Research on the friction of the pneumatic servo system

Abstract

The pneumatic servo system is used more and more widely in industrial production because of its advantages like low cost, easy maintenance and environmental cleaning. However, due to the compressibility of gases, the nonlinear of the flow rate through valves, and the friction between the piston and the cylinder, the precise control of the pneumatic servo system has many difficulties. Among them, the friction between the piston and the cylinder cause the overshoot, the phase retardation and other undesirable phenomena during the operation process of the pneumatic servo system. Especially in the low speed section, the friction is easy to make the servo system have low speed jerking motion, which makes the system unstable. In order to solve this problem, the pneumatic servo gun system is used as the research object in this paper. And, two kinds of friction observation methods are used to observe the friction during the process of the system. Furthermore, the friction can be compensated in order to have the ideal control effects. When the electrodes driven by the cylinder do not contact, the method of establishing a reduced order observer is used to observe the friction. The specific method is that: the movement speed of the electrode and the friction of the cylinder are used to be the state variables and construct the state equation. On this basis, the reduced order observer is established to observe the friction and compensate the control of the displacement. When the electrodes contact, the state equation of considerable is difficult to establish. So the Kalman filter observer is established to estimate the friction. The specific method is that: the friction of the cylinder and the derivative of the friction are used to be the state variables. The output force of the electrode is used to be the observation variables. The pressure of the cylinder is used to be input variables. On this basis, the Kalman filter can be established to estimate the friction and compensate the control of the output force. Through the simulation and experiments, the two methods both have good observation and compensation effects.