Self-weight-buckling of vertical circular cylindrical shells.

Abstract

in which a = b — 0.5 for laminar flow and a = 0.8, b = 0.2 for turbulent flow. In Eq. (5), the properties p3 U3, and /^3 correspond to the inviscid edge values in the case of full shear layer reattachment, and to limiting streamline properties in the case of partial shear layer reattachment. Comparisons of the predictions of the analysis with a sampling of previously published heat transfer measurements obtained on a variety of reattachment geometries are given in Ref. 1, including cavities, compression ramps, spike nosed body, and forward facing step. Representative comparisons are shown in Figs. 2 and 3 for laminar and turbulent separation/reattachment over compression ramps. In Fig. 2, the maximum laminar reattachment heating level measured by Holden is well predicted, as is the decay in the recirculation region. In Fig. 3, Eqs. (1) and (2) are seen to seriously underpredict the maximum turbulent reattachment heating level measured by Holden. However, the simple empirical relation of Eq. (3) yields reasonable agreement with the data. The method presented herein offers a generalized approach for predicting both laminar and turbulent shear layer reattachment heating. Previously published methods are confined to semiempirical correlations applicable for specific geometries and type of flow.