Improvement of the dynamic behavior of ball screw feed drives by a sliding mode controlled machine table (SliMoReK2)

• Applicant: Professor Dr.-Ing. Alexander Verl
• Subject Area: Metal-Cutting and Abrasive Manufacturing Engineering ; Production Automation and Assembly Technology
• Term: since 2020
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 438835664

Constantly increasing demands on the manufacturing accuracy and productivity of modern machine tools require highly dynamic feed drive systems with excellent tracking and disturbance characteristics. The dynamic behavior of ball screws, which are predominantly used for highly dynamic feed axes, can be significantly increased through control improvements. In the first funding period of this project, the suitability of a sliding-mode position controller was demonstrated while retaining the remaining cascade control: in comparison to the P-PI cascade control, a reduction of the absolute tracking error by 82 % and an increase of the bandwidth by 6.7 Hz to 16.4 Hz with a simultaneous improvement of the disturbance behavior could be experimentally demonstrated in the guidance behavior. The high robustness of the approach was verified with variations of the model parameters by ±20 % and variation of the observers. The commissioning effort remains comparable to that of a P position controller. Thus, the investigated SMC PI controller shows great potential for industrial use, especially for processes with high dynamic accuracy requirements.The linear control has also been extended as a nonlinear sliding-mode controller and the performance of the two controllers was compared. In the follow-up project applied for here (2nd funding period), two major hurdles to practical applicability are now to be overcome: On the one hand, the controller is to be transferred to a multi-axis milling machine with industrial control and validated in practice (e.g. in milling experiments) in order to prove the transferability of the control law to different axes and for industrial implementation. On the other hand, the automated, simultaneous controller parameterization (and model identification of the model-based controller law) of position and velocity controllers will be investigated using two different methods to obtain an optimal, robustly stable setting of the parameters of the cascades with respect to each other. A comparable methodology does not exist in the state of the art so far, since the model of the position controller depends on the setting of the speed controller and the setting of the two cascades is thus coupled.

• DFG Programme: Research Grants
• Co-Investigator: Dr.-Ing. Armin Lechler