Artificial neural network approach for detection and diagnosis of valve stiction

Abstract

Valve stiction or static friction in control loops is a common problem in modern industrial processes. Several recent studies have tried to understand, reproduce, and detect such issue; however, the actual quantification is still a challenge. Since the valve position (mv) is normally unknown in industrial process, the main challenge is to diagnose stiction knowing only the output signals of the process (pv) and the control signal (op). This paper presents an artificial neural network approach in order to detect and quantify the amount of static friction using only the pv and op information. This study was validated by a simulation process. The results show satisfactory measurements of stiction.