Multivariate Response Rotordynamic Modeling and Sensitivity Analysis of Tilting Pad Bearings

Achieving an optimal design of journal bearings is a very challenging effort due to the many input and output variables involved, including rotordynamic and tribological responses. This paper demonstrates the use of a multivariate response modeling approach based on response surface design of experiments (RSDOE) to design tilting pad bearings. It is shown that an optimal configuration can be achieved in the early stages of the design process while substantially reducing the amount of calculations. To refine the multivariate response model, statistical significance of the factors was assessed by examining the test's p-value. The effect coefficient calculation complemented the statistical hypothesis testing as an overall quantitative measure of the strength of factors, namely; main effects, quadratic effects, and interactions between variables. This provided insight into the potential nonlinearity of the phenomena. Once arriving at an optimized design, a sensitivity analysis was performed to identify the input variables whose variabilities have the greatest influence on the mean of a given response. Finally, an analysis of percent contribution of each input variable standard deviation to the actual response standard deviation was performed.