ZČU
Software for robust and optimal design of fixed structure controllers in electromechanical systems
Motion control loops are essential for modern mechatronic systems. They require proper design and tuning for the correct functionality and performance. Engineers and technicians have to adjust the parameters of motion controllers, which is a hard task. The controlled plant has unique dynamics for each machine, so the control tuning has to match the specific system. The tuning process is often manual and needs a lot of trial and error. The results are often suboptimal and depend on the people who do it.
Proper tuning is very important because industrial manufacturing systems have increasing performance demands that affect the control layer. For motion systems, this means strict requirements on bandwidth and tracking precision. High-performance requirements can lead to problems with vibrations when the bandwidth overlaps with the resonance modes of the controlled plant. Unwanted oscillations make the control system tuning more difficult. Automatic tuning methods are popular and successful in process control but not in mechatronic systems. Therefore, developing systematic methods that can assist with the commissioning process is very relevant for industrial practice.
We propose a new framework that can achieve both robustness and optimality in closed-loop motion systems.
Fig. 1. Assumed control setup – collocated motion system with different feedback and performance variables
We consider a feedback structure often called a "collocated control setup". This setup is common in mechatronics and motion control systems. The "collocated" term means that the actuator-sensor pair is physically at the same location of the controlled plant. The output y is usually provided by an encoder on the rotor shaft of an electric actuator. This signal is used to close the velocity or position feedback loop. However, the goal is often to control another physical variable z; usually the position or velocity of a reference point at the load-side moving part, such as a robot end-effector or a CNC machine tool spindle. If the load is ideally rigid, the working mechanism and the actuator move together. The outputs y and z are the same (except for possible scaling due to kinematic transform), and the control topology is a standard single-input-single-output (SISO) feedback loop. But mechanical flexibility in the load adds more degrees of freedom. More complex behavior with unwanted oscillations often occurs because of plant bending modes. This makes the feedback control design harder, as it inherently introduces a multivariable design problem. Good performance for feedback variable y does not automatically mean a good response for the variable of interest z.
The design method can be summarized as follows:
- Derive the model of the controlled plant from the experimental data using system identification or first principle / geometrical modelling, possibly including uncertainty for a robust control design
- Define feedback and performance variables y and z for the collocated control setup, forming a single-input-two-outputs system (the simpler SISO scenario is recovered by choosing y = z)
- Find a set of stabilizing controller of a given structure fulfilling certain robust stability condition
- From the admissible set, find an optimal vector of controller parameters with respect to the defined performance variable z and a chosen performance measure
- (optional) Provide a set of suboptimal reduced-gain controllers allowing in-situ fine-tuning of the controller during commissioning to find a suitable performance/cost trade-off
The algorithmic details were published in a conference paper
Goubej, M., Tvrz, J., Kubeš, B., Robust and optimal design of fixed structure controllers in collocated motion systems. In 2023 IEEE 28th International Conference on Emerging Technologies and Factory Automation (ETFA),
The control design algorithms were implemented in Matlab scripts in the m-language. C-code can be generated to employ them to various target platforms. The software can be used by means of a command-line interface, or through a developed graphical user interface (GUI) allowing the user to interfere and fine-tune the controller design process. The GUI is available in the form of a Matlab application or as a standalone application for Windows or Linux PCs.
Fig. 2. Graphical user interface of the developed control design tool
This project has received funding from the ECSEL Joint Undertaking (JU) under Grant Agreement No 101007311. The JU receives support from the European Union’s Horizon 2020 research and innovation programme and Netherlands, Czech Republic, Spain, Greece, Ireland, Italy, Belgium, Latvia, Portugal, Germany, Finland, Romania and Switzerland.
Licence
The software is free for non-commercial use. A license is required for commercial use.
Confirmation of usage
The software is used at the University of West Bohemia in Pilsen.
Contact form
This research/development result is protected by license. To download or get more information, please fill in the form below: