The aim of the diploma thesis is to develop and implement a virtual laboratory for control system design and to demonstrate its usability by designing a suitable controller for the multivariable laboratory pilot plant - helicopter CE150. The thesis first describes the method for choosing the most suitable model of the plant for control design. At this point some methods for multivariable system analysis are introduced, such as controllability, functional controllability and observability. Furthermore, step responses of the plant in the chosen operating points are performed and the measured data is compared to the model's response. According to the collected data and its analysis, the most suitable model of the system is chosen. Afterwards, a few suitable controllers an their tuning methodology is introduced. First, conventional PID controllers, which are used in the same way as in SISO systems are described, followed by the cross-link elimination controller based on Bode plots. At last, state controller with state observer is presented. The latest is due to unmeasured system states necessary when implementing such controllers. Finally, design and implementation of the virtual laboratory is presented. The virtual lab allows experimenting on the real systems in order to evaluate the suitability of the designed controllers. The overall solution is based on MariaDB database, which stores all the neccesary data for system operation, while providing real-time measurements of the plant operation to the user. PHP programming language is used on the server side, while the user side uses JavaScript and jQuery. Matlab and PoKeys I/O device are used to communicate with the pilot plant.
|
