An improved theory for flutter of cylindrical shells in cross-flow

Abstract

A series of experiments with cylindrical shells in cross-flow has shown that making the windward side of the shell rigid has little effect on ovalling, but if the same is done to the lee side, ovalling is suppressed, thus establishing the importance of what goes on in the wake in determining the onset of ovalling. Hence, the hybrid viscous-potential analytical model developed earlier, in which the phenomenon is modelled as one of aeroelastic flutter, has been refined, taking into account empirically determined changes in the base pressure as the shell is oscillating in a given mode, as well as introducing other refinements. Base pressure variations were determined by measuring the pressure around statically deformed shells, in given circumferential modes, and also measuring the phase lag between shell motion and pressure in the wake. Excellent qualitative and good quantitative agreement is achieved between the improved analytical model and experiment, making it more certain than ever that the mechanism of ovalling is one of aeroelastic flutter and that it is not related to vortex shedding. This realization made possible a more direct approach for determining the onset of ovalling, namely by establishing the balance of energy gained by the shell from the flowing fluid to the energy lost by dissipation. Agreement achieved between this method and experiment is excellent.