Component-based mixed reality environment for the control and design of servo-pneumatic system
Synopsis Considerable research efforts have been spent over the last two decades on improving the design, control, and modelling of pneumatic servo drive systems including the development of dedicated controllers and control valves. However, the commercial updates in employing pneumatic servos are still largely limited to laboratory research usage and the initiatives in developing seem to have lost their momentums. Although this situation has some to do with the rapid development and availability of cost effective electric servo technologies, one reason is considered to be a lack of design and simulation tools for employing pneumatic servo drives. This research has therefore been conducted to address these concerns, and to demonstrate how appropriate tools and environments can be developed and used to aid in the design, control and commissioning of pneumatic servo drives. Because of the inherent high nonlinearities associated with pneumatic systems, it would be highly desirable if the simulation environment could be run in time domain so that it can be mixed with the real system. This would make the simulation more accurate and reliable especially when dealing with such nonlinear systems. Unfortunately, the tools that are available in the market such as Propneu (Festo, 2005) and Hypneu (Bardyne, 2006) are dedicated for pneumatic circuit design only. This research is aimed at developing a mixed reality environment for the control and design of servo-pneumatic systems. Working with a mixed reality environment would include both the capability to model the system entirely as a simulation, the so-called "off-line", as well as being able to use real components running against simulations of others "on-line", or in a Mixed Reality (MR) manner. Component-based paradigm has been adopted, and hence the entire pneumatic system is viewed as a series of component modules with standardised linking variables. The mathematical model of each individual component has been implemented in simulation software which produces time domain responses in order to allow for mixing the simulation with the real system. The main outcome of this research can be seen as a successful development and demonstration of the Component-based Mixed Reality Environment (CMRE), which would facilitate the control and design of servo-pneumatic systems. On the one hand, the CMRE facilitates the identification of some nonlinear parameters such as frictional \I ynopsis parameters. These parameters could cause great difficulties in servo-pneumatic modelling and control. Accurate friction parameters would give the ability to attain an accurate model, and therefore provide more flexibility in applying different control and tuning strategies on the real system. On the other hand, the CMRE facilitates the design process by enabling the designer to evaluate the servo-pneumatic system off-line prior to building the system. This would reduce the design time, increase the reliability of the design, and minimize the design cost. The concept of the CMRE was validated by tests carried out on laboratory-based prototype servo-drive. Close agreement between the experimental and simulated responses was obtained showing that the models have represented the real system adequately. Case studies were then conducted to demonstrate the validity of the proposed methodology and environment. In these case studies, PIDVF controller and cascade control structure were successfully implemented, synthesised, and tuned. The results revealed that the CMRE is an easy, accurate and robust way of implementing different control and tuning strategies on servo-pneumatic systems. Furthermore, the research has shown how the CMRE can lead to significant improvements in certain life cycle phases of the system, e.g. commissioning, maintenance, etc. This research has contributed to knowledge in the following: (1) Adopting the mixed reality concept and the component-based approach in order to create a CMRE in facilitating the control and design of servo-pneumatic systems. (2) A method to identify the friction parameters of a single-axis pneumatic machine, (3) Encapsulate existing control methods within the CMRE to be applied on the real system. (4) A scheme for controller tuning, in which the controller is tuned off-line and then applied on the real system, and hence avoided on-line tuning which can be troublesome and time consuming. It is anticipated that the concept of the CMRE can be extended to include multi-axes servo-pneumatic system, servo-hydraulic, and servo-electric drives. Therefore, conceptual model structures have been introduced in this research which can be considered as the foundation for creating similar environments for those systems.
- PhD