Intelligent Control Valve Stiction Diagnosis Approach

Variations in process variables are one of the major concerns in industrial plants due to their impact on product quality and equipment performance, leading to higher operating and maintenance costs in addition to safety issues. One of the common sources of process variation is control valve stiction. Therefore, early detection and quantification of valve stiction is crucial for maintaining product quality and reducing maintenance costs. A smart valve equipped with a positioner simplifies stiction diagnosis, whereas the diagnosis process for an ordinary control valve is not straightforward. As the number of ordinary control valves installed in plants is still quite high, we developed in this research a smart approach for stiction detection and quantification for this type of control valve. In order to implement the proposed approach, we used state-of-the-art signal processing and machine learning techniques to develop a trustworthy strategy for stiction analysis applicable to a wide variety of operating setups, such as cascade loops and settings variations. Testing this approach on publicly available industrial data showed that it worked well for all tested scenarios to detect the stiction and provided an accurate estimate of the stiction value.