Modeling and control of industrial servo pneumatic actuator system

Abstract

Popularity of pneumatics is on the rise and they had received more attention since they offer many advantages over other type of force actuators such as costly effective, cleanliness and high power to weight ratio. Some of their engineering applications are included robotics, suspension system, haptic interface and etcetera. Servo pneumatic actuator system consists of a servo valve and a cylinder where its piston position needs to be controlled through servo valve as a commander in tracking desired trajectory along the stroke. The negative point of pneumatics which makes them difficult to control is their highly nonlinear behaviour. The objective of this study was to obtain mathematical model and control an industrial servo pneumatic system. Obtaining nonlinear mathematical model accurately to be used in controller design needs to determine all physical parameters of the real system which is very expensive and time consuming, to simplifying this procedure, model of system was analysed and obtained using system identification toolbox in Matlab. The system was excited with particular sine wave signal. Parametric approach using ARMAX structure was used to approximate the model. The best model was accepted based on the best fit criterion through SI toolbox. N-PID controller was designed for the model through the simulation. The results showed that N-PID controller provides better output than conventional PID controller. N-PID controller exhibits faster response to the system with desired transient error. But when the N-PID controller was applied on the real pneumatic system it showed very poor result because of existence of friction force. To improve the overall system output a friction compensator and a stabilizer attached to the N-PID controller. The system result illustrates that friction compensator and stabilized are very useful since they sufficiently enhance the controller performance. Self-tuning or robust controller beside of online system parameter estimation could be developed in future to increase the reliability of the controller.