International Journal of Engineering and Technology

Valve stiction is one of the prominent stand-alone reasons for oscillatory behavior in process industry. Stiction is a common problem in spring diaphragm type valves, which are widely used in the process industry. This behavior leads control valves to deliver accurate feed and thereby allowing wastage of utility. Recently, there have been many attempts to understand, define, model and detect stiction in control valves. Most of the available methods cannot simultaneously detect and quantify stiction. The paper presents an Adaptive Neuro-Fuzzy methodology for identification of stiction. For a vertical two tank level process with Kano’s model of stiction is considered to obtain necessary required data to formulate objective function. The proposed methods for detecting and quantifying stiction are applied on the flow control valve using MATLAB/Simulink platform.