Position Control of a Pneumatic Valve Using Nonlinear Model Predictive Control Based on Kalman Filter

The paper proposes the control algorithm for position control for a pneumatic valve using the predictive control based on Kalman filter. The difference between the actual valve opening and desired valve opening is caused by the stiction dynamic of the valve. To model the valve with stiction in the piecewise linearizing form, the higher order transfer function is used. The Kalman filter then is used for determining the states of the valve, thereby the predictive controller with predictive window is slided respect to the time axis. The optimal control signal for controlling the valve position is also determined directly in each sampling cycle. The experiment results indicate that the proposed control algorithm works well, and the delay phenomenon caused by the stiction is reduced.