Dynamic Analysis and Machine Learning Prediction of a Nonuniform Slot Air Bearing System

Abstract Nonuniform slot air bearing (NSAB) systems have two major advantages, the external supply and restrictor design, their inherent multidirectional supporting forces stiffness that provide excellent rotational stability. However, NSAB are prone to vibration from nonperiodic or chaotic motion caused by nonlinear pressure distribution within gas film, imbalance, simply inappropriate design. It is necessary determine under which conditions these motions arise, design a system will resist vibrations. The dynamic behavior of rotor supported an was studied using spectral response, bifurcation, Poincaré map, maximum Lyapunov exponent. numerical results showed chaos in occurred specific ranges mass number. For example, regions where exponents were greater than zero intervals 20.84 ≦ mf < 24.1 kg with number Λ = 3.45. In addition, coupling effect also investigated. To predict behavior, ensemble regression, back propagation neural network used forecast occurrence chaos. found regression dataset 26 × 121 gave best most accurate prediction for this system. may make valuable contribution use wide variety industrial applications.