• Journal of Korean Society of Water and Wastewater
• /
• v.35 no.6
• /
• pp.417-424
• /
• 2021

Algal bloom is an ongoing issue in the management of freshwater systems for drinking water supply, and the chlorophyll-a concentration is commonly used to represent the status of algal bloom. Thus, the prediction of chlorophyll-a concentration is essential for the proper management of water quality. However, the chlorophyll-a concentration is affected by various water quality and environmental factors, so the prediction of its concentration is not an easy task. In recent years, many advanced machine learning algorithms have increasingly been used for the development of surrogate models to prediction the chlorophyll-a concentration in freshwater systems such as rivers or reservoirs. This study used a light gradient boosting machine(LightGBM), a gradient boosting decision tree algorithm, to develop an ensemble machine learning model to predict chlorophyll-a concentration. The field water quality data observed at Daecheong Lake, obtained from the real-time water information system in Korea, were used for the development of the model. The data include temperature, pH, electric conductivity, dissolved oxygen, total organic carbon, total nitrogen, total phosphorus, and chlorophyll-a. First, a LightGBM model was developed to predict the chlorophyll-a concentration by using the other seven items as independent input variables. Second, the time-lagged values of all the input variables were added as input variables to understand the effect of time lag of input variables on model performance. The time lag (i) ranges from 1 to 50 days. The model performance was evaluated using three indices, root mean squared error-observation standard deviation ration (RSR), Nash-Sutcliffe coefficient of efficiency (NSE) and mean absolute error (MAE). The model showed the best performance by adding a dataset with a one-day time lag (i=1) where RSR, NSE, and MAE were 0.359, 0.871 and 1.510, respectively. The improvement of model performance was observed when a dataset with a time lag up of about 15 days (i=15) was added.

• Journal of Korea Multimedia Society
• /
• v.23 no.6
• /
• pp.747-755
• /
• 2020

In this paper, we deal with outlier data problems that occur when constructing a PM_2.5 fine dust forecasting system using a neural network. In general, when learning a neural network, some of the data are not helpful for learning, but rather disturbing. Those are called outlier data. When they are included in the training data, various problems such as overfitting occur. In building a PM_2.5 fine dust concentration forecasting system using neural network, we have found several outlier data in the training data. We, therefore, remove them, and then make learning 3 ways. Over_outlier model removes outlier data that target concentration is low, but the model forecast is high. Under_outlier model removes outliers data that target concentration is high, but the model forecast is low. All_outlier model removes both Over_outlier and Under_outlier data. We compare 3 models with a conventional outlier removal model and non-removal model. Our outlier removal model shows better performance than the others.

• Journal of Korean Society of Water and Wastewater
• /
• v.22 no.3
• /
• pp.389-396
• /
• 2008

The hybrid model can be used to predict the initial near field mixing and the far field transport of the buoyant jets, which are discharged from the submerged wastewater ocean outfall. In the near field, the jet integral model can be used for single port diffusers while the ${\sigma}$ transformed particle tracking model was used in the far field. In this study, the experimental study was performed to verify the developed hybrid model in the previous research. The developed hybrid model properly predict the surface and vertical concentration distribution of the single buoyant jets with various effluent and ambient conditions. The hybrid model can also simulate the surface concentration distribution of the rosette diffuser except for the parallel diffuser with the higher densimetric Froude number due to the assumption that dynamic effects of the effluent plumes are negligible in the far field. The application of the hybrid model to rosette diffusers can predict the concentration near the diffuser more accurately when the line-plume approximation is used.

• Environmental Engineering Research
• /
• v.14 no.2
• /
• pp.111-119
• /
• 2009

A prototype GIS-based decision support system (DSS) was developed by using a database management system (DBMS), a model management system (MMS), a knowledge-based system (KBS), a graphical user interface (GUI), and a geographical information system (GIS). The method of selecting a dispersion model or a modeling scheme, originally devised by Park and Seok, was developed using our GIS-based DSS. The performances of candidate models or modeling schemes were evaluated by using a single index(statistical score) derived by applying fuzzy inference to statistical measures between the measured and predicted concentrations. The fumigation dispersion model performed better than the models such as industrial source complex short term model(ISCST) and atmospheric dispersion model system(ADMS) for the prediction of the ground level $SO_2$ (1 hr) concentration in a coastal area. However, its coincidence level between actual and calculated values was poor. The neural network models were found to improve the accuracy of predicted ground level $SO_2$ concentration significantly, compared to the fumigation models. The GIS-based DSS may serve as a useful tool for selecting the best prediction model, even for complex terrains.

• Atmosphere
• /
• v.22 no.1
• /
• pp.73-85
• /
• 2012

The collision efficiency data for collision between graupel or hail particles and cloud drops that take into account the differences of particle density are applied to the Takahashi cloud model. The original setting assumes that graupel or hail collision efficiency is the same as that of the cloud drops of the same volume. The Takahashi cloud model is run with the new collision efficiency data and the results are compared with those with the original. As an initial condition, a thermodynamic profile that can initiate strong convection is provided. Three different CCN concentration values and therefore three initial cloud drop spectra are prescribed that represent maritime (CCN concentration = 300 $cm^{-3}$), continental (1000 $cm^{-3}$) and extreme continental (5000 $cm^{-3}$) air masses to examine the aerosol effects on cloud and precipitation development. Increase of CCN concentration causes cloud drop sizes to decrease and cloud drop concentrations to increase. However, the concentration of ice particles decreases with the increase of CCN concentration because small drops are difficult to freeze. These general trends are well captured by both model runs (one with the new collision efficiency data and the other with the original) but there are significant differences: with the new data, the development of cloud and raindrop formation are delayed by (1) decrease of ice collision efficiency, (2) decrease of latent heat from riming process and (3) decrease of ice crystals generated by ice multiplication. These results indicate that the model run with the original collision efficiency data overestimates precipitation rates.

• Journal of The Korean Society of Agricultural Engineers
• /
• v.46 no.7
• /
• pp.35-44
• /
• 2004

In this study, a ventilation model was developed to determine a ventilation rate for the balance of heat, moisture and $CO_{2}$ in a mushroom house. Internal and external temperature, relative humidity and $CO_{2}$ concentration were measured and used to validate the ventilation model. The effects of various environmental factors on physiological responses of mushroom were also investigated. The verified model was simulated under the observed ventilation rates with a difference of$ 0.001{\~}0.065\;m^{3}{\cdot}S^{-1}$ (relative error of $0.3{\~}18.9\%$) when external temperature varied 22.5 to $24.8^{circ}C$ and average ventilation rates was $0.35m^{3}{\cdot}S^{-1}$. The optimal conditions for mushroom growth (internal temperature $22 ^{circ}C$, relative humidity $80\%$, $CO_{2}$ concentration 1,000 ppm) were used for the model application with external temperature, relative humidity and $CO_{2}$ concentration of $27.5{\~}33.5^{circ}C$, $60\%$, and 355 ppm, respectively. Thermal balance was a important factor for an optimum ventilation up to the external temperature of $32^{circ}C$, while $CO_{2}$ concentration balance was more important over $32^{circ}C$. This suggests that humidification for moisture balance is required to maintain temperature and $CO_{2}$ concentration at an optimal level by ventilation in a mushroom house.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.29 no.1B
• /
• pp.73-81
• /
• 2009

This paper presents a numerical investigation of the impact of the secondary currents on solute transport in open-channel flows. The RANS model with Reynolds stress model is used for flow modeling, and the GGDH(generalized gradient diffusion hypothesis) model is used to close the scalar transport equation. Using the developed model, the impact of secondary currents on solute transport in open channel flows over smooth-rough strip is investigated. Through numerical experiments, the secondary currents are found to affect the solute spreading, leading a movement of the position of the peak concentration and a skewed distribution of solute concentration. Due to the lateral flow of secondary currents near the free surface, the concentration at the rough strip is found to be larger than that at the smooth strip bed. The solute at the rough strip is more rapidly transported than smooth bed. A magnitude analysis of the solute transport rate in scalar transport equation is also carried out to investigate the effect of secondary currents and scalar flux on the concentration distribution.

• Journal of Korean Society of Water and Wastewater
• /
• v.31 no.5
• /
• pp.421-430
• /
• 2017

In this study, a model was developed to predict for Disinfection By-Products (DBPs) generated in water supply networks and consumer premises, before and after the introduction of advanced water purification facilities. Based on two-way ANOVA, which was carried out to statistically verify the water quality difference in the water supply network according to introduce the advanced water treatment process. The water quality before and after advanced water purification was shown to have a statistically significant difference. A multiple regression model was developed to predict the concentration of DBPs in consumer premises before and after the introduction of advanced water purification facilities. The prediction model developed for the concentration of DBPs accurately simulated the actual measurements, as its coefficients of correlation with the actual measurements were all 0.88 or higher. In addition, the prediction for the period not used in the model development to verify the developed model also showed coefficients of correlation with the actual measurements of 0.96 or higher. As the prediction model developed in this study has an advantage in that the variables that compose the model are relatively simple when compared with those of models developed in previous studies, it is considered highly usable for further study and field application. The methodology proposed in this study and the study findings can be used to meet the level of consumer requirement related to DBPs and to analyze and set the service level when establishing a master plan for development of water supply, and a water supply facility asset management plan.

• Archives of Pharmacal Research
• /
• v.13 no.3
• /
• pp.221-226
• /
• 1990

A physiological pharmacokinetic hybrid model was developed in order to predict the disposition kinetics of diphenylhydantoin (DPH) in the brain from the plasma conentration data of DPH. The model was constructed under the assumptions of well-stirred, plasma flow-limited and lienar tissue diposition kinetics of DPH. DPH was administered intravenously to the rats at a dose of 10 mg/kg together with/without sodium salicylate (SA;10 mg/kg) and the DPH concentrations in the plasma and brain were determined. Plasma protein binding of DPH concentrations in the plasma and brain were determined. Plasma protein binding of DPH was also determined using equilibrium dialysis technique. Then the model was tested for its predictability of DPH concentrations in the brian from the plasma data of DPH. It was found that the predicted values of DPH concentrations in the brian were in fair agreement with the experimental values in the rats of both treatments. The 2-fold increase in the brain concentration of DPH by SA-coadinistration was predicted well from the plasma concentration and plasma free fraction ($f_p$) data of DPH using the model. Therefore, the hybrid model was concluded to be very useful for the prediction of the concentrations of DPH in the brain from the plasma concentration data. Finally, DPH concentrations in the human brian was calculated using this model from plasma DPH data in the literature, yet the scale-up of this model to the human is not convinced.

• Proceedings of the KSME Conference
• /
• 2001.06e
• /
• pp.136-141
• /
• 2001

In this study, we analysed tree surface flow by using the experimental and numerical method with a different surfactant concentration. We compared numerical solution with experimental results for one-dimensional model. The result shows that in general the tree surface velocity can well be reproduced by the one-dimensional model for various surfactant concentration.