No abstract

A theoretical investigation into the next stage of dynamic stall, concerning the beginnings of eddy shedding from the boundary layer near an aerofoil's leading edge, is described by means of the unsteady viscous-inviscid interacting marginal separation of the boundary layer. The fully nonlinear stage studied in the present paper is shown to match with a previous ‘weakly nonlinear’ regime occurring in the earlier development of the typical eddy from its initially slender thin state. Numerical solutions followed by linear and nonlinear analysis suggest that with confined initial conditions the strong instabilities in the present unsteady flow tend to force a breakdown within a finite time. This leads on subsequently to an unsteady predominantly inviscid stage where the eddy becomes non-slender, spans the entire boundary layer and its evolution then is governed by the Euler equations. This is likely to be followed by the shedding of the eddy from the boundary layer.

No abstract

Ice accretion is considered in the impact of a supercooled water droplet on a smooth or rough solid surface, the roughness accounting for earlier icing. In this theoretical investigation the emphasis and novelty lie in the full nonlinear interplay of the droplet motion and the growth of the ice surface being addressed for relatively small times, over a realistic range of Reynolds numbers, Froude numbers, Weber numbers, Stefan numbers and capillary underheating parameters. The Prandtl number and the kinetic under-heating parameter are taken to be order unity. The ice accretion brings inner layers into play forcibly, affecting the outer flow. (The work includes viscous effects in an isothermal impact without phase change, as a special case, and the differences between impact with and without freezing.) There are four main findings. First, the icing dynamically can accelerate or decelerate the spreading of the droplet whereas roughness on its own tends to decelerate spreading. The interaction between the two and the implications for successive freezings are found to be subtle. Second, a focus on the dominant physical effects reveals a multi-structure within which restricted regions of turbulence are implied. The third main finding is an essentially parabolic shape for a single droplet freezing under certain conditions. Fourth is a connection with a body of experimental and engineering work and with practical findings to the extent that the explicit predictions here for ice-accretion rates are found to agree with the experimental range.

An extended version of the interactive boundary-layer approach which has been used widely in steady-flow calculations is applied here to the linear and nonlinear stability properties of channel flows and boundary layers in the moderate-to-large Reynolds-number regime. This is the regime of most practical concern. First, for linear stability the agreement found between the interactive approach and Orr-Sommerfeld results remains fairly close even at Reynolds numbers as low as about$\frac{1}{10}$of the critical value for plane Poiseuille flow, or$\frac{1}{5}$for Blasius flow. Secondly, nonlinear unsteady calculations and comparisons with full solutions obtained by enlarging the same method are also presented. Overall the work suggests that, at the finite Reynolds numbers where real interest lies, the dominant physical processes of instability in channel flow and boundary layers are of boundary-layer form, with interaction, and it suggests also an alternative numerical technique for determining those processes. This alternative technique uses the interactive boundary-layer method as the central means for obtaining full unsteady Navier-Stokes solutions.