Preparation and impact of operations on the use of the membrane aerated biofilm reactor (MABR) fingerprint soft sensor

ABSTRACT

This study investigated how the diurnal loading variation and physical sensor precision would affect the use of the MABR fingerprint soft sensor to generate valid signals in indicating biofilm thickness changes.

Membrane aerated biofilm reactors (MABRs) are increasingly being implemented to intensify processes in existing wastewater treatment facilities. Biofilm thickness is a key parameter for maintaining high-level performance in MABRs; however, frequent sampling and direct measurement are challenging in full-scale MABRs. We developed an indirect method using off-gas oxygen content (% O2) and bulk ammonia concentration (mg N/L) as an MABR fingerprint soft sensor to indicate biofilm thickness in full-scale MABRs. This study evaluated the conditions required for the MABR fingerprint soft sensor to generate precise, reproducible signals. The analysis of practical full-scale MABR processes reveals that sufficient diurnal loading variations are required to generate valid signals from the fingerprint soft sensor, considering the typical precisions of oxygen and ammonia sensors. Monte Carlo analyses were conducted to assess the precision of fingerprint soft sensors across different combinations of diurnal oxygen and ammonia variability levels, with typical physical sensor precisions. Results indicate that the MABR fingerprint soft sensor achieves an error of ≤10% when diurnal variability in oxygen and ammonia exceeds five times their respective sensor precisions. Introducing controlled loading variability into the full-scale MABR influent can enhance the precision of the fingerprint soft sensor and create more opportunities to evaluate MABR performance.