Experimentation for Fault Detection of Process Loop Components

Control valve performance needs to be monitored and maintained for a plant to operate at peak efficiency. During off-line outages, control valve inspection and maintenance may be insufficient and expensive. Testing the valves while they are in place and functioning normally can lower rebuild costs, enhance process performance, and concentrate maintenance efforts. There are many faults of control valves causes to degrade overall operation performance. The most common are stiction, hysteresis and backlash, improper valve sizing, air leakage and fault at control valve accessories. Even output pressure of current-to-pressure converter also depends on the installation, which may affect valve performance. Based on real-time recorded data, we can build the model for the components. We observe deviation in the actual and predicted signal occurred because of faults. Detection of faults and failures allows the maintenance team to take the required action, thus reducing the number of failures.