Experimental investigation of laminar separation bubbles and comparison with theory

Abstract

Experimental and computational results from flow in and around laminar separation bubbles are given, with emphasis on the reattachment region. The experimental bubble parameters are correlated with existing formulas, utilizing a novel method of measuring with dense spacing in and behind the rear part of the bubble. Removal of the separation bubble by means of a transition trip is discussed. HE prediction of laminar separation bubbles creates numerous problems for aircraft design and develop- ment. The existence of laminar separation, bubble length, or failure to reattach, as well as implications concerning drag, have to be considered. Computational codes used, for exam- ple, in high-lift design1'2 are critically dependent on the validity of empirical criteria. This is an important issue also because a leading-edge stall for an airfoil may either be associated with a laminar bubble burst or, as pointed out by van den Berg,3 be a turbulent separation. At low Reynolds number model tests the laminar bubble burst may take place, whereas the stall in flight may occur because of turbulent boundary-layer separation. Despite considerable efforts, the engineering formulas used to predict the characteristics of laminar separation bub- bles are unreliable even for two-dimensional flows. The physical modeling and description of the bubble structure is rather vague. This is to some extent a consequence of the small physical dimensions of bubbles. The total height is often a fraction of a millimeter with bubble lengths of a few millimeters. Any probing of the bubble with, for example, hot wires or pitot probes, may seriously affect the entire flowfield. As the bubble is located close to the leading edge of the airfoil, very few pressure taps can be located in the bubble region, and generally it is difficult to obtain data of sufficient quantity and quality to improve the mathematical description of short, laminar bubbles. Furthermore, most engineering criteria use the momentum thickness Bs at laminar separation (i.e., the very start of the bubble) as scaling length, and com- puted values are often employed as reliable experimental values are very difficult to obtain. A number of theories and criteria suitable for engineering predictions are available, as exemplified by Refs. 4-7. Distinction is made between long and short bubbles, and whether the bubble is close to or far from bursting. The bub- ble itself is split into a laminar part associated with the laminar shear layer from laminar separation back to transi- tion and a turbulent part where the turbulent layer reattaches.