Enhanced representation of the nonlinear dynamic characteristics of ball screw feed drive system through developing a three-state model

Abstract

This paper develops a three-state model to represent the nonlinear dynamic characteristics of the ball screw feed drive system (BSFDS), including the impact of backlash during velocity reversal. The three states are distinguished as a rigid model during unidirectional motion, a rigid model capturing the backlash phenomenon during velocity reversal, and a flexible model capable of describing the torsional vibrations. Impact of backlash on friction and inertia is considered in the three-state model, with a special friction model established to depict the effect of backlash. To identify the parameters involved in the model, a recursive maximum likelihood (RML) algorithm is proposed, based on the distinct signals designed specially for exciting the three states. The accuracy of parameter identification is enhanced through the combined use of iterative learning (IL). Based on the identified friction results using the IL-based RML, the model parameters are finally determined through least squares fitting. Experiments and comparisons are conducted to verify the effectiveness of this work.