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.

Separately tracking the sources of hydrophobic and hydrophilic dissolved organic matter during a storm event in an agricultural watershed.

The hydrophobicity of dissolved organic matter (DOM) affects various aspects of its environmental impacts in terms of water quality, sorption behaviors, interactions with other pollutants, and water treatment efficiency. In this study, source tracking of river DOM was conducted separately for hydrophobic acid (HoA-DOM) and hydrophilic (Hi-DOM) fractions using end-member mixing analysis (EMMA) in an agricultural watershed during a storm event. EMMA with optical indices of bulk DOM revealed larger contributions of soil (24 %), compost (28 %), and wastewater effluent (23 %) to riverine DOM under high versus low flow conditions. Molecular level analysis of bulk DOM revealed more dynamic features, showing an abundance of CHO and CHOS formulae in riverine DOM under high- and low flow conditions. CHO formulae originated from soil (78 %) and leaves (75 %) and contributed to the increasing CHO abundance during the storm event, whereas CHOS formulae likely originated from compost (48 %) and wastewater effluent (41 %). The characterization of bulk DOM at the molecular level demonstrated that soil and leaves are the dominant contributors for the high-flow samples. However, in contrast to the results of bulk DOM analysis, EMMA with HoA-DOM and Hi-DOM revealed major contributions from manure (37 %) and leaf DOM (48 %) during storm events, respectively. The results of this study highlight the importance of individual source tracking of HoA-DOM and Hi-DOM for the proper evaluation of the ultimate roles of DOM in affecting river water quality and for a better understanding of DOM dynamics and transformation in natural and engineered systems.

Effects of organic matter on the aggregation of anthropogenic microplastic particles in turbulent environments.

Biofilm-coated microplastics are omnipresent in aquatic environments, carrying different organic matter (OM) that in turn influences the flocculation and settling of microplastic aggregates. In this study, the effects of chitosan, guar gum, humic acid, and xanthan gum on the flocculation of anthropogenic microplastics are examined under controlled shear through the mixing chamber experiments. The results show that all of the selected OMs have positive effects on biofilm culturing and thus enhance the growth of microplastic flocs, with more evident promoting effects for cationic and neutral OMs (i.e., chitosan and guar gum) than anionic OMs (i.e., humic acid and xanthan). No critical shear rate is observed in the size vs. shear relationship based on our measurements. In addition, the quadrature-based two-class population balance model is employed to track the development of bimodal floc size distributions (FSDs) composed of small and large microplastic flocs. The model predictions show reasonable agreement with the observed FSDs. The largest error of settling flux from the two-class model is 7.8% in contrast with the reference value measured by the camera-based FSDs with 30 bins. This study highlights the role of different OMs on microplastic flocculation and indicates that a two-class model may be sufficient to describe microplastic transport processes in estuaries.

Flocculation kinetics and mechanisms of microalgae- and clay-containing suspensions in different microalgal growth phases.

Interplays between microalgae and clay minerals enhance biologically mediated flocculation, thereby affecting the sedimentation and transportation of suspended particulate matter (SPM) in water and benthic environments. This interaction forms larger flocs with a higher settling velocity and enhances SPM sinking. The aim of this study was to investigate the flocculation kinetics of microalgae and clay in suspension and to elucidate the mechanisms associated with such interactions. Standard jar test experiments were conducted using various mixtures of kaolinite and microalgal samples from batch cultures (Chlorella vulgaris) to estimate biologically mediated flocculation kinetics. The organic matter (OM) composition secreted by the microalgae was characterized using a liquid chromatography - organic carbon detection system, and quantitative analysis of transparent exopolymer particles was conducted separately. A two-class flocculation kinetic model, based on the interaction between flocculi and flocs, was also adopted to quantitatively analyze the experimental data from flocculation. Results from the flocculation kinetic tests and OM analyses, in association with other data analyses (i.e., floc size distribution and flocculation kinetic model), showed that flocculation increased with OM concentration during the growth phase (10-20 d). However, on day 23 during the early stationary phase, flocculation kinetics started decreasing and substantially declined on day 30, even though the amount of OM (mainly biopolymers) continued to increase. Our results indicate that an adequate quantity of biopolymers produced by the microalgal cells in the growth phase enhanced floc-to-floc attachment and hence flocculation kinetics. In contrast, an excessive quantity of biopolymers and humic substances in the stationary phase enhanced the formation of polymeric backbone structures and flocculation via scavenging particles but simultaneously increased steric stabilization with the production of a large number of fragmented particles.

Flocculation with heterogeneous composition in water environments: A review.

Flocculation is a key process for controlling the fate and transport of suspended particulate matter (SPM) in water environments and has received considerable attention in the field of water science (e.g., oceanography, limnology, and hydrology), remaining an active area of research. The research on flocculation has been conducted to elucidate the SPM dynamics and to diagnose various environmental issues. The flocculation, sedimentation, and transportation of SPM are closely linked to the compositional and structural properties of flocs. In fact, flocs are highly heterogeneous in terms of composition. However, the lack of comprehensive research on floc composition and structure has led to misconceptions regarding the temporal and spatial dynamics of SPM. This review summarizes the current understanding of the heterogeneous composition of flocs (e.g., minerals, organic matter, metals, microplastic, engineered nanoparticles) and its effect on their structure and on their fate and transport within aquatic environments. Furthermore, the effects of human activities (e.g., pollutant discharge, construction) on floc composition are discussed.

A tri-modal flocculation model coupled with TELEMAC for estuarine muds both in the laboratory and in the field.

Estuarine and coastal regions are often characterized by a high variability of suspended sediment concentrations in their waters, which influences dredging projects, contaminant transport, aquaculture and fisheries. Although various three-dimensional open source software are available to model the hydrodynamics of coastal water with a sediment module, the prediction of the fate and transport of cohesive sediments is still far from satisfied due to the lack of an efficient and robust flocculation model to estimate the floc settling velocity and the deposition rate. Single-class and sometimes two-class flocculation models are oversimplified and fail to examine complicated floc size distributions, while quadrature-based or multi-class based flocculation models may be too complicated to be coupled with large scale estuarine or ocean models. Therefore, a three-class population balance model was developed to track the sizes and number concentrations of microflocs, macroflocs and megaflocs, respectively. With the assumption of a fixed size of microflocs and megaflocs, only four tracers are needed when coupled with the open-source TELEMAC system. It enables better settling flux estimates and better addresses the occurrence and concentration of larger megaflocs. This tri-modal flocculation model was validated with two experimental data sets: (1) 1-D settling column tests with the Ems mud and (2) in-situ measurements at the WZ Buoy station on the Belgian coast. Results show that the flocculation properties of cohesive sediments can be reasonably simulated in both environments. It is also found that the number of macroflocs created, when a larger macrofloc breaks up, is a statistical mean value and may not be an integer when applying the model in the field.

The effects of the menstrual cycle on the static balance in healthy young women.

[Purpose] The purpose of this study was to investigate the effects of the menstrual cycle on the static balance of healthy young women. [Subjects and Methods] Eighteen healthy young subjects (mean age 19.1 years; weight 57.5 kg; height 159.9 cm) participated in this study. The Good Balance system was used to measure the postural sway speed and velocity moment of subjects in the static standing posture. Subjects were measured for static balance between 1 and 3 days after menstruation and 13 days after menstruation. [Results] The velocity moment of postural sway was significantly higher at 13 days after menstruation. [Conclusion] Our results indicate that the menstrual cycle affects the static balance of healthy subjects. During the menstrual cycle, intensity for balance exercises in females should be carefully controlled for injury prevention.

The effects of sitting with the right leg crossed on the trunk length and pelvic torsion of healthy individuals.

[Purpose] The purpose of this study was to determine the effects on the trunk length and pelvic torsion of healthy individuals that arise from crossing the right leg while sitting. [Subjects and Methods] The subjects in this study were 30 healthy individuals consisting of 18 males and 12 females. The subjects were instructed to sit on a chair, the height of which was adjustable, so that their knee and hip joints were bent at 90°. For the study, they sat stripped to the waist, with the back and hips bare. They were then instructed to perform a one-leg-crossed sitting posture by placing the right leg on the top of the left knee. A spinal posture test was performed to measure the subjects' trunk length and pelvic torsion by using a three-dimensional image-based spinal diagnostic system. [Results] The results of the three-dimensional spine examination showed statistically significant decreases in trunk length and pelvic torsion after the one-leg-crossed sitting posture. [Conclusion] In this study, the right leg-crossed sitting posture led to a decrease in the right trunk length with time and, in terms of pelvic torsion, increased the posterior rotation of the right pelvis when compared with the left pelvis.

The effects of horse riding simulation exercise with blindfolding on healthy subjects' balance and gait.

[Purpose] The study was conducted to determine the effect of horse riding simulation combined with blindfolding on healthy individuals' balance and gait. [Subjects and Methods] Thirty subjects were randomly divided into an experimental group (n=15) and a control group (n=15). The subjects in the experimental group covered their eyes using a blindfold, climbed onto a horse riding simulator, and performed the horse riding simulation exercise. The control group took part in the horse riding exercises without a blindfold. All of the subjects performed the 20 minutes long exercise once a day, five times a week, over a four-week period. [Results] The experimental group showed significant improvement in static balance, dynamic balance, velocity, and cadence compared to pre-intervention measurements. In addition, the control group showed significant improvement in static balance, dynamic balance, single support, and cadence compared to pre-intervention measurements. Significant differences in post-training gains in static balance, dynamic balance, and cadence were observed between the experimental group and the control group. [Conclusion] Subjects that performed horse riding simulation exercise after blindfolding showed significant improvements in balance and cadence compared to the control group.

Measurement and mathematical modelling of competition between fast- and slow-growing ordinary heterotrophic organisms in low and high substrate-loaded systems.

Ordinary heterotrophic organisms (OHO) of an activated sludge wastewater treatment system showed an atypical growth behaviour when they are inoculated to batch aerobic growth tests with a high substrate-loaded condition. For example, the OHO maximum specific growth rates on readily biodegradable substrates (µ H) increased with a high ratio of substrate concentration to OHO active biomass concentration (So/Xo), although they were assumed to be constant in a conventional microbial growth kinetic model with a single OHO population group. We, therefore, set a hypothesis in that the change of OHO maximum specific growth rates in the batch test condition is caused by turnover of fast-growing OHO population against slow-growing OHO population. And, a competitive microbial growth kinetic model of the fast- and slow-growing OHO population groups was developed and validated with model-data fitting analysis for the batch test results. The competitive microbial growth kinetic model of process selection, rather than that of kinetic selection, was capable of simulating microbial growth kinetics in high substrate-loaded dynamic systems (i.e., batch tests) and low substrate-loaded steady-state systems (i.e., continuously operated wastewater treatment systems), better than the conventional non-competitive growth kinetic model.

Competition between kaolinite flocculation and stabilization in divalent cation solutions dosed with anionic polyacrylamides.

Divalent cations have been reported to develop bridges between anionic polyelectrolytes and negatively-charged colloidal particles, thereby enhancing particle flocculation. However, results from this study of kaolinite suspensions dosed with various anionic polyacrylamides (PAMs) reveal that Ca(2+) and Mg(2+) can lead to colloid stabilization under some conditions. To explain the opposite but coexisting processes of flocculation and stabilization with divalent cations, a conceptual flocculation model with (1) particle-binding divalent cationic bridges between PAM molecules and kaolinite particles and (2) polymer-binding divalent cationic bridges between PAM molecules is proposed. The particle-binding bridges enhanced flocculation and aggregated kaolinite particles in large, easily-settleable flocs whereas the polymer-binding bridges increased steric stabilization by developing polymer layers covering the kaolinite surface. Both the particle-binding and polymer-binding divalent cationic bridges coexist in anionic PAM- and kaolinite-containing suspensions and thus induce the counteracting processes of particle flocculation and stabilization. Therefore, anionic polyelectrolytes in divalent cation-enriched aqueous solutions can sometimes lead to the stabilization of colloidal particles due to the polymer-binding divalent cationic bridges.

A two-class population balance equation yielding bimodal flocculation of marine or estuarine sediments.

Bimodal flocculation of marine and estuarine sediments describes the aggregation and breakage process in which dense microflocs and floppy macroflocs change their relative mass fraction and develop a bimodal floc size distribution. To simulate bimodal flocculation of such sediments, a Two-Class Population Balance Equation (TCPBE), which includes both size-fixed microflocs and size-varying macroflocs, was developed. The new TCPBE was tested by a model-data fitting analysis with experimental data from 1-D column tests, in comparison with the simple Single-Class PBE (SCPBE) and the elaborate Multi-Class PBE (MCPBE). Results showed that the TCPBE was the simplest model that is capable of simulating the major aspects of the bimodal flocculation of marine and estuarine sediments. Therefore, the TCPBE can be implemented in a large-scale multi-dimensional flocculation model with least computational cost and used as a prototypic model for researchers to investigate complicated cohesive sediment transport in marine and estuarine environments. Incorporating additional biological and physicochemical aspects into the TCPBE flocculation process is straight-forward also.

Evaluating alternative electrostatic potential models for polyacrylamide-co-acrylate in aqueous solution.

The capabilities of three simplified analytical equations to accurately model electrostatic interactions during proton binding and release by linear anionic polyelectrolytes in aqueous solution were evaluated. The impermeable sphere (IS), Donnan (DN), and cylindrical (CY) electrostatic models were fit to experimental acid-base titration curves of linear polyacrylamide-co-acrylate having ionizable site densities ranging from ca. 10-35%. The titrations were conducted in 0.003-0.12M NaCl solutions and the sum of squared errors from modeled and experimental data was used as a comparative index of each model's capability. In addition, the relative size of each polyelectrolyte was estimated from its measured specific viscosity and then compared against the values obtained from the fitting procedure for the size parameter that each model contained. Although the IS and DN electrostatic models could be used to obtain reasonably good fits to each titration curve, the size parameter values obtained by each model were not reflective of the actual polyelectrolyte sizes, indicating that the models had limited physical meaning and that the size parameter was essentially just an additional fitting parameter in each model. In contrast, the CY model was not only more effective in its ability to fit the titration data but also provided a better physical representation of the polyelectrolyte size. Therefore, for polyelectrolytes that remain essentially linear or are only loosely coiled such that counter ions are free to travel throughout the polymer structure, we conclude that the CY model and its morphological representation of a cylindrical polyelectrolyte are more valid and realistic than the IS and DN models and their representation of polyelectrolytes as spheres.