Bifurcation Analysis of a Rotor Supported by Five-Pad Tilting Pad Journal Bearings Using Numerical Continuation
Academic Article
Overview
Research
Identity
Additional Document Info
Other
View All
Overview
abstract
This paper presents analytical bifurcations analysis of a Jeffcott type rigid rotor supported by five-pad tilting pad journal bearings (TPJBs). Numerical techniques such as nonautonomous shooting/arc-length continuation, Floquet theory, and Lyapunov exponents are employed along with direct numerical integration (NI) to analyze nonlinear characteristics of the TPJB-rotor system. A rocker pivot type five-pad TPJB is modeled with finite elements to evaluate the fluid pressure distribution on the pads, and the integrated fluid reaction force and moment are utilized to determine coexistent periodic solutions and bifurcations scenarios. The numerical shooting/continuation algorithms demand significant computational workload when applied to a rotor supported by a finite element bearing model. This bearing model may be significantly more accurate than the simplified infinitely short-/long-bearing approximations. Consequently, the use of efficient computation techniques such as deflation and parallel computing methods is applied to reduce the execution time. Loci of bifurcations of the TPJB-rigid rotor are determined with extensive numerical simulations with respect to both rotor spin speed and unbalance force magnitude. The results show that heavily loaded bearings and/or high unbalance force may induce consecutive transference of response in forms of synchronous to subsynchronous, quasi-periodic responses, and chaotic motions. It is revealed that the coexistent responses and their solution manifolds are obtainable and stretch out with selections of pad preload, pivot offset, and lubricant viscosity so that the periodic doubling bifurcations, saddle node bifurcations, and corresponding local stability are reliably determined by searching parameter sets. In case the system undergoes an aperiodic state, the rate of divergence/convergence of the attractor is examined quantitatively by using the maximum Lyapunov exponent (MLE).
