No abstract

Biofilm fouling is known to impact the data quality of sensors, but little is known about the exact effects. We studied the effects of artificial and real biofilm fouling on dissolved oxygen (DO) sensors in full-scale water resource recovery facilities, and how this can automatically be detected. Biofilm fouling resulted in different drift direction and bias magnitudes for optical (OPT) and electrochemical (MEC) DO sensors. The OPT-sensor was more affected by biofilm fouling compared to the MEC-sensor, especially during summer conditions. A bias of 1 mg/L was detected by analysing the impulse response (IR) of the automatic air cleaning system in the DO sensor. The IR is an effect of a temporal increase in DO concentration during the automatic air cleaning. The IRs received distinct pattern changes that were matched with faults including: biofilm fouling, disturbances in the air supply to the cleaning system, and damaged sensor membrane, which can be used for fault diagnosis. The results highlight the importance of a condition-based sensor maintenance schedule in contrast to fixed cleaning intervals. Further, the results stress the importance of understanding and detecting bias due to biofilm fouling, in order to maintain a robust and resource-efficient process control.

Monitoring and fault detection methods are increasingly important to achieve a robust and resource efficient operation of wastewater treatment plants (WWTPs). The purpose of this paper was to evaluate a promising machine learning method, Gaussian process regression (GPR), for WWTP monitoring applications. We evaluated GPR at two WWTP monitoring problems: estimate missing data in a flow rate signal (simulated data), and detect a drift in an ammonium sensor (real data). We showed that GPR with the standard estimation method, maximum likelihood estimation (GPR-MLE), suffered from local optima during estimation of kernel parameters, and did not give satisfactory results in a simulated case study. However, GPR with a state-of-the-art estimation method based on sequential Monte Carlo estimation (GPR-SMC) gave good predictions and did not suffer from local optima. Comparisons with simple standard methods revealed that GPR-SMC performed better than linear interpolation in estimating missing data in a noisy flow rate signal. We conclude that GPR-SMC is both a general and powerful method for monitoring full-scale WWTPs. However, this paper also shows that it does not always pay off to use more sophisticated methods. New methods should be critically compared against simpler methods, which might be good enough for some scenarios.