MPC for RPD

The drive control of the electrically preloaded system with a direct measurement system is performed by the industry-standard cascade controller. However, this controller is reaching its dynamic limits. To further improve workpiece quality, the drive control must therefore be optimized. A promising approach to increasing dynamics and accuracy is model-based control with predictive properties. The predictive behavior of the controller allows position errors to be detected earlier and compensated for. It is not yet known which control structure and which internal model are most suitable for the drive system with electrically preloaded drives, nor how the system-specific properties can be taken into account.

To improve the drive control of an electrically preloaded RPD, a model predictive controller (MPC) is applied to replace parts of the existing cascade controller. This control law determines the system behavior by predicting future behavior using an internal model. Due to its predictive properties, this control method can achieve a significant increase in dynamics and accuracy. The cascade controller used in industry for the drive control of an electrically preloaded RPD is shown in Figure 2. Three different MPC formulations, which replace various control loops of this cascade controller (see Figure 2), are investigated. For each control structure, different internal prediction models result, which must be identified and parameterized in advance and are intended to account for the compliant behavior of the system. The goal of the project is to investigate various MPC structures and to analyze their respective potential for improving dynamics and accuracy.