Scaling and structure of the momentum and scalar fields within a turbulent boundary layer perturbed by an effusion film and its comparison to asymptotic behaviour of pipe flow momentum

The first part of this thesis presents a new scaling paradigm for the intermediate region of pressure-driven, fully developed turbulent boundary layers. The log law assumption is shown to be invalid for scaling the location of Reynolds shear stress maximum, which predicts $ y^+_{m\tau} \propto Re_{\tau}^{\beta}$ with $\beta = 0.5$. An outer peak in premultiplied production is observed to be further from the wall than the Reynolds shear stress peak, indicating that $\beta > 0.5$. Fitting to high Reynolds number turbulent pipe and channel data produces $\beta = 0.529 \pm 0.005$. An intermediate velocity scale $\Delta U^+_m \propto Re_{\tau}^{2\beta - 1}$ is derived. The magnitude and location of the outer peak in streamwise velocity variance are shown to scale well with $\Delta U^+_m$. A second-order Taylor expansion about $ y^+_{m\tau}$ is used to predict the location of the outer peak in premultiplied production, which has good agreement with high Reynolds number data. The second part of this work presents a large-scale experimental study, where a canonical turbulent boundary layer is perturbed by an effusion film - applicable to gas turbine blade cooling. Lower temperature air is injected into the boundary layer through a staggered array of small holes. The injection-freestream velocity ratio ($VR$) is varied from $0.10 < VR < 0.37$. Three-component laser Doppler anemometry and oil film interferometry characterise the momentum field, and simultaneous x-wire and cold-wire characterise the joint momentum-scalar field. The boundary layer scaling behaviour changes about a critical velocity ratio $VR_c \approx 0.20$, which is attributed to bulk film lift-off. Large spanwise differences in the temperature and momentum fields are observed. Turbulence is significantly enhanced, particularly in the outer region and the near-wall spanwise fluctuations. Neither the local equilibrium hypothesis or Reynolds' analogy are valid.