Development of quantification algorithm for control valve stiction

Control valve stiction is the most commonly found valve problem in the process industry. Though many methods have been developed in detecting stiction, quantification of the actual amount of stiction presents in a loop is still an open research area. In this paper, Neural-network techniques are investigated in the development of a quantification algorithm for control valve stiction. It is shown that in the presence of a well-tuned controller, satisfactory performance of the proposed Neural-network based quantification algorithm can be achieved.