Application of convolutional neural networks for prediction of disinfection by-products

Fluorescence spectroscopy can provide high-level chemical characterization and quantification that is suitable for use in online process monitoring and control. However, the high-dimensionality of excitation–emission matrices and superposition of underlying signals is a major challenge to implementation. Herein the use of Convolutional Neural Networks (CNNs) is investigated to interpret fluorescence spectra and predict the formation of disinfection by-products during drinking water treatment. Using deep CNNs, mean absolute prediction error on a test set of data for total trihalomethanes, total haloacetic acids, and the major individual species were all < 6 µg/L and represent a significant difference improved by 39–62% compared to multi-layer perceptron type networks. Heat maps that identify spectral areas of importance for prediction showed unique humic-like and protein-like regions for individual disinfection by-product species that can be used to validate models and provide insight into precursor characteristics. The use of fluorescence spectroscopy coupled with deep CNNs shows promise to be used for rapid estimation of DBP formation potentials without the need for extensive data pre-processing or dimensionality reduction. Knowledge of DBP formation potentials in near real-time can enable tighter treatment controls and management efforts to minimize the exposure of the public to DBPs.

Effects of Filtration Mode on the Performance of Gravity-Driven Membrane (GDM) Filtration: Cross-Flow Filtration and Dead-End Filtration

Gravity-driven membrane (GDM) filtration technology has been extensively in the employed drinking water treatment, however, the effect filtration mode (i.e., dead-end mode vs. cross-flow mode) on its long-term performance has not been systematically investigated. In this study, pilot-scale GDM systems were operated using two submerged filtration mode (SGDM) and cross-flow mode (CGDM) at the gravity-driven pressures 120 mbar and 200 mbar, respectively. The results showed that flux stabilization was observed both in the SGDM and CGDM during long-term filtration, and importantly the stabilized flux level of CGDM was elevated by 3.5–67.5%, which indicated that the filtration mode would not influence the occurrence of flux stability, but significantly improve the stable flux level. Interestingly, the stable flux level was not significantly improved with the increase of driven pressure, and the optimized driven pressure was 120 mbar. In addition, the GDM process conferred effective removals of turbidity, UV254, CODMn, and DOC, with average removals of 99%, 43%, 41%, and 20%, respectively. With the assistance of cross flow to avert the overaccumulation of contaminants on the membrane surface, CGDM process exhibited even higher removal efficiency than SGDM process. Furthermore, it can be found that the CGDM system can effectively remove the fluorescent protein-like substances, and the intensities of tryptophans substance and soluble microbial products were reduced by 64.61% and 55.08%, respectively, higher than that of the SGDM. Therefore, it can be determined that the filtration mode played an important role in the flux stabilization of GDM system during long-term filtration, and the cross-flow filtration mode can simultaneously improve the stabilized flux level and removal performance.

Predicting the distribution of arsenic in groundwater by machine learning technique in two worst hit districts of Assam, India: a risk to public health

Arsenic (As) is a well-known human carcinogen and a significant chemical contaminant in groundwater. The spatial heterogeneity in the distribution of As in groundwater makes it difficult to predict the location of safe areas for tube well installations for consumption and agricultural use. Geospatial machine learning techniques have been used to predict the location of safe and unsafe areas of groundwater As contaminations. Here we used a similar machine learning approach to determine the risk and extent of As >10 ug/L in groundwater at a finer spatial resolution (250m x 250m) in two worst-hit districts of Assam, India, to advise policymakers for targeted campaigning for mitigation. Random Forest Model was employed in Python environments to predict probabilities of the occurrences of As at concentrations >10 ug/L using several intrinsic and extrinsic predictor variables. The selection of predictor variables was based on their inherent relationship with the occurrence of As in groundwater. The relationships between predictor variables and proportions of As occurrences >10 ug/L follow the well-documented processes leading to As release in groundwater. We identified extensive areas of potential As hotspots based on the probability of 0.7 for As >10 ug/L. These identified areas include areas that were not previously surveyed and extended beyond previously known As hotspots. Twenty-five percent of the land area (1,500 km2) was identified as a high-risk zone with an estimated population of 155,000 potentially consuming As through drinking water or food cooked with water containing As >10 ug/L. The ternary hazard map (i.e., high, moderate, and low risk for As >10 ug/L) could inform the policymakers to target the regions by establishing newer drinking water treatment plants and supplying safe drinking water.

The Analysis of Decision-making Items for Operation Management of Digital Twin Simulators for Sewage and Water Treatment Facilities

This paper addresses the derivation of decision-making factors for the operation and management of the Cyber Physical systems (CPS)-linked wastewater treatment plants (WWTPs) and drinking water treatment plants (DWTPs) simulator. The analytical hierarchy process (AHP) method was applied to evaluate the importance of each influencing factor on the operational elements targeting experts. Here, the experts were 37 people working for WWTPs and 30 people working for DWTPs, respectively. The analysis factors for decision-making were influent load, reactor capacity, treatment process, occurrence of high turbidity influent, risk, urgency, response, and recovery, and their relative importance was analyzed. Among the operational elements of WWTPs, influent fluctuations have been shown as the most important factor. Conversely, the possibility of occurrence was found to be the most important influencing factor. In the case of DWTPs, the inflow of high turbidity was found to be the main operating factor, and the influencing factors were the probability of occurrence and the degree of risk. Based upon the results obtained, this study is expected to contribute to the establishment of a stable system of both WWTPs and DWTPS by identifying influent fluctuations, which are a major influencing factor, and by controlling operation factors connected with the establishment of a digital twin simulation.

The Impact of Extreme Weather Events on Bacterial Communities and Opportunistic Pathogens in a Drinking Water Treatment Plant

Drinking water treatment processes are highly effective at improving water quality, but pathogens can still persist in treated water, especially after extreme weather events. To identify how extreme weather events affected bacterial populations in source and treated water, water samples were collected from the Yangtze River Delta area and a local full-scale drinking water treatment plant. Bacterial community structure and the occurrence of pathogens were investigated in samples using 16S rRNA sequencing and qPCR techniques. In this study, the results show that intense rainfall can significantly increase levels of bacteria and opportunistic pathogens in river and drinking water treatment processes (p < 0.05); in particular, the relative abundance of Cyanobacteria increased after a super typhoon event (p < 0.05). The biological activated carbon (BAC) tank was identified as a potential pathogen reservoir and was responsible for 52 ± 6% of the bacteria released downstream, according to Bayesian-based SourceTracker analysis. Our results provide an insight into the challenges faced by maintaining finished water quality under changing weather conditions.

Novel helical or coiled flocculator for turbidity reduction in drinking water treatment: a performance study

Helical or spiral coiled flocculator have not been applied in drinking water treatment yet in Indonesia. There were only a few articles discussed it with different themes like hydrodynamic, floc characteristic, and performance. This study was done to know the efficiency (performance) of helical flocculator with parameters velocity gradient, pipe and helical diameter, flowrate, detention time, coagulant dose. The study was divided into two steps: Jar test to determine the optimum dose of coagulant and flocculation experiments to evaluate the helical flocculator efficiency. Efficiencies were in the range of medium to high. On flowrate 13 ml/second was obtained good results for two pipe sizes but different in helical diameters. In 0.5 inch pipe with 0.8 m helical diameter the turbidity reduction efficiencies were 72.4% and 73.9% and sediment volume were 18.3 ml and 20.0 ml. In 0.625 inch pipe with 0.4 m helical diameter the turbidity reduction efficiencies were 76.7% and 78.5% and sediment volume were 14.3 ml and 19.7 ml. The optimum velocity gradient about 64.9–69.6 persecond and detention time about 438–649 seconds. The results showed that helical flocculator was effective for floc formation. Flowrate, pipe diameter, helical diameter were three key parameters to perform helical flocculator.

River water turbidity removal using new natural coagulant aids: case study of Euphrates River, Iraq

For turbidity removal, most of drinking water treatment plants are using coagulants due to the presence of suspended and colloidal materials at the coagulation and flocculation units. Aluminium and sulphates salts are the widely used coagulants, such as Aluminium sulphate (Alum) and ferric chloride. However, several researches have linked Alzheimer's disease to the use of Aluminium sulphate. Hence, scholars have conducted several researches on the possibility to reduce the amount of Aluminium sulphate by using natural material/plants base as coagulant aids. In this study, Mallow's Leaves Extracts (MLE) and Carob's Pods Extracts (CPE) were used as an alternative coagulant aid. Couples of coagulation tests were implemented to find the optimal dosage of Aluminium Sulphates were used as coagulants. The results displayed that the maximum turbidity removal efficiency by adding 100% of each coagulant (i.e., Alum, MLE and CPE) were (61.16%, 51.175% and 37.12%), respectively. In addition, the minimum residual turbidity and maximum turbidity removal efficiency were 4.56 NTU and 97.72% by adding 22.5 Alum and 7.5 MLE presenting 30 mg/l dosing. Further, the minimum residual turbidity and maximum turbidity removal efficiency were 15.4 NTU and 92.3% by adding 22.5 Alum and 7.5 CPE presenting 30 mg/l dosing.