three methods, which are direct measurement, Preston tube and Sublayer plate. A modified log-law considered the effects of the inertia and pressure gradient terms in the boundary layer equation gives better representation for the mean velocity profile subjected to favorable pressure gradient at finite Reynolds number. The velocity scale u_s applied to the modified log-law representation seems to be more relevant for turbulent intensity profile as well as mean velocity profile.

The roughness sublayer beneath a turbulent boundary layer has been investigated analytically and experimentally, and the effect has been discussed on the mean velocity profile throughout the layer. The rough surface consists of two-dimensional square roughness elements. The roughness height is proportional to streamwise distance. It is one of conditions of an equilibrium boundary layer proposed by Rotta J.C. The roughness pitch ratio is 4. The measurement within/outside the roughness sublayer was made using LDV and CTA, respectively. The Reynolds number based on the momentum thickness is 6000. The roughness Reynolds number is 149 and the roughness is in complete rough regime. The analytical solutions of streamwise mean velocity and Reynolds shear stress profiles were derived from the space averaged Reynolds equation in the cavity region between roughness elements and the mixing length model. The solutions were well consistent with the experimental results. By use of the solutions and experimental results, the center height of drag moment and the constant of the roughness function can be represented as a function of the roughness pitch ratio for the equilibrium boundary layer. Also, from the momentum integral equation, Coles's wake law, solution and experimental results, the wake parameter will be given as a function of the friction parameter and relative roughness height. Finally, the local skin friction coefficient will be formulated with a function of the roughness pitch ratio and relative roughness height.