Friction-resilient position control for machine tools—Adaptive and sliding-mode methods compared. (June 2018)

• Record Type: Journal Article
• Authors: Papageorgiou, Dimitrios; Blanke, Mogens; Niemann, Hans Henrik; Richter, Jan H.
• Is Part Of: Control engineering practice. Volume 75(2018)
• Page Start: 69
• Page End: 85
• Publication Date: 2018-06
• Subjects: Machine tools -- Mechanical drive-train -- Robustness to friction -- Nonlinear adaptive control -- Sliding-mode control -- Controller comparison -- High-accuracy positioning; Automatic control -- Periodicals; 629.89
• DOI: 10.1016/j.conengprac.2018.03.017
• Languages: English
• ISSNs: 0967-0661

Abstract

Abstract: Robust trajectory tracking and increasing demand for high-accuracy tool positioning have motivated research in advanced control design for machine tools. State-of-the-art industry solutions employ cascades of Proportional (P) and Proportional–Integral (PI) controllers for closed-loop servo control of position and velocity of the machine axes. Although these schemes provide the required positioning accuracy in nominal conditions, performance deteriorates with increased friction and wear of the machine. With conventional control, re-tuning is necessary during the lifetime if specified accuracy shall be maintained. This paper investigates whether nonlinear and adaptive controllers can cope with typical levels of friction increase without loss of performance. It evaluates the performance of a state-of-art industry solution with that obtainable with adaptive and sliding mode positioning controls. The main finding is that an adaptive backstepping control is resilient to unknown and increasing friction at realistic levels of wear, where the P-PI control fall short with respect to accuracy. A single-axis test rig with adjustable friction is used to assess the performance of different controllers. Highlights: Design & comparison of three nonlinear position controllers for machine tools. Performance evaluation with respect to accuracy and resilience to unknown friction. Comparison based on accuracy metrics and design and implementation complexity. Adaptive backstepping controller is superior to all other considered schemes. High order sliding mode controllers perform well only for slow position references.