Online friction parameter estimation for machine tools. Issue 1 (25th February 2020)

• Record Type: Journal Article
• Authors: Papageorgiou, Dimitrios; Blanke, Mogens; Henrik Niemann, Hans; Richter, Jan H.
• Is Part Of: Advanced control for applications. Volume 2:Issue 1(2020)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2020-02-25
• Subjects: adaptive estimator -- cascaded systems -- experimental validation -- friction estimation -- machine tool drive -- nonlinear parameterization -- stiction -- wear assessment; Automatic control -- Periodicals; Automatic control; Periodicals; Electronic journals; 629.8
• DOI: 10.1002/adc2.28
• Languages: English
• ISSNs: 2578-0727

Abstract

Abstract : Accurate description of frictional phenomena is essential in applications that require high‐accuracy control. Online monitoring of friction parameters in machine tools greatly improves the control accuracy making condition‐based feedforward compensation possible and, at the same time, facilitates equipment wear assessment, which enables efficient scheduling of maintenance. Existing friction estimation methods that use detailed dynamical models offer accurate description of friction phenomena, but often rely on a priori knowledge of the static friction parameters, which have to be identified offline. This article suggests an adaptive estimation strategy suitable for online use while the machine works in its normal production cycle. Smooth approximations are introduced to account for stiction, viscous and bidirectional Coulomb friction in order to make online estimation possible. A parallel architecture is used with two adaptive estimators that segregate the frictional phenomena that dominate in different parts of the motion regime. Stability properties are analyzed and performance is experimentally validated on a single‐axis state‐of‐the‐art industrial test rig. Abstract : This study pursues online friction parameter estimation for machine tool drive axes. Smoothened first‐principle models of low complexity describe the different frictional phenomena, i.e. Stiction, Coulomb and viscous friction. The friction coefficients are identified online by a cascaded scheme comprising acceleration estimation, motion regime separation and two adaptive estimators. The proposed method is tested on state‐of‐the‐art industrial equipment and it is demonstrated that accuracy corresponding to estimation errors of 1% of the true values is achievable under appropriate excitation conditions.