IoT and Machine Learning based Valve Stiction Estimation in Process Industry

Stiction in control valves represents an undesirable nonlinear phenomenon. Product quality and process safety frequently suffer due to the occurrence of valve stiction within process control systems utilizing control valves as their final control elements. This paper introduces a straightforward and...

Full description

Saved in:
Bibliographic Details
Published in2024 Control Instrumentation System Conference (CISCON) pp. 1 - 6
Main Authors V, Vimal Kumar, K, Ragesh G., George, Aparna, S, Priya
Format Conference Proceeding
LanguageEnglish
Published IEEE 02.08.2024
Subjects
Control systems
Internet of Things
Machine learning
MATLAB
Maximum likelihood estimation
Neural Network
Neural networks
Preventive maintenance
Process control
Safety
Thing Speak
Training
Valve Stiction
Valves
Online AccessGet full text

Cover

More Information
Summary:Stiction in control valves represents an undesirable nonlinear phenomenon. Product quality and process safety frequently suffer due to the occurrence of valve stiction within process control systems utilizing control valves as their final control elements. This paper introduces a straightforward and pragmatic Machine Learning technique, supported by Internet of Things (IoT) technology, for identifying and assessing the initiation of stiction in all pertinent control loops on a periodic basis. Furthermore, this paper suggests an IoT-based innovative approach to facilitate collaboration among industrial experts across multiple process plants, enabling them to explore new avenues for preventive maintenance concerning stiction in control valves.
DOI:10.1109/CISCON62171.2024.10696423