Search

This study focuses on the potential capability of numerous machine learning models, namely CatBoost, GradientBoosting, HistGradientBoosting, ExtraTrees, XGBoost, DecisionTree, Bagging, light gradient boosting machine (LGBM), GaussianProcess, artificial neural network (ANN), and light long short-term memory (LightLSTM). These models were investigated to predict the photocatalytic degradation of malachite green from wastewater using various NM-BiFeO

Capacitive deionization (CDI) is an alternative desalination technology that uses electrochemical ion separation. Although several attempts have been made to maximize the energy efficiency and productivity of CDI with conventional control methods, it is difficult to optimize the CDI processes because of the complex correlation between the operational conditions and the composition of feed water. To address these challenges, we applied deep reinforcement learning (DRL) to automatically control the membrane capacitive deionization (MCDI) process, which is one of the representative CDI processes, to accomplish high energy efficiency while desalinating water. In the DRL model, the numerical model is combined as the environment that provides states according to the actions. The feed water conditions, that is, the input state of the DRL, were assumed to have a random salt concentration and constant foulant concentration. The model was constructed to minimize energy consumption and maximize desalted water volume per cycle. After training of 1,000 episodes, the DRL model achieved a 22.07% reduction in specific energy consumption (from 0.054 to 0.042 kWh m

Enhancing engineering efficiency and reducing operating costs are permanent subjects that face all engineers over the world. To effectively improve the performance of filtration systems, it is necessary to determine an optimal operating condition beyond conventional methods of periodic and empirical operation. Herein, this paper proposes an effective approach to finding an optimal operating strategy using deep reinforcement learning (DRL), particularly for an ultrafiltration (UF) system. Deep learning was developed to represent the UF system utilizing a long-short term memory and provided an environment for DRL. DRL was designed to control three actions; operating pressure, cleaning time, and cleaning concentration. Ultimately, DRL proposed the UF system to actively change the operating pressure and cleaning conditions over time toward better water productivity and operating efficiency. DRL denoted ∼20.9% of specific energy consumption can be reduced by increasing average water flux (39.5-43.7 L m

Isolating dissolved organic matter (DOM) is a preliminary step that improves the accuracy of its characterization. In this study, DOM in brackish water was clearly separated and evaluated by multiple characterization analyses. The sample was divided into three fractions by preparative high-performance liquid chromatography (preparative HPLC) according to molecular size. The homogeneity of each fraction was estimated by analytical size exclusion chromatography (SEC) and fluorescence excitation-emission matrix (FEEM). Pyrolysis gas chromatography-mass spectrometry (Py-GC/MS) and liquid chromatography-organic carbon detection (LC-OCD) were used to characterize the physicochemical properties of each fraction. Py-GC/MS revealed that Fraction 1 consisted of evenly distributed organic matter in order polysaccharides, proteins, polyhydroxy aromatics, lignins, and lipids. However, Fraction 2 was primarily composed of dominant lipids and low portion of proteins, and Fraction 3 was composed predominantly of lignins and lipids. The LC-OCD results showed that Fractions 1 and 2 had similar organic carbon (OC) compositions: a humic substance (ca. 37%), building blocks (ca. 10%), and neutrals (ca. 37%), whereas Fraction 3 contained a high proportion of neutrals (62%). In the fouling experiments, the distinct DOM characteristics in each fraction resulted in different declining flux behaviors, ranked as: Fraction 2 > Fraction 1 > Fraction 3.

Resistance-in-series models have been applied to investigate fouling behavior. However, it is difficult to model the influence of morphology on fouling behavior because resistance is indirectly calculated from the water flux and transmembrane pressure. In this study, optical coherence tomography (OCT) was applied to evaluate the resistance of the fouling layer based on fouling morphology. Sodium alginate, humic acid, and bovine serum albumin (BSA) with high salts concentrations (conductivity: 23 mS/cm) were used as model foulants. At the same total fouling resistance, BSA showed the highest cake layer thickness (BSA (114.5 μm) > humic acid (53.5 μm) > sodium alginate (20.0 μm)). However, a different order was found for the cake layer resistance (BSA > sodium alginate > humic acid). This indicates that fouling thickness is not correlated with cake layer resistance. According to the Carman–Kozeny equation, fouling layer porosity decreased in the following order: humic acid (0.30) > BSA (0.21) > sodium alginate (0.20). In addition, we provided a specific value that was calculated using the ratio between the fouling thickness and cake layer resistance. The results show that alginic acid induced a stronger cake layer resistance, despite its thin fouling layer, whereas BSA showed a relatively low potential for inducing cake layer resistance. The results obtained in this study could be used for estimating and predicting fouling behavior.

We studied the influence of bioavailability of organic matter on membrane fouling layer development by comparing the filtration performance of two feed waters (wetland water and graywater). Dissolved organic carbon (DOC) concentration, size exclusion chromatography (SEC), and fluorescence excitation-emission matrix (FEEM) were used to characterize the bioavailability of organic matter in these water samples during the nanofiltration process. The wetland sample contained a high proportion of humic acid- and fulvic acid-like matter with low bioavailability, whereas the graywater sample comprised substantial amounts of aromatic proteins and microbial byproduct-like matter with high bioavailability. In addition, the molecular size distribution revealed that the wetland sample contained a large portion of recalcitrant organic matter, whereas the graywater sample contained easily bioavailable organic matter. After the filtration experiment, the DOC of the wetland sample decreased to 4.8 mgC/L, whereas the graywater sample resulted in a lower DOC concentration of 3.4 mgC/L. Optical coherence tomography (OCT) illustrated real-time variations in the fouling layer morphology, providing both 2D and 3D images. In addition, confocal laser scanning microscopy (CLSM) quantified the bacterial volume in the fouling layer. The wetland sample yielded a bacterial volume of 11.8 µm3/μm2 from a total fouling volume of 103 µm3/μm2, whereas the graywater sample yielded a bacterial volume of 53.2 µm3/μm2 from a total fouling layer volume of 134 µm3/μm2. Fitting of the two-phase Monod model to the fouling layer growth on the membrane resulted in lower-yield coefficients (i.e., the volumes produced per unit amount of substrate, Y xs ) of 7.46 and 27.95 µm3/μm2 in wetland water and higher-yield coefficients of 13.17 and 47.53 µm3/μm2 in the graywater at first and second phase, respectively. This study addresses the quantitative evaluation of the organic matter bioavailability in terms of membrane fouling using OCT images and a two-phase Monod model.

Practical modeling for assessing the efficiency of a full-scale reverse osmosis (RO) system may be a challenging task. This is because the operating conditions of RO systems can change significantly in actual practice owing to high seasonal variations and different progress of membrane fouling during long-term filtration. Accordingly, it is difficult to reliably model the RO performance if such conditions are excluded. In this study, we model a full-scale installation of a RO membrane system, considering actual operations of the industrial water treatment plant. A numerical model is built to describe spatiotemporal behavior of (water, salt, and foulant) mass transport inside a full-dimension pressure vessel. By performing a global sensitivity analysis, we evaluate the relative importance of key influential factors on model accuracy and specific energy consumption (SEC). The model and its parameters are optimized based on the sensitivity result and validated using best-fitted time-series measurement data of 3875 h. The results demonstrate the practical behaviors of fouling development and separation performance of the primary RO process. A regression tree analysis of SEC for 27 different operational scenarios in simulations may benefit decision making for energy efficient RO. Results reveal the high dependence of SEC on cleaning frequency in the feed temperature range.

The seawater battery (SWB) is a promising desalination technology that utilizes abundant sodium ions as an energy storage medium. Recently, the alternative desalination system, seawater battery desalination (SWB-D), was developed by placing an SWB next to the desalination compartment. This SWB-D system can desalt water while charging the SWB next to it. However, only a fixed catholyte solution has been investigated, although the catholytes impact the overall SWB-D performance. Therefore, we evaluated the effect of different catholytes on the desalination performance. High-saline reverse osmosis (RO) concentrate or brackish water exhibited excellent salt removal capability (>85.3% of sodium and >76.6% of chloride ions) with relatively short operation times (36.4 h for RO concentrate and 39.5 h for brackish water) upon charging, whereas the relatively low-saline river water showed the longest operation time (81.0 h), implying that river water should be excluded as a potential catholyte. The amount of desalinated water was marginally reduced due to osmosis through the anion exchange membrane; however, the amount of treated salt was >82.9% even after the reduction in water volume. These findings suggest that the catholyte with a resistance of >0.041 kΩ·cm can be ideal for the SWB-D.

Biofouling, a critical issue in membrane filtration, is influenced by several factors such as membrane characteristics and feed water quality. The organic matter (OM) composition is known to significantly influence biofilm formation, but few studies on this subject have been reported. Optical coherence tomography (OCT) allows direct monitoring of biofilm development on the membrane surface without the need for membrane autopsy. The purposes of the present study are 1) to quantify biofouling formation on a membrane surface using OCT; 2) to monitor the temporal variation of OM composition during membrane formation; and 3) to investigate the variation of OM composition. Substantial variations in thickness was observed from OCT images. It demonstrates that biovolume quantified from a single 2D cross-sectional image could result in inaccurate quantification. Here, we quantified foulants on the membrane surface using 3D images and validated the results using confocal laser scanning microscopy (CLSM). As well, variations in the OM composition driven by bacterial activity were observed by fluorescence excitation–emission matrix analysis and size exclusion chromatography. The biovolume estimated by the OCT system was 103.7 μm 3 /μm 2 . CLSM demonstrated that 10% of the foulants on the membrane surface was composed of live and dead bacteria.

This study examines South Korean print media coverage of two traditional Korean clan villages, Hahoe and Yangdong, both designated World Heritage Site (WHS) in 2010 after nomination to the tentative list in 2001. The study explores the representation of these villages by Korean mass media leading up, and after the process of WHS designation. Findings suggest that the evolution of articulations of heritage and tourism require an event or concern that disrupts the existing interests and needs, and calls upon social agents to consider future implications and opportunities. Some articulations remain for longer periods of time if they continue to serve the needs of the present. The sociocultural significance and implications of the findings are discussed.

Membrane capacitive deionization (MCDI) is a prospective desalination technology that removes ions using an electric potential difference across charged porous carbon electrodes. Natural organic matter (NOM) in feed water could influence the electrochemical process by leading to pore-blockages or forming a cake layer on the ion-exchange membrane-coated porous carbon electrode, thereby hindering ion removal. In this study, we explored the influence of different types of NOM, namely, humic acid (HA) and tannic acid (TA), on the MCDI desalination process for feed waters with inorganic salts (NaCl and CaCl2). HA significantly interfered with the adsorption process and reduced the salt removal rate by up to 68% in the case of NaCl-based feed water. However, the influence of HA on salt removal in the case of CaCl2-based feed water was marginal owing to the formation of a charge-neutralized complex, which was caused by the egg-box effect between Ca2+ and HA. TA reduced removal rates of salts (NaCl and CaCl2) by 37% and 60%, respectively. This is because of the lower molecular weight and smaller hydrodynamic diameter of TA relative to that of HA, owing to which TA exhibits a stronger adhesion to the electrode pore structure. Furthermore, as TA substantially reduces MCDI performance with regard to the adsorption of inorganic salts, its presence in feed water results in higher electrical resistance and energy consumption.

The development of a superior feed channel spacer is one among the problems that have to be resolved to effectively mitigate foulant accumulation in the membrane system and improve filtration performance. In view of this, a novel honeycomb-shaped spacer, whose hexagonal form is the most stable and economical structure observed in nature, is proposed. 3D printing enables us to manufacture the honeycomb-shape spacer. Then, the performance of such spacer is demonstrated by comparing its filtration results with those of a standard diamond-shaped spacer in nanofiltration. The proposed structure is observed to have higher fouling mitigation performance under various fouling conditions (i.e., low and high fouling potentials and different organic foulants). Optical coherence tomography demonstrates that the foulant layer formed by utilizing honeycomb-shaped spacers (119.0 μm) is significantly thinner than that resulting from the use of standard spacers (175.5 μm); thereby, the permeate flux of the honeycomb-shaped spacers was 16.0% greater than that of the standard spacers. Hydraulic cleaning tests reveal that honeycomb-shaped spacers have a higher potential for mitigating fouling resistances driven by the concentration polarization layer (Rcp) and cake layer (Rc), leading to higher permeate production than that generated by filtration using standard spacers. The use of computational fluid dynamics simulation affords better insights into the hydrodynamic effects of these spacers on the feed channel. It is observed that honeycomb-shaped spacers have superior performance that is attributable to the generation of high-magnitude turbulent kinetic energy in the areas enclosed by spacer filaments.

Typical reverse osmosis (RO) systems achieve a high level of performance in removing salt from feed water; however, they are relatively energy-intensive even for brackish water applications, due to the high water pressure required to overcome their high membrane resistance. Thus, we propose an energy-efficient hybrid system in which nanofiltration (NF) is sequentially coupled with membrane capacitive deionization (MCDI). The performance of the NF-MCDI hybrid system was demonstrated through experiments and modeling under various operating conditions. First, we experimentally evaluated the impacts of feed concentration, NF recovery rate, and MCDI flow rate on removal rate and energy consumption of the individual NF and MCDI processes. Then, a response surface methodology (RSM) analysis demonstrated the significant dependence of salt removal and energy consumption on the feed concentration, NF recovery, and MCDI flow rate within the NF-MCDI system. Although NF alone is insufficient to remove a large amount of salt from salty water, the proposed system configuration using MCDI was found to improve salt removal by up to 95%. Furthermore, it was found that the optimized NF-MCDI system (at ≤10 g/L feed concentration) can outperform conventional RO systems in terms of energy efficiency while meeting drinking water standards (≤0.5 g/L).

Water-stressed countries have been shifting their sources of clean water from inland freshwater to seawater. This led to a comprehensive exploration of seawater desalination processes to address water scarcity; however, membrane processes have expensive operational costs and high energy consumption. In this regard, this study presented a novel energy self-sufficient desalination system design that incorporates rechargeable seawater batteries as an additional energy storage system. Experimental data were projected using the reverse osmosis system analysis model to determine the configuration that achieved the lowest energy consumption and highest charging rate. The results show that the seawater battery achieved a satisfactory desalination performance with >90% and 74%−82% removal of sodium and chloride ions from actual water samples, respectively. Among the configurations, using ultrafiltration as pretreatment and applying 1.8 mA as initial current yielded the lowest energy consumption (1.35 kWh/m3) and the highest energy charging rate (1.01). Compared to the conventional reverse osmosis desalination plants (2.83 kWh/m3), the seawater battery-desalination system has a huge potential in addressing the major disadvantages of current desalination technologies.

Many studies have demonstrated a close relationship between fouling and fluid dynamics to control fouling formation. However, few have been carried out to validate the relationship in practice, owing to the lack of techniques to evaluate fouling with hydrodynamics. We investigated the organic fouling of humic acid (HA) and tannic acid (TA) at three crossflow velocities using a fluorescence-labelled organism bioimaging instrument (FOBI). Higher crossflow velocity generally enhanced filtration performance; however, the type of foulants induced different fouling formation patterns on the membrane. The monitoring system and computational model revealed that HA fouling had strong (R2 = 0.82) correlation with the shear rate (obtained from a velocity of 10.42 cm/s), while TA had weak correlation (R2 = 0.13). HA fouling showed a high correlation with velocity profile at 10 cm/s, but TA fouling occurred in the over 20 cm/s velocity zone. This is because TA had relatively high adhesive force acting to the membrane, resulting in irreversible fouling that was hardly mitigated by crossflow control. Therefore, this study demonstrated that fouling development was predominantly controlled by hydrodynamic conditions, but the actual fouling distribution could be influenced by the physicochemical properties of dissolved organic matters.

Mathematical models have been developed to obtain a better understanding of membrane fouling mechanisms. However, those models could not simulate the membrane fouling behaviors accurately because of the large number of fitting parameters related to feed water quality and flow pattern in a membrane filtration system. In this study, we developed a deep neural network (DNN) to model membrane fouling during nanofiltration (NF) and reverse osmosis (RO) filtration using in-situ fouling image data from optical coherence tomography (OCT). The performance of the DNN model was compared with that of existing mathematical models. In total, 13,708 high-resolution fouling layer images were used to develop the DNN model and validate the model performance. The DNN model was trained to simulate both organic fouling growth and flux decline, and it reproduced two- or three-dimensional images of the organic fouling growth. The DNN model demonstrated better predictive performance than the existing mathematical models. It achieved an R2 value of 0.99 and RMSE of 2.82 μm for the fouling growth simulation and R2 of 0.99 and RMSE of 0.30 Lm−2h−1 for the flux decline simulation. Therefore, the data-driven approach is an alternative way to model the membrane fouling and flux decline processes under high-pressure filtrations.

Because of recent advances in wireless communication technologies, the world of mobile computing is flourishing with a variety of applications. In this study, we present an integrated architecture for a personal digital assistant (PDA)-based mobile medical display system that supports collaborative work between remote users. We aim to develop a system that enables users in different regions to share a working environment for collaborative visualization with the potential for exploring huge medical datasets. Our system consists of three major components: mobile client, gateway, and parallel rendering server. The mobile client serves as a front end and enables users to choose the visualization and control parameters interactively and cooperatively. The gateway handles requests and responses between mobile clients and the rendering server for efficient communication. Through the gateway, it is possible to share working environments between users, allowing them to work together in computer supported cooperative work (CSCW) mode. Finally, the parallel rendering server is responsible for performing heavy visualization tasks. Our experience indicates that some features currently available to our mobile clients for collaborative scientific visualization are limited due to the poor performance of mobile devices and the low bandwidth of wireless connections. However, as mobile devices and wireless network systems are experiencing considerable elevation in their capabilities, we believe that our methodology will be utilized effectively in building quite responsive, useful mobile collaborative medical systems in the very near future.

While 2D texture mapping is one of the most effective of the rendering techniques that make 3D objects appear visually interesting, it often suffers from visual artifacts produced when 2D image patterns are wrapped onto the surfaces of objects with arbitrary shapes. On the other hand, 3D texture mapping generates highly natural visual effects in which objects appear carved from lumps of materials rather than laminated with thin sheets as in 2D texture mapping. Storing 3D texture images in a table for fast mapping computations, instead of evaluating procedures on the fly, however, has been considered impractical due to the extremely high memory requirement. In this paper, we present a new effective method for 3D texture mapping designed for real-time rendering of polygonal models. Our scheme attempts to resolve the potential texture memory problem by compressing 3D textures using a wavelet-based encoding method. The experimental results on various nontrivial 3D textures and polygonal models show that high compression rates are achieved with few visual artifacts in the rendered images and a small impact on rendering time. The simplicity of our compression-based scheme will make it easy to implement practical 3D texture mapping in software/hardware rendering systems including real-time 3D graphics APIs such as OpenGL and Direct3D.