Monitoring of fouled DO-sensors with active fault detection

In this study, the automatic air cleaning system for dissolved oxygen (DO) sensors was used to monitor the sensor’s status. The response from an air cleaning impulse changed shape when artificial fouling was applied to the sensor.

Summary

In this study, the automatic air cleaning system for dissolved oxygen (DO) sensors was used to monitor the sensor’s status. The response from an air cleaning impulse changed shape when artificial fouling was applied to the sensor. In addition, the shape of the IR changed when other relevant sensor faults were introduced. Two fault detection methods were used to interpret the change in impulse response, Time constant estimation and Gaussian process regression.

Both methods were evaluated for their receiver operation characteristics and had high detection rate at a low false alarm rate in the conducted experiments. Both methods were sensitive enough to detect a fouling that resulted in a bias smaller than -0.2 mg/L. The results indicate that the suggested approach is promising and can increase the reliability of DO-measurements at WRRFs.

Coworkers: Oscar Samuelsson, Anders Björk

Year: 2017.0

Report number: A2278

Authors: Oscar Samuelsson, Anders Björk, Bengt Carlsson Jesús Zambrano

Published in: 12th IWA Specialized Conference on Instrumentation, Control and Automation (ICA), Quebec, Canada