Radial basis function artificial neural network able to accurately predict disinfection by-product levels in tap water: Taking haloacetic acids as a case study

Role of BP-ANN in simulating greenhouse gas emissions from global aquatic ecosystems via carbon component-environmental factor coupling

Inland waters (IW), estuarine areas (EA), and offshore areas (OA) function as aquatic systems in which the transport of carbon components results in the release of greenhouse gases (GHGs). Interconnected subsystems exhibit a greater greenhouse effect than individual systems. Despite this, there is a lack of research on how carbon loading and its components impact GHG emissions in various aquatic systems. In this study, we analyzed 430 aquatic sites to explore trade-off mechanisms among dissolved organic carbon (DOC), particulate organic carbon, dissolved inorganic carbon (DIC), and GHGs. The results revealed that IW emerged as the most significant GHG source, possessing a comprehensive global warming potential (GWP) of 0.78 ± 0.08 (10-2 Pg CO2-ep ha-1 year-1) for combined carbon dioxide, methane, and nitrous oxide. This surpassed the cumulative potentials of EA and OA (0.35 ± 0.05 (10-2 Pg CO2-ep ha-1 year-1)). Additionally, structural equation modeling indicated that GHG emissions resulted from a combination of carbon component loading and environmental factors. DOC exhibited a positive correlation with GWPs when influenced by biodegradable DOC. Total alkalinity and pH influenced DIC, leading to elevated pCO2 in aquatic systems, thereby enhancing GWPs. Predictive modeling using backpropagation artificial neural networks (BP-ANN) for GWPs, incorporating carbon components and environmental factors, demonstrated a good fit (R2 = 0.6078, RMSEaverage = 0.069, p > 0.05) between observed and predicted values. Enhancing the estimation of aquatic region feedback to GHG changes was achieved by incorporating corresponding water quality parameters. In summary, this study underscores the pivotal role of carbon components and environmental factors in aquatic regions for GHG emissions. The application of BP-ANN to estimate greenhouse effects from aquatic regions is highlighted, providing theoretical and experimental support for future advancements in monitoring and developing policies concerning the influence of water quality on GHG emissions.

Application of artificial intelligence in (waste)water disinfection: Emphasizing the regulation of disinfection by-products formation and residues prediction

Water/wastewater ((waste)water) disinfection, as a critical process during drinking water or wastewater treatment, can simultaneously inactivate pathogens and remove emerging organic contaminants. Due to fluctuations of (waste)water quantity and quality during the disinfection process, conventional disinfection models cannot handle intricate nonlinear situations and provide immediate responses. Artificial intelligence (AI) techniques, which can capture complex variations and accurately predict/adjust outputs on time, exhibit excellent performance for (waste)water disinfection. In this review, AI application data within the disinfection domain were searched and analyzed using CiteSpace. Then, the application of AI in the (waste)water disinfection process was comprehensively reviewed, and in addition to conventional disinfection processes, novel disinfection processes were also examined. Then, the application of AI in disinfection by-products (DBPs) formation control and disinfection residues prediction was discussed, and unregulated DBPs were also examined. Current studies have suggested that among AI techniques, fuzzy logic-based neuro systems exhibit superior control performance in (waste)water disinfection, while single AI technology is insufficient to support their applications in full-scale (waste)water treatment plants. Thus, attention should be paid to the development of hybrid AI technologies, which can give full play to the characteristics of different AI technologies and achieve a more refined effectiveness. This review provides comprehensive information for an in-depth understanding of AI application in (waste)water disinfection and reducing undesirable risks caused by disinfection processes.

Identification and verification of PCDD/Fs indicators from four typical large-scale municipal solid waste incinerations with large sample size in China

Monitoring PCDD/Fs emissions from municipal solid waste incinerations (MSWIs) is of paramount importance, yet it can be time-consuming and labor-intensive. Predictive models offer an alternative approach for estimating their levels. However, robust models specific to PCDD/Fs were lacking. In this study, we collected 190 PCDD/Fs samples from 4 large-scale MSWIs in China, with the average PCDD/Fs levels and TEQ levels of 0.987 ng/m3 and 0.030 ng TEQ/m3, respectively. We developed and evaluated predictive models, including traditional statistical methods, e.g., linear regression (LR) as well as machine learning models such as back propagation-artificial neural networks (BP ANN) and random forest (RF). Correlation analysis identified 2,3,4,7,8-PeCDF, 1,2,3,6,7,8-HxCDF, 2,3,4,6,7,8-HxCDF were better indicator congeners for PCDD/Fs estimation (R2 > 0.9, p < 0.001). The predictive results favored the RF model, exhibiting a high R2 value and low root mean square error (RMSE) and mean absolute error (MAE). Additionally, the RF model showed excellent prediction ability during external validation, with low absolute relative error (ARE) of 10.9 %-12.6 % for the three indicator congeners in the normal PCDD/F TEQ levels group (<0.1 ng TEQ/m3) and slightly higher ARE values (13.8 %-17.9 %) for the high PCDD/F TEQ levels group (>0.1 ng TEQ/m3). In conclusion, our findings strongly support the RF model's effectiveness in predicting PCDD/Fs TEQ emission from MSWIs.

The combination of multiple linear regression and adaptive neuro-fuzzy inference system can accurately predict trihalomethane levels in tap water with fewer water quality parameters

Artificial Neural Network (ANN) models are accurate in predicting the levels of disinfection by-products (DBPs) in drinking water. However, these models are not yet practical due to the large number of parameters involved, which should take a significant amount of time and cost to detect. Developing accurate and reliable prediction models of DBPs with fewest parameters is essential in the management of drinking water safety. This study used the adaptive neuro-fuzzy inference system (ANFIS) and radial basis function artificial neural network (RBF-ANN) to predict the levels of trihalomethanes (THMs), the most abundant DBPs in drinking water. Two water quality parameters identified by multiple linear regression (MLR) models were used as model inputs, and the quality of the models was assessed based on criteria such as correlation coefficient (r), mean absolute relative error (MARE), and the percentage of predictions with absolute relative error less than 25% (NE<25%) and over than 40% (NE>40%), etc. The results showed that the ANFIS models had higher correlation coefficients (r = 0.853-0.898) and prediction accuracy (NE<25% = 91%-94%) compared to RBF-ANN models (r = 0.553-0.819; NE<25% = 77%-86%) and traditional MLR models (r = 0.389-0.619; NE<25% = 67%-77%). Conversely, the prediction error, as indicated by MARE and NE>40%, showed the opposite trend: ANFIS models (MARE = 8%-11%; NE>40% = 0-5%) < RBF-ANN models (MARE = 15%-18%; NE>40% = 5%-11%) < MLR models (MARE = 19%-21%; NE>40% = 11%-17%). The present study provided a novel approach for constructing high-quality prediction models of THMs in water supply systems using only two parameters. This method holds promise as a viable alternative for monitoring THMs concentrations in tap water, thereby contributing to the improvement of water quality management strategies.

The mixed-order chlorine decay model with an analytical solution and corresponding trihalomethane generation model in drinking water

Ensuring effective drinking water disinfection, remaining a certain amount of residual chlorine, and controlling disinfection by-product formation were very important for guarantying water quality safety and protecting public health; thus, the chlorine decay model and corresponding disinfection by-product formation model were necessary. This paper proposed a mixed-order chlorine bulk decay model (two parameters) based on Taylor's formula and derived its analytical solution. The accuracy of the mixed-order model was evaluated by comparing it with the nth-order model. To optimize the model and reduce the number of parameters required to be calibrated, the relationship of parameters with temperature, initial chlorine concentration, TOC and inorganic substance (ammonia nitrogen and iodide ion) was explored. The result proved that one of the parameters could be regarded as temperature dependent only. Meanwhile, the temperature equation of the model parameters was established by the Arrhenius formula. Subsequently, this paper selected trihalomethane as the target and study the linear relationship between chlorine consumption and trihalomethane formation. The results indicated that the liner slope had little correlation with initial chlorine concentration and temperature. On this basis, the corresponding trihalomethane model was built and its performance was proven to be good. The modeling developed in this work could be applied to drinking water distribution systems for residual chlorine and trihalomethane prediction, and provided a reference for the decision involving water quality.

Artificial neural network implementation for dissolved organic carbon quantification using fluorescence intensity as a predictor in wastewater treatment plants

Although spectroscopic methods provide a fast and cost-effective means of monitoring dissolved organic carbon (DOC) in natural and engineered water systems, the prediction accuracy of these methods is limited by the complex relationship between optical properties and DOC concentration. In this study, we developed DOC prediction models using multiple linear/log-linear regression and feedforward artificial neural network (ANN) and investigated the effectiveness of spectroscopic properties, such as fluorescence intensity and UV absorption at 254 nm (UV₂₅₄), as predictors. Optimum predictors were identified based on correlation analysis to construct models using single and multiple predictors. We compared the peak-picking and parallel factor analysis (PARAFAC) methods for selecting appropriate fluorescence wavelengths. Both methods had similar prediction capability (p-values >0.05), suggesting PARAFAC was not necessary for choosing fluorescence predictors. Fluorescence peak T was identified as a more accurate predictor than UV₂₅₄. Combining UV₂₅₄ and multiple fluorescence peak intensities as predictors further improved the prediction capability of the models. The ANN models outperformed the linear/log-linear regression models with multiple predictors, achieving higher prediction accuracy (peak-picking: R² = 0.8978, RMSE = 0.3105 mg/L; PARAFAC: R² = 0.9079, RMSE = 0.2989 mg/L). These findings suggest the potential to develop a real-time DOC concentration sensor based on optical properties using an ANN for signal processing.

New methods based on a genetic algorithm back propagation (GABP) neural network and general regression neural network (GRNN) for predicting the occurrence of trihalomethanes in tap water

Monitoring trihalomethanes (THMs) levels in water supply systems is of great significance in ensuring drinking water safety. However, THMs detection is a time-consuming task. Developing predictive THMs models using parameters that are easier to obtain is an alternative. To date, there is still no application of optimization algorithms and general regression neural networks in predicting disinfection by-products levels. This study was to explore the feasibility of back propagation neural network (BPNN), genetic algorithm back propagation (GABP) neural network and general regression neural network (GRNN) for predicting THMs occurrence in real water supply systems. The results showed that the BPNN models' prediction ability was limited (test r_p = 0.571-0.857, N₂₅ = 61.5 %-91.5 %). Optimized by the genetic algorithm (GA), GABP models were generated and exhibited better prediction performance (test r_p = 0.573 and 0.696-0.863, N₂₅ = 68.2 %-93.6 %). However, GABP models took a lot of time and their prediction performance was unstable. A GRNN was then used to generate simpler neural network models, and the resulting prediction performance was excellent (total trihalomethanes and bromodichloromethane: test r_p = 0.657-0.824, N₂₅ = 81.8 %-100 %). In general, GRNN was the best at predicting THMs concentrations among the three models. However, it is worth noting that the prediction accuracy of bromodichloromethane (BDCM) was not high, which may be due to the absence of key parameters affecting BDCM formation. Accurate predictions of THMs by GRNN with these nine water parameters made THMs monitoring in real water supply systems possible and practical.

Membrane separation of antibiotics predicted with the back propagation neural network

Antibiotics and antibiotic resistance genes (ARGs) have been frequently detected in the aquatic environment and are regarded as emerging pollutants. The prediction models for the removal effect of four target antibiotics by membrane separation technology were constructed based on back propagation neural network (BPNN) through training the input and output. The membrane separation tests of antibiotics showed that the removal effect of microfiltration on azithromycin and ciprofloxacin was better, basically above 80%. For sulfamethoxazole (SMZ) and tetracycline (TC), ultrafiltration and nanofiltration had better removal effects. There was a strong correlation between the concentrations of SMZ and TC in the permeate, and the R2 of the training and validation processes exceeded 0.9. The stronger the correlation between the input layer variables and the prediction target was, the better the prediction performances of the BPNN model than the nonlinear model and the unscented Kalman filter model were. These results showed that the established BPNN prediction model could better simulate the removal of target antibiotics by membrane separation technology. The model could be used to predict and explore the influence of external conditions on membrane separation technology and provide a certain basis for the application of the BPNN model in environmental protection.

Appraisal of machine learning techniques for predicting emerging disinfection byproducts in small water distribution networks

Monitoring emerging disinfection byproducts (DBPs) is challenging for many small water distribution networks (SWDNs), and machine learning-based predictive modeling could be an alternative solution. In this study, eleven machine learning techniques, including three multivariate linear regression-based, three regression tree-based, three neural networks-based, and two advanced non-parametric regression techniques, are used to develop models for predicting three emerging DBPs (dichloroacetonitrile, chloropicrin, and trichloropropanone) in SWDNs. Predictors of the models include commonly-measured water quality parameters and two conventional DBP groups. Sampling data of 141 cases were collected from eleven SWDNs in Canada, in which 70 % were randomly selected for model training and the rest were used for validation. The modeling process was reiterated 1000 times for each model. The results show that models developed using advanced regression techniques, including support vector regression and Gaussian process regression, exhibited the best prediction performance. Support vector regression models showed the highest prediction accuracy (R² =0.94) and stability for predicting dichloroacetonitrile and trichloropropanone, and Gaussian process regression models are optimal for predicting chloropicrin (R² =0.92). The difference is likely due to the much lower concentrations of chloropicrin than dichloroacetonitrile and trichloropropanone. Advanced non-parametric regression techniques, characterized by a probabilistic nature, were identified as most suitable for developing the predictive models, followed by neural network-based (e.g., generalized regression neural network), regression tree-based (e.g., random forest), and multivariate linear regression-based techniques. This study identifies promising machine learning techniques among many commonly-used alternatives for monitoring emerging DBPs in SWDNs under data constraints.

Prediction of monthly precipitation using various artificial models and comparison with mathematical models

Precipitation (PP) prediction is an interesting topic in the meteorology or hydrology field since it is directly related to agriculture, the management of water resources in hydrologic basins, and water scarcity. Selecting the right model to predict precipitation has always been a challenge because it could help researchers to use the proper model for their purposes. Accordingly, the performance of five artificial models (feed-forward neural network, cascade forward neural network, Elman neural network, multi-layer perceptron neural network, and radial basis neural network) and three mathematical models (Poisson regression model (PRM), quadratic model, and multiple linear regression) were evaluated for their ability to predict the monthly precipitation in Mediterranean coastal cities located in Eastern part of Mediterranean Sea for the first time. Twenty-seven Mediterranean coastal cities are considered case studies. For this aim, scenario 1 and scenario 2 with various input variables are proposed. Scenario 1 is developed using the number of months (MN), maximum temperature (Tmax), minimum temperature (Tmin), downward radiation (DR), wind speed (WS), vapor pressure (VP), and actual evapotranspiration (AE). Scenario 2 is developed by adding geographical coordinates (latitude, longitude, and altitude) to the global meteorological data to see the impact of geographical coordinates on the accuracy of the prediction of monthly precipitation. This study utilized the monthly data, which were obtained from TerraClimate for the period from 2010 to 2021. Based on the performance indexes, the PRM model performed best for the prediction of monthly precipitation in all selected locations compared to other models. Moreover, the results indicate that scenario 2 ([Formula: see text]) has shown higher prediction accuracy compared to scenario 1 ([Formula: see text]). In conclusion, PRM with the combination of [[Formula: see text]] had RMSE value that was lower by 12% relative to PRM with the combination of [[Formula: see text]]. Consequently, the PRM model can be recommended for modeling the complexity of interactions for precipitation-climate conditions-geographical coordinates and predicting precipitation.

Recent advances in shelf life prediction models for monitoring food quality

Each year, 1.3 billion tons of food is lost due to spoilage or loss in the supply chain, accounting for approximately one third of global food production. This requires a manufacturer to provide accurate information on the shelf life of the food in each stage. Various models for monitoring food quality have been developed and applied to predict food shelf life. This review classified shelf life models and detailed the application background and characteristics of commonly used models to better understand the different uses and aspects of the commonly used models. In particular, the structural framework, application mechanisms, and numerical relationships of commonly used models were elaborated. In addition, the study focused on the application of commonly used models in the food field. Besides predicting the freshness index and remaining shelf life of food, the study addressed aspects such as food classification (maturity and damage) and content prediction. Finally, further promotion of shelf life models in the food field, use of multivariate analysis methods, and development of new models were foreseen. More reliable transportation, processing, and packaging methods could be screened out based on real-time food quality monitoring.

Model for halo-acetic acids formation in bulk water of water supply systems

With increased understanding of the differences in toxicity between species of haloacetic acids (HAAs) and the possibility of more stringent regulations, the ability to predict individual HAA species formation is important. Nine different haloacetic acids are regulated and their total concentration is referred to as HAA9. A mathematical model to predict concentrations of HAA species was proposed and tested using independent data sets. The amount of HAA9 formed per unit amount of chlorine consumed (μg-HAA9/mg-consumed chlorine) remained constant throughout the reaction times in each sample. Similarly, the fraction of a given HAA species largely remained constant during most of the reaction time. Thus, each HAA species was assumed to have its own yield with respect to consumed chlorine in a given water sample. The parallel second-order (2R) model describing chlorine decay kinetics was then extended to predict HAA species formation kinetics. The combined chlorine and HAA species model closely predicts all tested HAA species and its sum with standard error ≤ 5 μg/L. Within the tested waters having Cl₂/N mass ratio ≥ 10.7 (g-Cl₂/g-N), ammonia did not impact the mass yield. The mass yield of each HAA species can be calculated from three measurements (e.g. at 0, 4 and 24 h) of HAA species and chlorine. Once the yield is known, HAA species concentrations could be predicted for up to 120 h with only chlorine measurements. The model extends the previous work of predicting the trihalomethane species formation kinetics to HAA species formation kinetics. Further research is needed to understand how the yield varies with source water quality, treatment and in distribution systems.

Insight into the Behavior of Mortars Containing Glass Powder: An Artificial Neural Network Analysis Approach to Classify the Hydration Modes

In this paper, an artificial neural network (ANN) model is proposed to predict the hydration process of a new alternative binder. This model overcomes the lack of input parameters of physical models, providing a realistic explanation with few inputs and fast calculations. Indeed, four mortars are studied based on ordinary Portland cement (CEM I), cement with limited environmental impact (CEM III), and glass powder (GP) as the cement substitution. These mortars are named CEM I + GP and CEM III + GP. The properties of the mortars are characterized, and their life cycle assessment (LCA) is established. Indeed, a decrease in porosity is observed at 90 days by 4.6%, 2.5%, 12.4%, and 7.9% compared to those of 3 days for CEMI, CEMIII, CEMI + GP, and CEMIII + GP, respectively. In addition, the use of GP allows for reducing the mechanical strength in the short term. At 90 days, CEMI + GP and CEMIII + GP present a decrease of about 28% and 57% in compressive strength compared to CEMI and CEMIII, respectively. Nevertheless, strength does not cease increasing with the curing time, due to the continuous pozzolanic reactions between Ca(OH)₂ and silica contained in GP and slag present in CEMIII as demonstrated by the thermo-gravimetrical (TG) analysis. To summarize, CEMIII mortar provides similar performance compared to mortar with CEMI + GP in the long term. This can later be used in the construction sector and particularly in prefabricated structural elements. Moreover, the ANN model used to predict the heat of hydration provides a similar result compared to the experiment, with a resulting R² of 0.997, 0.968, 0.968, and 0.921 for CEMI, CEMIII, CEMI + GP, and CEMIII + GP, respectively, and allows for identifying the different hydration modes of the investigated mortars. The proposed ANN model will allow cement manufacturers to quickly identify the different hydration modes of new binders by using only the heat of hydration test as an input parameter.

Multi-input multi-output temporal convolutional network for predicting the long-term water quality of ocean ranches

The prediction of water quality parameters is of great significance to the control of marine environments and provides a scientific decision-making basis for maintaining the stability of water environments and ensuring the normal survival and growth of marine aquatic products. However, the water quality in ocean ranches is affected by the complex, dynamic, and changeable environments of open water, which have complex nonlinear relationships, poor accuracy, high time complexity, and poor long-term predictability. Therefore, in this paper, a multi-input multi-output end-to-end prediction model based on a temporal convolutional network (MIMO-TCN) is proposed to predict water quality. A ConvNeXt module and TCN module were used as the model encoder and decoder, respectively. ConvNeXt was used to extract the features of the input data, and the TCN used the extracted feature data to achieve improved prediction accuracy. The model adds skip connections between its modules to solve the gradient disappearance problem as the number of network layers increases. To prove the effectiveness of the proposed method, a model robustness and prediction ability evaluation was conducted in this paper based on the dissolved oxygen in multiple ocean pasture validation samples. Compared with other learning models, the mean absolute error (MAE), root mean square error (RMSE), and mean absolute percentage error (MAPE) of the MIMO-TCN prediction results were reduced by 60.77%, 30.88%, and 52.45% on average, respectively, and the R² improved by 6.07% on average over those of other models. The experimental results show that the proposed method has higher forecasting accuracy than competing approaches.

Prediction of suicidal ideation among Chinese college students based on radial basis function neural network

Background

Suicide is one of the leading causes of death for college students. The predictors of suicidal ideation among college students are inconsistent and few studies have systematically investigated psychological symptoms of college students to predict suicide. Therefore, this study aims to develop a suicidal ideation prediction model and explore important predictors of suicidal ideation among college students in China.

Methods

We recruited 1,500 college students of Sichuan University and followed up for 4 years. Demographic information, behavioral and psychological information of the participants were collected using computer-based questionnaires. The Radial Basis Function Neural Network (RBFNN) method was used to develop three suicidal ideation risk prediction models and to identify important predictive factors for suicidal ideation among college students.

Results

The incidence of suicidal ideation among college students in the last 12 months ranged from 3.00 to 4.07%. The prediction accuracies of all the three models were over 91.7%. The area under curve scores were up to 0.96. Previous suicidal ideation and poor subjective sleep quality were the most robust predictors. Poor self-rated mental health has also been identified to be an important predictor. Paranoid symptom, internet addiction, poor self-rated physical health, poor self-rated overall health, emotional abuse, low average annual household income per person and heavy study pressure were potential predictors for suicidal ideation.

Conclusions

The study suggested that the RBFNN method was accurate in predicting suicidal ideation. And students who have ever had previous suicidal ideation and poor sleep quality should be paid consistent attention to.

Water Quality Prediction Based on SSA-MIC-SMBO-ESN

Water pollution threatens the safety of human production and life. To quickly respond to water pollution, it is important for water management staff to predict water quality changes in advance. Drawing on the temporality of water quality data, the leaky integrator echo state network (ESN) was introduced to construct the water quality prediction models for dissolved oxygen (DO), permanganate index (CODMn), and total phosphorus (TP), respectively. First, the missing values were filled by the linear trend method of adjacent points, and the outliers were detected and corrected by the Z-score method and the linear trend method. Second, the singular spectrum analysis (SSA) was performed to denoise the original monitoring data, such that the predicted data catch up with the real data, and the model accuracy is not affected by the hidden noise in the data. Third, the correlation between water quality indices was measured by the maximum information coefficient (MIC), and the strongly correlated indices were imported to the prediction model. Finally, according to these strong correlation indicators, the water quality prediction models based on multiple features were constructed, respectively, using the offline and online learning algorithms of the ESN. The hyperparameters of the models were optimized through the sequential model-based optimization (SMBO). Experimental results show that the proposed water quality prediction models, namely, SSA-MIC-SMBO-Offline ESN and SSA-MIC-SMBO-Online ESN, predicted DO, CODMn, and TP accurately, providing suitable tools for practical applications.

Precursor characteristics of mono-HAAs during chlorination and cytotoxicity of mono-HAAs on HEK-293T cells

Monohaloacetic acids (mono-HAAs), a class of disinfection by-products widely occurred in drinking water, receives significant attention due to their extremely high toxicity. Many studies on the biological toxicity of mono-HAAs have been reported, yet the toxic effects of mono-HAAs on human renal cells (kidney is one of the target organs for disinfection by-products) has not been involved. Studies on organic precursors for mono-HAAs formation were also very limited due to their lower levels as compared to di-HAAs and tri-HAAs. Based on this, the formation of mono-HAAs after chlorination of some typical source water samples and their relationship with water quality parameters were investigated. Meanwhile, the cytotoxicity of monochloroacetic acid (MCAA), monobromoacetic acid (MBAA), and monoiodoacetic acid (MIAA) were tested using human embryonic kidney cells (HEK-293 T cells). The results showed that the levels of mono-HAAs formed during chlorination of source water samples were between 0.44 and 0.87 μg/L. Formation of MBAA positively (p < 0.05) correlated with bromide ion and dissolved organic carbon, but negatively (p < 0.01) correlated with SUVA₂₅₄ (specific UV absorbance at 254 nm), while formation of MCAA was only positively (p < 0.05) related with SUVA₂₅₄. These results suggested that although MCAA and MBAA both belong to the mono-HAAs, the characteristics of their organic precursors differ significantly. MCAA precursors have high aromaticity and are more hydrophobic, yet MBAA precursors have low aromaticity and are more hydrophilic. The half-lethal concentrations (LC₅₀) of MCAA, MBAA, and MIAA on HEK293T cells were 1196-1211 μM, 16.07-18.96 μM, and 6.08-6.17 μM, respectively. An in-depth analysis showed that the cytotoxicity of mono-HAAs on HEK 293 T cells could not be explained by the parameters concerning cellular uptake (e.g., logP and pKa), but the S_N2 reaction of C-X bond with cellular molecules (e.g., glyceraldehyde-3-phosphate dehydrogenase, etc) may be the relevant cause for the cytotoxicity of mono-HAAs on HEK 293 T cells.

Sensor Data Fusion as an Alternative for Monitoring Chlorate in Electrochlorination Applications

As chlorate concentrations have been found to be harmful to human and animal health, governments are increasingly demanding strict control of the chlorate concentration in drinking water. Since there are no chlorate sensors available, the current solution is sampling and laboratory analysis. This is costly and time consuming. The aim of this work was to investigate Sensor Data Fusion (SDF) as an alternative approach, with a focus on chlorate formation in the electrochlorination process, and design an observer for the real-time estimation of chlorate. The pH, temperature and UV-a absorption were measured in real time. A reduced-order nonlinear model was derived, and it was found to be detectable. An Extended Kalman Filter (EKF), based on this model, was then used to estimate the chlorate formation. The EKF algorithm was verified experimentally and was found to be capable of accurately estimating chlorate concentrations in real time. Electrochlorination is an emerging and efficient method of disinfecting drinking water. Soft sensing of chlorate concentrations, as proposed in this paper, may help to better control and manage the process of electrochlorination.

A Review on Machine Learning, Artificial Intelligence, and Smart Technology in Water Treatment and Monitoring

Artificial-intelligence methods and machine-learning models have demonstrated their ability to optimize, model, and automate critical water- and wastewater-treatment applications, natural-systems monitoring and management, and water-based agriculture such as hydroponics and aquaponics. In addition to providing computer-assisted aid to complex issues surrounding water chemistry and physical/biological processes, artificial intelligence and machine-learning (AI/ML) applications are anticipated to further optimize water-based applications and decrease capital expenses. This review offers a cross-section of peer reviewed, critical water-based applications that have been coupled with AI or ML, including chlorination, adsorption, membrane filtration, water-quality-index monitoring, water-quality-parameter modeling, river-level monitoring, and aquaponics/hydroponics automation/monitoring. Although success in control, optimization, and modeling has been achieved with the AI methods, ML models, and smart technologies (including the Internet of Things (IoT), sensors, and systems based on these technologies) that are reviewed herein, key challenges and limitations were common and pervasive throughout. Poor data management, low explainability, poor model reproducibility and standardization, as well as a lack of academic transparency are all important hurdles to overcome in order to successfully implement these intelligent applications. Recommendations to aid explainability, data management, reproducibility, and model causality are offered in order to overcome these hurdles and continue the successful implementation of these powerful tools.

Identification of environmental microorganism using optimally fine-tuned convolutional neural network

To not only optimize the hyper-parameters of the classification layer of dense convolutional network with 201 convolutional layers (DenseNet-201) but also use data augmentation processes could enhance the performance of DenseNet-201, and DenseNet-201 is rarely applied to the identifications of the environmental microorganism (EM) images. Hence, this study was to propose the optimally fine-tuned DenseNet-201 (OFTD) with data augmentation to better classify the EM images on Environmental Microorganism Dataset (EMDS). The training dataset was composed of 70% Environmental Microorganism Dataset (EMDS) images and so was mainly used to fit the parameters of convolutional layers of optimally fine-tuned DenseNet-201 (OFTD). Meanwhile, the other EMDS images were considered as the testing dataset and used to qualify the performance of the OFTD. Also, gradient-weighted class activation mapping method (Grad-CAM) was adopted to visually illustrate the dominant features of the EM images. Based on the results, the OFTD model with data augmentation achieved the highest classification accuracy of 98.4%. In this case, so its stability and accuracy were guaranteed. Besides, the optimally fine-tuned classification layer is considered a more efficient method than the data augmentation technique adopted in this study when it comes to the improvement of the performance in DenseNet-201 implemented on EMDS. Grad-CAM highlighted the coarse EM features identified effectively by the OFTD; for example, foot and stalk were considered as the dominated features of Rotifera and Vorticella, respectively. In summary, the proposed OFTD with data augmentation could provide an efficient solution for the EM detection in digital microscope.

Efficient reductive and oxidative decomposition of haloacetic acids by the vacuum-ultraviolet/sulfite system

Haloacetic acids (HAAs), as a representative category of halogenated disinfection byproducts, are widely detected in disinfected water. In this work, the vacuum ultraviolet (VUV)/sulfite process under N₂ saturated conditions was proposed to eliminate a series of HAAs (i.e., monochloroacetic acid (MCAA), difluoroacetic acid (DFAA), trifluoroacetic acid (TFAA), dichloroacetic acid (DCAA), etc.). The in situ generated hydrated electron (e_aq^-) demonstrated to be the main species to fulfill the initial degradation and dechlorination of MCAA, while hydroxyl radicals (˙OH) were in charge of the mineralization of MCAA. This means that the VUV/sulfite system is a combination of advanced reduction and oxidation processes (ARPs and AOPs). A significant enhancement of MCAA removal was observed with increasing pH values from 6.0 to 10.0, and surprisingly, k_obs correlated well with the proportion of SO₃^2- as the pH changed. This can be explained by the production of e_aq^- from VUV irradiation of SO₃^2- rather than HSO₃^- and also due to e_aq^- being more stable under alkaline conditions. Increasing the sulfite dosage also elevated the degradation of MCAA. However, the addition of certain anions (i.e., chloride (Cl^-), bicarbonate (HCO₃^-), and nitrate (NO₃^-)) and dissolved organic matter (DOM) inhibited the removal of MCAA to varying degrees. The VUV/sulfite system was effective toward various types of halogenated disinfection byproducts, supporting its broad applicability. Nevertheless, even in real waters, the VUV/sulfite system was also promising for the simultaneous abatement of HAAs and other oxyanions.

New insights into the role of molecular structures on the fate and behavior of antibiotics in an osmotic membrane bioreactor

Osmotic membrane bioreactors (OMBRs) have been applied to enhance removal of antibiotics, however, information on the effects of molecular structures on the behavior of antibiotics is still lacking. Herein, adsorption kinetics, transformation pathways, and membrane rejection mechanisms of OMBRs were investigated by adding two typical antibiotics (i.e., sulfadiazine, SDZ, and tetracycline hydrochloride, TC-HCl). 80.70-91.12% of TC-HCl was removed by adsorption and biodegradation, while 17.50-75.14% of SDZ was removed by membrane rejection; this depended on its concentration due to reduced electrostatic interactions and hydrophobic adsorption. The adsorption capacity of TC-HCl (i.e., 1.34±0.01 mg/g) was significantly higher than that of SDZ (i.e., 0.18±0.03 mg/g) due to enhanced π-π interactions, hydrogen bonding and improved electrostatic interactions. The abundant production of polysaccharide-like substances from TC-HCl biodegradation contributed to microbial metabolism and thus enhanced microbial function during TC-HCl biotransformation. The primary degradation pathways were determined by microbial function analysis, and the primary intermediates from TC-HCl degradation were less toxic than those from SDZ degradation due to the different reactions of amino groups. These results and the corresponding mechanism provide a theoretical foundation for the further development of OMBR technology for highly efficient treatment of antibiotic wastewater.

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.

Using simple and easy water quality parameters to predict trihalomethane occurrence in tap water

Monitoring of disinfection by-products (DBPs) in water supply system is important to ensure safety of drinking water. Yet it is a laborious job. Developing predictive DBPs models using simple and easy parameters is a promising way. Yet current models could not be well applied into practice because of the improper dataset (e.g. not from real tap water) they used or involving the parameters that are difficult to measure or require expensive instruments. In this study, four simple and easy water quality parameters (temperature, pH, UVA₂₅₄ and Cl₂) were used to predict trihalomethane (THMs) occurrence in tap water. Linear/log linear regression models (LRM) and radial basis function artificial neural network (RBF ANN) were adopted to develop the THMs models. 64 observations from tap water samples were used to develop and test models. Results showed that only one or two parameters entered LRMs, and their prediction ability was very limited (testing datasets: N₂₅ = 46-69%, r_p = 0.334-0.459). Different from LRM, the prediction accuracy of RBF ANNs developed with pH, temperature, UVA₂₅₄ and Cl₂ can be improved continuously by tweaking the maximum number of neuron (MN) and Gaussian function spread (S) until it reached best. The optimum RBF ANNs of T-THMs, TCM and BDCM were obtained when setting MN = 20, S = 100, 100.1 and 60, respectively, where the N₂₅ and r_p values for testing datasets reached 85-92% and 0.813-0.886, respectively. Accurate predictions of THMs by RBF ANNs with these four simple and easy parameters paved an economic and convenient way for THMs monitoring in real water supply system.

Effect of different fish feeds on water quality and growth of crucian carp (Carassius carassius) in the presence and absence of prometryn

Few data are available regarding comprehensive or quantitative assessment of fish feed considering both the environmental and feeding impacts. Aiming to fill the gap, an experimental study to investigate the effects of three fish feeds on concentrations of nutrients and crucian carp (Carassius carassius) growth was conducted in laboratory aquariums in the presence and absence of prometryn. Results showed that weight gain rates of crucian carp treated with Tong Wei (TW) feed were 106.3% and 2.0% higher than that of Zhong Shan (ZS) and Zhong Liang (ZL) feeds, a possible explanation was that the quality of protein in TW feed was highest as evidenced by the protein efficiency ratios. Meanwhile, TW feed posed relatively lighter effects on water qualities (between ZL and ZS). Prometryn significantly inhibited the growth of crucian carp and thus affected concentrations of nutrients in water indirectly. The relationships between weight gain rates of fish and concentrations of nutrients in water (R² = 0.929-0.990) were developed. In sum, this study suggested that it is realizable to obtain better fish growth performance with lesser degrading effects on water qualities by producing and selecting appropriate feed regardless of prometryn existence, and the developed equations could be used as a basis for future studies.

Prediction of tuberculosis cases based on sociodemographic and environmental factors in gombak, Selangor, Malaysia: A comparative assessment of multiple linear regression and artificial neural network models

Background

Early prediction of tuberculosis (TB) cases is very crucial for its prevention and control. This study aims to predict the number of TB cases in Gombak based on sociodemographic and environmental factors.

Methods

The sociodemographic data of 3325 TB cases from January 2013 to December 2017 in Gombak district were collected from the MyTB web and TB Information System database. Environmental data were obtained from the Department of Environment, Malaysia; Department of Irrigation and Drainage, Malaysia; and Malaysian Metrological Department from July 2012 to December 2017. Multiple linear regression (MLR) and artificial neural network (ANN) were used to develop the prediction model of TB cases. The models that used sociodemographic variables as the input datasets were referred as MLR1 and ANN1, whereas environmental variables were represented as MLR2 and ANN2 and both sociodemographic and environmental variables together were indicated as MLR3 and ANN3.

Results

The ANN was found to be superior to MLR with higher adjusted coefficient of determination (R²) values in predicting TB cases; the ranges were from 0.35 to 0.47 compared to 0.07 to 0.14, respectively. The best TB prediction model, that is, ANN3 was derived from nationality, residency, income status, CO, NO₂, SO₂, PM₁₀, rainfall, temperature, and atmospheric pressure, with the highest adjusted R² value of 0.47, errors below 6, and accuracies above 96%.

Conclusions

It is envisaged that the application of the ANN algorithm based on both sociodemographic and environmental factors may enable a more accurate modeling for predicting TB cases.

Artificial neural network modeling of full-scale UV disinfection for process control aimed at wastewater reuse

Accurate modeling of wastewater ultraviolet disinfection is fundamental as support for process optimization and control. Detailed modeling of hydrodynamics and fluence rate via computational fluid dynamics, coupled to laboratory studies of inactivation kinetics, are usually the preferred approach for UV disinfection modeling. Despite this approach often provides accurate predictive performance, it requires significantly high computational time, making it unfeasible for real-time process control. In this study, to enable an effective process control, black-box regression models were assessed as a modeling alternative for UV disinfection, synthesizing hydrodynamics, fluence rate and inactivation kinetics. UV disinfection of a full-scale wastewater treatment plant in Italy was monitored for 10 months, measuring influent and effluent E. coli concentration, turbidity, absorbance at 254 nm, temperature and flow rate at different UV doses. Considering the usually observed distribution of effluent E. coli concentration and the zero inflation of the collected dataset, Poisson, zero-inflated Poisson and Hurdle generalized linear models were tested, as well as two-part models coupling a classifier describing the E. coli zero-count events and a regressor estimating the magnitude of E. coli concentrations in positive-count events. The two-part artificial neural network model showed the best predictive performance, being able of both describing nonlinearities and handling the high proportion of null values in the dataset. The deployment of this model to control ultraviolet disinfection was simulated, estimating a plausible 63% energy saving.

Radial basis function-artificial neural network (RBF-ANN) for simultaneous fluorescent determination of cysteine enantiomers in mixtures

The determination of chiral compounds is critically important in chemical and pharmaceutical sciences. Cysteine amino acid is one of the important chiral compounds where each enantiomer (L and D) has different effects on fundamental physiological processes. The unique optical properties of nanoparticles make them a suitable probe for the determination of different analytes. In this work, the water-soluble thioglycolic acid (TGA)-capped cadmium-telluride (CdTe) quantum dots (QDs) were applied as optical nanoprobe for the simultaneous determination of cysteine enantiomers. The difference in the kinetics of the interactions between L- and D-cysteine with CdTe QDs is used for multivariate quantitative analysis. Multivariate methods are superior to univariate methods in determining the concentration of each enantiomer in the mixture without the information about the total chiral analyte concentration. As a nonlinear calibration method the radial basis function -artificial neural network (RBF-ANN) model was more successful in predicting L-and D-cysteine concentrations than the linear partial least squares regression (PLS) model.

Biological mechanism of alleviating membrane biofouling by porous spherical carriers in a submerged membrane bioreactor

In this study, porous spherical carriers were fixed around the hollow fiber membrane module to mitigate membrane biofouling. Two MBRs (R1 without carriers, R2 with carriers) were operated for 31 days under identical operating conditions to investigate the effects of the carriers on the reactor performances, the production of extracellular polymeric substances (EPS), the level of N-acyl-homoserine lactones (AHLs), and the microbial communities. The results showed that the presence of carriers in MBR was conducive to nitrogen removal and decreased the total membrane filtration resistance by about 1.7 times. Slower transmembrane pressure (TMP) rise-up, thinner bio-cakes, lower EPS production, and fewer tryptophan and aromatic proteins substances on the membrane surface were observed in R2. The polysaccharides secretion of EPS in bio-cakes was mainly regulated by C4-HSL and 3OC6-HSL in the presence of carriers. The microbial community analysis revealed that carriers addition reduced the relative abundance of EPS and AHL producing bacteria in the membrane bio-cakes and enriched the accumulation of functional bacteria conducive to nutrient removal in the mixed liquor. This study provided an in-depth understanding for the application of porous spherical carriers to alleviate membrane biofouling.

Machine learning in natural and engineered water systems

Water resources of desired quality and quantity are the foundation for human survival and sustainable development. To better protect the water environment and conserve water resources, efficient water management, purification, and transportation are of critical importance. In recent years, machine learning (ML) has exhibited its practicability, reliability, and high efficiency in numerous applications; furthermore, it has solved conventional and emerging problems in both natural and engineered water systems. For example, ML can predict various water quality indicators in situ and real-time by considering the complex interactions among water-related variables. ML approaches can also solve emerging pollution problems with proven rules or universal mechanisms summarized from the related research. Moreover, by applying image recognition technology to analyze the relationships between image information and physicochemical properties of the research object, ML can effectively identify and characterize specific contaminants. In view of the bright prospects of ML, this review comprehensively summarizes the development of ML applications in natural and engineered water systems. First, the concept and modeling steps of ML are briefly introduced, including data preparation, algorithm selection and model evaluation. In addition, comprehensive applications of ML in recent studies, including predicting water quality, mapping groundwater contaminants, classifying water resources, tracing contaminant sources, and evaluating pollutant toxicity in natural water systems, as well as modeling treatment techniques, assisting characterization analysis, purifying and distributing drinking water, and collecting and treating sewage water in engineered water systems, are summarized. Finally, the advantages and disadvantages of commonly used algorithms are analyzed according to their structures and mechanisms, and recommendations on the selection of ML algorithms for different studies, as well as prospects on the application and development of ML in water science are proposed. This review provides references for solving a wider range of water-related problems and brings further insights into the intelligent development of water science.

High flux and complete dyes removal from water by reduced graphene oxide laminate on Poly Vinylidene Fluoride/graphene oxide membranes

Dyes molecules are the most common pollutants of wastewater in the environment from the textile industry to numbers of technologies include dyeing, printing, and painting procedures. Among membrane-based separation approaches as established methods in the water treatment industry, polymers attracted massive attention in the production of membranes due to their low cost and high-performance filtration of pollutants. However, hydrophobicity and low speed of filtration along with limited decontamination performance against some of the dyes, demand new approaches and membranes to overcome drawbacks points. Herein, a new design introduced including a support layer made by Poly Vinylidene Fluoride (PVDF)/Graphene Oxide (PGO) composite membrane via immersion precipitation process and a thin layer (≤100 nm) of reduced graphene oxide (rGO) deposited (as an active layer) through a simple vacuum filtration method. It has been observed that the presence of the GO sheets in the PGO composite improved the hydrophilicity of the membrane, water flux (from ~90 L m^-2 h^-1 bar^-1 in pristine PVDF to ~1690 L m^-2 h^-1 bar^-1 in PGO), and anti-fouling property. By deposition of rGO laminate on PGO support, dyes separation as high as ~99% can be achieved for most of the cationic and anionic dyes due to electrostatic adsorption, π-π interactions and molecular sieving. This approach opens new insight on hybrid designs for graphene-polymers based membrane toward efficient and fast removal of pollutants from wastewater.

Trihalomethanes in Water Supply System and Water Distribution Networks

The formation of trihalomethanes (THMs) in natural and treated water from water supply systems is an urgent research area due to the carcinogenic risk they pose. Seasonal effects and pH have captured interest as potential factors affecting THM formation in the water supply and distribution systems. We investigated THM occurrence in the water supply chain, including raw and treated water from water treatment plants (coagulation, sedimentation, sand filtration, ClO₂-disinfection processes, and distribution pipelines) in the Chiang Mai municipality, particularly the educational institute area. The effects of two seasons, rainy (September–November 2019) and dry (December 2019–February 2020), acted as surrogates for the water quality profile and THM occurrence. The results showed that humic acid was the main aromatic and organic compound in all the water samples. In the raw water sample, we found a correlation between surrogate organic compounds, including SUVA and dissolved organic carbon (DOC) (R² = 0.9878). Four species of THMs were detected, including chloroform, bromodichloromethane, dibromochloromethane, and bromoform. Chloroform was the dominant species among the THMs. The highest concentration of total THMs was 189.52 μg/L. The concentration of THMs tended to increase after chlorination when chlorine dioxide and organic compounds reacted in water. The effect of pH on the formation of TTHMs was also indicated during the study. TTHM concentrations trended lower with a pH ≤ 7 than with a pH ≥ 8 during the sampling periods. Finally, in terms of health concerns, the concentration of TTHMs was considered safe for consumption because it was below the standard (<1.0) of WHO’s Guideline Values (GVs).

Integrated neural network model with pre-RBF kernels

To improve the network performance of radial basis function (RBF) and back-propagation (BP) networks on complex nonlinear problems, an integrated neural network model with pre-RBF kernels is proposed. The proposed method is based on the framework of a single optimized BP network and an RBF network. By integrating and connecting the RBF kernel mapping layer and BP neural network, the local features of a sample set can be effectively extracted to improve separability; subsequently, the connected BP network can be used to perform learning and classification in the kernel space. Experiments on an artificial dataset and three benchmark datasets show that the proposed model combines the advantages of RBF and BP networks, as well as improves the performances of the two networks. Finally, the effectiveness of the proposed method is verified.

New methods based on back propagation (BP) and radial basis function (RBF) artificial neural networks (ANNs) for predicting the occurrence of haloketones in tap water

Haloketones (HKs) is one class of disinfection by-products (DBPs) which is genetically toxic and mutagenic. Monitoring HKs in drinking water is important for drinking water safety, yet it is a time-consuming and laborious job. Developing predictive models of HKs to estimate their occurrence in drinking water is a good alternative, but to date no study was available for HKs modeling. This study was to explore the feasibility of linear, log linear regression models, back propagation (BP) as well as radial basis function (RBF) artificial neural networks (ANNs) for predicting HKs occurrence (including dichloropropanone, trichloropropanone and total HKs) in real water supply systems. Results showed that the overall prediction ability of RBF and BP ANNs was better than linear/log linear models. Though the BP ANN showed excellent prediction performance in internal validation (N₂₅ = 98-100%, R² = 0.99-1.00), it could not well predict HKs occurrence in external validation (N₂₅ = 62-69%, R² = 0.202-0.848). Prediction ability of RBF ANN in external validation (N₂₅ = 85%, R² = 0.692-0.909) was quite good, which was comparable to that in internal validation (N₂₅ = 74-88%, R² = 0.799-0.870). These results demonstrated RBF ANN could well recognized the complex nonlinear relationship between HKs occurrence and the related water quality, and paved a new way for HKs prediction and monitoring in practice.