Solving Sensor Reading Drifting Using Denoising Data Processing Algorithm (DDPA) for Long-Term Continuous and Accurate Monitoring of Ammonium in Wastewater

Abstract

Sensor reading drifting caused by sensor property deterioration is a major problem of long-term continuous monitoring in wastewater and hinders wide-range application of online wastewater management. This study aims to tackle this problem by developing denoising data processing algorithm (DDPA) for a typical electrochemical sensor, solid-state ion-selective membrane (S-ISM) sensor. Based on data mining and electrochemical principles, DDPA was designed by combining digital filter and outlier analysis to differentiate actual sensor readings from background noise when the S-ISM sensitivity declined over time. The sensor sensitivity was raised from 21 mV/dec to 55 mV/dec after the reading processing, without compromising the detection limit (7 × 10–6 mol/L). Furthermore, long-term accuracy of S-ISM sensors in wastewater was enhanced by adding hydrophobic polytetrafluoroethylene (PTFE) into polymer matrix. The sensitivity (57 mV/dec) of PTFE-loaded S-ISM sensors was the near-theoretical value on the first day and still higher than 35 mV/dec after 24 days in wastewater, providing an excellent stable baseline for DDPA. Combination of sensor material enhancement (adding PTFE) with sensor reading processing (using DDPA) assured the stable and high sensitivity (55 mV/dec after 24 days) and high detection limit (<5 × 10–5 mol/L) for wastewater monitoring. The study demonstrates a new route toward long-term accurate wastewater monitoring and smart wastewater sensor networks by establishing a strong correlation between multiorder derivatives of sensor readings and electrochemical responses with DDPA as an efficient data analysis approach.

Electrospraying Zwitterionic Copolymers as an Effective Biofouling Control for Accurate and Continuous Monitoring of Wastewater Dynamics in a Real-Time and Long-Term Manner

“…The newly developed DDPA exhibited a lower discrepancy (1.2 mg/L for 50 days and 2.5 mg/L for 55 days) against the lab-based validation results compared with the original data of the PTFEMA- r -SBMA-loaded S-ISE sensor (1.8 mg/L for 50 days and 3.2 mg/L for 55 days) (Figure c). These results explicitly demonstrate that the “sensor material improvement plus data processing using DDPA” package achieved long-term continuous and accurate detection of contaminants (with NH 4 + as the model ion in this study) in wastewater …”

“…, temperature and pH , ) cause the fluctuation and drifting of S-ISE sensor readings. We had developed DDPA to differentiate actual sensor readings from background noise when the sensor sensitivity declined in wastewater over time . In this study, a new DDPA was developed to improve the accuracy of the processed sensor data because previous DDPAs failed to identify and correct the random signal fluctuations generated from data transmission malfunction.…”

“…, minute-scale or even second-scale precision). Our previous study demonstrated that DDPA accurately eliminated the noise not coming from electrochemical responses by filtering the terms with higher-order derivatives in Taylor series . However, the Taylor series-based DDPA misjudged steady fluctuations of minor signals and declined concentration gradients when the sensor readings deteriorated over time.…”

“…Our previous study demonstrated that DDPA accurately eliminated the noise not coming from electrochemical responses by filtering the terms with higher-order derivatives in Taylor series. 23 However, the Taylor series-based DDPA misjudged steady fluctuations of minor signals and declined concentration gradients when the sensor readings deteriorated over time. To eliminate this effect, we redesigned the DDPA kernel using the Lagrange approximation to capture and correct the errors of data transmission between the sensor and electrical interface by taking data from sensor readings with a rolling standard deviation window.…”

Forward-Looking Roadmaps for Long-Term Continuous Water Quality Monitoring: Bottlenecks, Innovations, and Prospects in a Critical Review

“…For example, a denoising data processing algorithm has been employed to quantify the impacts of polytetrafluoroethylene on potentiometric ammonium sensors in terms of accuracy and lifespan during a 20 day test in WW. 282 Furthermore, the data generated from LTCM delivers months/years monitoring scopes and minute- or even second-temporal resolutions, which can maximize the computing ability, predict sensor reading variation, and eliminate sensor data drift. 283 Finally, we also suggest applying sensor arrays consisting of different types of miniature sensors for multiplexed detection of a broad spectrum of water parameters and using data fusion/ML algorithms to achieve high-resolution profiling in water systems.…”