Rolling Stream Trails: An Alternative Cavitation Analysis

The Rolling Stream Trails (RST) model introduces new interpretation of photographic data of the cavitation process in an eccentric journal bearing gap. It stipulates a 3-D flow structure for transition from the Filled Fluid Film (FFF) to a cross-void fluid transportation process. The transition starts as a two-component composite rupture front and becomes an Adhered Film (AF) that is masked by rolled-on stream trails, which are drawn from the rupture front. The AF moves with the journal surface across the void to feed the FFF at the formation boundary. The rupture front comprises a first component of periodic streamers separated by a second component of interspersing wet pockets. Olsson’s Equation for flow continuity relative to a moving cavitation boundary (OEM) [1] yields a moving speed of the rupture front that is proportional to the reciprocal of the width fraction of the wet pockets multiplied into the FFF pressure gradient. Both rupture and formation boundaries move with finite speeds. The formation boundary speed in RST is determined by overall mass conservation in the bearing gap. RST formulation of the cavitation of a journal bearing fluid film yields an initial value time-dependent problem for the fluid flow in the bearing gap. The latter includes both the FFF and the AF that are joined at rupture and formation boundaries. The initial fluid content in the void span is bracketed between a dry void and a freshly cavitated wet void. As time progresses, transportation of AF across the void space and boundary motions form a coupled evolution process. Formulas for the boundary motions indicate that the formation boundary would become stationary simultaneously as the rupture boundary approaches the Swift-Stieber (SS) condition [2, 3]. Sample 1-D RST calculations illustrate temporal evolutions of the FFF pressure profiles, cavitation boundary locations and fluid content in the void span (AF profile). Resulting asymptotic pressure profiles are substantially different from published ECA calculations.