An integrated bearing parameter identification approach using a nonlinear optimisation scheme

Abstract

In system identification process, efficient global optimisation schemes play a vital role. In the present work, an effective optimisation-based identification methodology of bearing stiffness and damping coefficients using the bearing response data is proposed. The flexible rotor is initially analysed by finite element model with nonlinear bearing forces. The dynamic equations of rotor are solved to obtain the bearing responses in frequency domain at different operating speeds. The equivalent speed dependent parameters of stiffness and damping are then obtained by solving an error-based optimal formulation via modified particle-swarm optimisation scheme. The robustness of the approach is tested with added input noise. The approach is illustrated with the simulation results of rotor supported on full floating ring bearings and using an experimental data obtained from a rotor supported on two similar journal bearings. The error in identification is not exceeding five percent. The modified particle swarm optimisation is converging faster, and it is taking less computational time. The results are shown in the forms of graphs and tables.