Tilting Pad Journal Bearings: On Bridging the Hot Gap Between Experimental Results and Model Predictions

Abstract

The accurate prediction of the forced performance of tilting pad journal bearings (TPJBs) relies on coupling a fluid film model that includes thermal energy transport, and on occasion fluid inertia, to the structural stiffness of the pads' pivots and the thermomechanical deformation of the pads' surfaces. Often enough, the flexibility of both pads and pivots is ignored prior to the bearing actual operation; practice dictating that force coefficients, damping in particular, decrease dramatically due to pivot flexibility. Even in carefully conducted experiments, components' flexibilities are invoked to explain dramatic differences between measurements and predictions. A multiple-year test program at TAMU has demonstrated the dynamic forced response of TPJBs can be modeled accurately with matrices of constant stiffness K, damping C, and added mass M coefficients. The K-C-M model, representing frequency independent force coefficients, is satisfactory for excitation frequencies less or equal to the shaft synchronous speed. However, as shown by San Andrés and Tao (2013, “The Role of Pivot Stiffness on the Dynamic Force Coefficients of Tilting Pad Journal Bearings,” ASME J. Eng. Gas Turbines Power, 135, p. 112505), pivot flexibility reduces the applicability of the simple constant parameter model to much lower excitation frequencies. Presently, a fluid film flow model predicts the journal eccentricity and force coefficients of a five-pad rocker-back TPJB tested at TAMU under a load-between-pad (LBP) configuration. The predictions agree well with the test results provided the model uses actual hot bearing clearances and an empirical characterization of the pivot stiffness. A study follows to determine the effects of pad preload, r¯P = 0.0, 0.27 (test article), and 0.50, as well as the load orientation, LBP, and load-on-pad (LOP), on bearing performance with an emphasis on ascertaining the configuration with most damping and stiffness, largest film thickness, and the least drag friction. In the study, a rigid pivot and two flexible pivots are considered throughout. Further examples present the effective contribution of the pads' mass and mass moment of inertia and film fluid inertia on the bearing force coefficients. To advance results of general character, predictions are shown versus Sommerfeld number (S), a design parameter proportional to shaft speed and decreasing with applied load. Both LBP and LOP configurations show similar performance characteristics; the journal eccentricity increasing with pivot flexibility. For LBP and LOP bearings with 0.27 preload, pivot flexibility decreases dramatically the bearing damping coefficients, in particular, at the low end of S, i.e., large loads. The model and predictions aid to better design TPJBs supporting large specific loads.