Sensor Fault Isolation in a Liquid Flow Process Using Kalman Filter

Abstract: Investigation of the effect of sensor fault in liquid process loop and designing a suitable fault isolation technique using the Kalman filter is proposed in this work. The objective of the work is to design an observer to estimate the process parameters, so as to compare with sensor functionality and thus identify and isolate faults in sensor. Based on the derived system model, observer is designed using Kalman filter approach. The proposed system with observer is subject to test in simulation and validated using physical system. The designed observer was successfully able to isolate drift, short circuit and open circuit faults in sensor. Practical validation shows the system with observer was able to track the set point with the root mean square of 0.986% error, even with faulty sensor.