Simulation and experimental research of non-linear friction compensation for high-precision ball screw drive system

Non-linear friction is one of the main factors that effect both dynamic and static performance of high-precision mechanical servo system, and model-based friction compensation is the basic method of eliminating the impact of friction. This paper focuses on modeling, identification, and compensation of non-linear friction with simulation and experiment on ball screw drive system. Firstly, a simulation platform of the servo system according to high-precision experimental platform is established, based on dynamic model of the ball screw drive system and the LuGre friction model. Secondly, both static and dynamic parameters of the friction model are identified using genetic algorithm. Finally, the simulation and experiments of the friction feed-forward compensation are performed based on the friction model, utilizing by both the simulation platform and experimental platform. It shows that the experimental results are basically consistent with the simulation results without and with compensation. The feasibility and advantage of LuGre model based feed-forward compensation is verified by simulation and experiment, and it can decrease the tracking error of the system and improve the positioning accuracy.