Optimal shape design of spherical squeeze film bearings using genetic algorithms

Abstract

The optimal shape design of hydrodynamic axisymmetric spherical bearings under conditions of transient axisymmetric pure squeeze is investigated using genetic algorithms (GA). The optimization problem is to determine the radial clearance shape which maximizes the load impulse resulting from transient ball motion to a target minimum film thickness specification. For fully-hemispherical cups, non-dimensional analysis shows that GA-based optimal designs yield significantly larger non-dimensional impulse values when compared with those obtained either randomly or from optimal ellipsoidal geometry. A sample design application with specified load history shows that an optimal GA-inspired composite clearance specification amenable to fabrication can be constructed from preferred features obtained from the GA-based optimal shapes. The optimal composite shape design is shown to provide significantly better bearing performance, as measured by minimum film thickness, after a specified time interval than an optimal ellipsoidal design. The optimal composite shape design is also found to be less sensitive to realistic deviations in clearance shape, load impulse, and viscosity variation when compared with similar variations to an optimal ellipsoidal cup design.