A well-established compensation technique to remove oscillations caused by control valve stiction is the two-moves method. However, the actual versions of this technique present major drawbacks, as the long time for implementation and, mostly, some strong assumptions on the valve position in oscillation. A recent version of two-moves compensation has proven to reduce the time of execution. Nevertheless, this method does not allow the control loop to handle set point tracking and disturbance rejection. The present paper proposes a revised version of two-moves stiction compensation method, which overcomes previous limitations. This new approach is based on the estimation of controller output associated with the desired valve position at the steady-state, by using the amplitude of oscillation before compensation and through the estimate of valve stiction, obtainable with specific techniques. In this case, fast responses are possible as well as a complete removal of the oscillation. In addition, set point tracking and disturbance rejection are guaranteed, by monitoring the control error and by switching temporarily to a standard PI(D) controller. Simulation examples and applications to a pilot plant show the effectiveness of the proposed method.
A Revised Technique of Stiction Compensation for Control Valves
BACCI DI CAPACI, RICCARDO
;SCALI, CLAUDIO;
2016-01-01
Abstract
A well-established compensation technique to remove oscillations caused by control valve stiction is the two-moves method. However, the actual versions of this technique present major drawbacks, as the long time for implementation and, mostly, some strong assumptions on the valve position in oscillation. A recent version of two-moves compensation has proven to reduce the time of execution. Nevertheless, this method does not allow the control loop to handle set point tracking and disturbance rejection. The present paper proposes a revised version of two-moves stiction compensation method, which overcomes previous limitations. This new approach is based on the estimation of controller output associated with the desired valve position at the steady-state, by using the amplitude of oscillation before compensation and through the estimate of valve stiction, obtainable with specific techniques. In this case, fast responses are possible as well as a complete removal of the oscillation. In addition, set point tracking and disturbance rejection are guaranteed, by monitoring the control error and by switching temporarily to a standard PI(D) controller. Simulation examples and applications to a pilot plant show the effectiveness of the proposed method.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.