ELASTOHYDRODYNAMIC CONTACT MODEL FOR CALCULATION OF AXIAL AND ANGULAR STIFFNESS IN THRUST BEARING
Thrust bearing model is developed for fluid flow calculation and for determi-nation of bearing integral characteristics in the presence of sliding surfaces closure and shaft angular displacements. The model is based on the coupled solution of the problem of incompressible fluid flow between the sliding surfaces and the problem of bearing and shaft elements deformation under the action of the fluid film pressure. Verification of the bearing model results is carried out by the comparison versus the fluid flow calculation results obtained by STAR-CD software and the experimental and theoretical results represented in the certain literature. Thrust bearing charac-teristics are determined versus sliding surfaces closure and rotating disk (runner) angular displacements. The contribution of the sliding surfaces deformations into bearing integral characteristics is estimated.
1. Introduction
Hydrodynamic thrust bearing is used for compensation of the unbalanced axial force acting on rotor. Thrust bearings with hydrodynamic lubrication are used in high- and low-speed rotor supports of the stationary gas turbine units and testing rigs. Hydrodynamic lubricant in bearing has a considerably nonlinear stiffness and damping characteristics and hence influences station-ary gas turbine unit rotor stiffness and damping characteristics.
Rotor rotating produces angular displacements as a rigid body, as well as bending displacements, which also cause the runner’s angular displace-ments. So, despite the fact that the thrust bearing function is compensation of the rotor axial deflection, the runner’s angular displacement leads to the production of a moment preventing this displacement that has an influence on the rotor’s bending stiffness characteristics when it oscillates. The thrust
454 MIKHAIL TEMIS, ALEXANDER LAZAREV
bearing, due to its nonlinear stiffness and damping properties, may consider-ably affect rotor nonlinear vibrations pattern. The influence pattern that the thrust bearing has on rotor dynamics is insufficiently investigated at present times. It may be investigated after the development of an adequate thrust bearing model. Therefore, one of the main problems that exists in the static and dynamic analysis of a slide-bearings supported rotor is the calculation of thrust bearing stiffness and damping characteristics in the presence of the axial closure between the bearing sliding surfaces and with runner angular displacements taken into account. Experimental and numerical investigations of the lubricant flow in a thrust bearing with rigid sliding surfaces and the runner parallel to the thrust plate are presented in [1].
Despite all that – the gap between runner and thrust plate for a high axial loading is sufficiently small for providing sufficient fluid film pressure values in bearing. In such a case, deformations of sliding surfaces, produced by the fluid film pressure, are comparable with the fluid film thickness values. Particularly, this may have considerable effect on bearings with antifriction coating with low modulus of elasticity values that are widely used in rotor supports. As shown in papers [2-4], for radial bearings, changes in fluid film thickness for radial bearings due to radial deformations of the bearing work-ing surfaces considerably influence the carrying force value and direction, and hence the rotor dynamic characteristics. Similar phenomenon must be considered for the thrust bearing (Fig. 1). The following symbols are used in Fig. 1: R1 – internal radius; R2 – external radius; – lobe angle; hmin and hmax – minimal and maximal gap thicknesses, respectively; – shaft axis rotation angle; ! – shaft revolution speed; – angle of the shaft axis rotation plane; ' – angular coordinate; r – radial coordinate.
In the paper, the methodology for bearing stiffness calculation is de-veloped on the example of a 6-lobe thrust bearing with working surfaces deformability and runner angular displacements taken into account. The pre-sented methodology was verified versus experimental and numerical data published by Mote et al. [1] and calculations results obtained via STAR-CD software.