An Experimental Investigation of the Dynamic Performance of a Vertical-Application Three-Lobe Bearing

Abstract

Dynamic performance test results are provided for a vertical-application three-lobe bearing, geometrically similar to a three-lobe bearing tested by Leader [1] to stabilize a vertical sulfur pump. The bearing has the following specifications: 100° pad arc angle, 0.64 preload, 100% offset, 101.74 mm bore diameter, 0.116 mm radial pad clearance, 76.3 mm axial length, and 100° static load orientation from the leading edge of the loaded pad. The bearing is tested at 2000 rpm, 4400 rpm, 6750 rpm, and 9000 rpm. This bearing is tested in the no-load condition and with low unit loads of 58 kPa and 117 kPa. The dynamic performance of this bearing is evaluated to determine (1) whether a fully (100%) offset three-lobe bearing configuration is more stable than a standard plain journal bearing (0.5 whirl-frequency ratio), and (2) whether a fully offset three-lobe bearing provides a larger direct stiffness than a standard fixed-arc bearing. Hot and cold clearances are measured for this bearing. Dynamic measurements include frequency-independent stiffness and damping coefficients. Bearing stability characteristics are evaluated using the whirl-frequency ratio (WFR). Test results are compared to numerical predictions obtained from a fixed-arc bearing Reynolds equation solver. Dynamic tests show that the vertical-application three-lobe bearing does not improve stability over conventional fixed-arc bearings. The measured WFRs for the vertical-application bearing are approximately 0.4–0.5 for nearly all test cases. Predicted WFRs are 0.46 at all test points. The vertical-application bearing dimensionless direct stiffness coefficients were compared to those for a 70% offset three-lobe bearing. Dimensionless direct stiffness coefficients at 0 kPa are larger for the vertical-application bearing by 45–48% in the loaded direction and larger by 15–26% in the unloaded direction. Thus, the vertical-application bearing does impart a larger centering force to the journal relative to the 70% offset bearing, in the no-load condition. Predictions using both the measured hot clearance and measured cold clearance as inputs to the code are compared to the measured dynamic data. In general, the predicted direct stiffness coefficients using both the hot and cold clearances as inputs were higher than measured direct stiffnesses. The two sets of predicted cross-coupled stiffness coefficients straddle the measured cross-coupled stiffness coefficients. Predicted direct damping coefficients using both solutions were higher than measured values in most cases, but agreement between predictions and measurements improved significantly at high speeds and when applying light loads.