Teaching advanced topics in control engineering and understanding nonlinear modeling and modern control design methods are fascinatingand challenging tasks for teachers and students in a technical field. Teachers often combine classical lectures with laboratory experiments to teach these topics; however, commercially available experiments are typically expensive and limited in adaptations or changes that the user can make. Furthermore, systems engineers are required to work in project teams.
To meet the defined challenges and appropriately prepare our students for an industrial career, we worked with a student group to develop a laboratory device that is now used as a challenging laboratory experiment to design and implement state space controllers or more complicated control laws. In contrast to existing levitation experiments, the main advantages of this system include the ability to position an object in a 2D controlled B-field as well as the dimensions of the controllable region (10 x 10cm).
Figure 1. The Magnetic Levitation System
The Control Task and the Main Components
The main control task monitors the position of an object in a magnetic field. Therefore, the central element of the proposed levitation system is given by two in-house constructed solenoids with iron cores to produce a controllable magnetic field (B-field) by adjusting the current in the electromagnet though an input voltage. A power amplifier supplies the coil with a current that is proportional to the input voltage.
To implement a control loop, we determined the vertical and horizontal ball position by evaluating the signals of two line scans. As the main control unit of the system, we used a CompactRIO controller to fulfill real-time requirements of the process. In addition to implementing the control law, the real-time controller calculates the ball position from the measurement signals and communicates to the NI TPC-2006 touch panel device. Using a touch panel computer, we realized that we needed a human machine interface (HMI) to implement a comfortable interaction with the system. The initialization of the line scans and visualization of the current object position provide the graphical user interface. Furthermore, the touch screen allows the user to manually apply a reference input and specify various options during the control process.
In addition to choosing the relevant components for building the proposed system, we had to describe the dynamic behavior of the system, and we needed an appropriate dynamic model of the system for a successful control design. In our case, the system is both nonlinear and unstable. While developing the system, our students learned how geometric conditions of a coil affect the corresponding magnetic field and that nonlinearity is caused by the nonlinear extension of the magnetic field created by the coils. The most challenging requirement to meet was the long distance (10 cm) for the proposed tracking problem because of the highly nonlinear relationship between the field and distance. We learned that it was beneficial to use magnetic objects to achieve the required specifications. Concerning geometrical form of the object balls avoid difficulties when the object is turning around.
We can approximately describe the system by a third-order nonlinear state space model. The corresponding state variables are the distance from the coil, the velocity of the object, and the current through the coils. Applying a linearization procedure around an operating point (object position) yields in a linear, time-invariant, and unstable state space model. Based on this model, we successfully applied mathematical description classical state space controller design methods and conducted the design steps using the LabVIEW Control Design and Simulation Module.
By conducting practical experiments, we realized that our system required some immediate improvements. First, we extended the state space controller concept by finding controllers for at least three operating points and applying a simple switching strategy; and then we added a shiftable integrating element in the control loop. We quickly implemented these improvements, and, while testing the experiment, observed that due to the inevitable high current, the heating process in the coils cannot be neglected for a sufficiently long operation time. Therefore, we conducted temperature measurements of the coils. Some changes in the mathematical model led to improved temperature dependent control laws and guaranteed the stabilization of the object for a longer time period.
System Flexibility and Successful Development
Because our proposed system will be applied in student laboratory conditions, ruggedness and flexibility are required properties, and monitoring specific signals during the control process is one of the most helpful methods for students when testing a control law. Furthermore, we designed an interface board to transfer measurement data to an external PC because we now have the option to implement the control law in software running on the external PC, which gives us the ability to use math-script-based tools for controller implementation. The system is also prepared to test many other control design methods for finding an appropriate control method for the object's position. Using NI hardware and software, we developed this flexible laboratory experiment in a short period of time to increase academic practice and knowledge of nonlinear system modeling and control.
About the Author
Wolfgang Werth at the Carinthia University of Applied Sciences, can be reached at: firstname.lastname@example.org.
Related Case Studies