BP Network Model for Nonlinear Oil-film Force on Hydrodynamic Bearing

秦平,孟志强,朱均
DOI: https://doi.org/10.3321/j.issn:1004-0595.2002.03.016
2002-01-01
Abstract:The database method to compute nonlinear oil-film force of the finite width hydrodynamic journal bearings can solve the contradiction between accuracy and efficiency, but the database and formulae of oil-film force are subsections. Positions and velocities of bearings are synthesized into three basic parameters and the state space transformation is used to change discontinuous oil-film force databases to consecutive oil-film force databases. So the integrative formula and the BP network model based on consecutive oil-film force databases are established to obtain the oil-film force under any movement state of axes with high accuracy. By means of the computation example of cylinder journal hydrodynamic bearings, the finite differential method, the database method and the BP neural network model are employed to calculate the oil-film force and the orbit of shaft center in the transient analysis of the rotor-bearing system. Results show that the result achieved with the BP network model is more approximate to that achieved with the numerical computation method and the BP network model can raise remarkably the computation efficiency of bearings.
What problem does this paper attempt to address?