• Journal of Korean Society of Water and Wastewater
• /
• v.34 no.1
• /
• pp.35-43
• /
• 2020

Turbidity has various effects on the water quality and ecosystem of a river. High turbidity during floods increases the operation cost of a drinking water supply system. Thus, the management of turbidity is essential for providing safe water to the public. There have been various efforts to estimate turbidity in river systems for proper management and early warning of high turbidity in the water supply process. Advanced data analysis technology using machine learning has been increasingly used in water quality management processes. Artificial neural networks(ANNs) is one of the first algorithms applied, where the overfitting of a model to observed data and vanishing gradient in the backpropagation process limit the wide application of ANNs in practice. In recent years, deep learning, which overcomes the limitations of ANNs, has been applied in water quality management. LSTM(Long-Short Term Memory) is one of novel deep learning algorithms that is widely used in the analysis of time series data. In this study, LSTM is used for the prediction of high turbidity(>30 NTU) in a river from the relationship of turbidity to discharge, which enables early warning of high turbidity in a drinking water supply system. The model showed 0.98, 0.99, 0.98 and 0.99 for precision, recall, F1-score and accuracy respectively, for the prediction of high turbidity in a river with 2 hour frequency data. The sensitivity of the model to the observation intervals of data is also compared with time periods of 2 hour, 8 hour, 1 day and 2 days. The model shows higher precision with shorter observation intervals, which underscores the importance of collecting high frequency data for better management of water resources in the future.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2007.05a
• /
• pp.31-39
• /
• 2007

In this study we developed a turbidity management system to support the operation for effective turbid water management. The decision-making system includes various models for prediction of turbid water inflow, effective reservoir operation using the selective withdrawal facility, analysis of turbid water discharge in the downstream. The system is supported by the intensive monitoring devices installed in the upstream rivers, reservoirs, and downstream rivers. SWAT and HSPF models were constructed to predict turbid water flows in the Imha and Andong catchments. CE-QUAL-W2 models were constructed for turbid water behavior prediction, and various analyses were conducted to examine the effects of the selective withdrawal operation for efficient high turbid water discharge, turbid water distribution under differing amount and locations of turbid water discharge. A 1-dimensional dynamic water quality model was built using Ko-Riv1 for simulation of turbidity propagation in the downstream of the reservoirs, and 2-dimensional models were developed to investigate the mixing phenomena of two waters discharged from the Andong and Imha reservoirs with different temperature and turbidity conditions during joint dam operation for reducing the impacts of turbid water.

• International conference on construction engineering and project management
• /
• 2009.05a
• /
• pp.168-172
• /
• 2009

Shihmen reservoir was started in May 1963. The main purposes of Shihmen reservoir are for agriculture, power supply, flood control and tourism. Shihme Asn dam is an earth dam. Its crown height is 133m above mean sea level, with length 360 m, watershed 763.4 km², and maximum volume 309 million cms. Turbidity in Shihmen dam was severely affected by typhoons Aere (2004) and Masa (2005). Increased deposition after Aere was 28 million cms. Turbidity at Shihmen Canal Inlet is 3000 NTU (Nephelometry Turbidity Unit). Sediment sluicing strategies for downstream channel are demanded. Therefore, diversionary sediment preventing channel is planned in the upstream of Shihmen reservoir. Finally, turbid flow in tunnel channel is bypassed and diverted its flow down to downstream.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.05a
• /
• pp.263-264
• /
• 2021

The importance of evaluation and management of drinking water quality is emerging among the impacts of industrial changes and environmental destruction. Currently, the turbidity-related laws in Korea are regulated, and the low-concentration turbidity of 1.0 NTU or less must be maintained for process management, and control technology remains a necessary task. In this study, absorbance experiments according to turbidity were conducted using 470nm, 670nm, and 850nm, and water quality was measured using a light source of 850nm and a light receiving device of 820nm.

• Journal of Korean Society of Water and Wastewater
• /
• v.35 no.1
• /
• pp.83-91
• /
• 2021

The increased turbidity in rivers during flood events has various effects on water environmental management, including drinking water supply systems. Thus, prediction of turbid water is essential for water environmental management. Recently, various advanced machine learning algorithms have been increasingly used in water environmental management. Ensemble machine learning algorithms such as random forest (RF) and gradient boosting decision tree (GBDT) are some of the most popular machine learning algorithms used for water environmental management, along with deep learning algorithms such as recurrent neural networks. In this study GBDT, an ensemble machine learning algorithm, and gated recurrent unit (GRU), a recurrent neural networks algorithm, are used for model development to predict turbidity in a river. The observation frequencies of input data used for the model were 2, 4, 8, 24, 48, 120 and 168 h. The root-mean-square error-observations standard deviation ratio (RSR) of GRU and GBDT ranges between 0.182~0.766 and 0.400~0.683, respectively. Both models show similar prediction accuracy with RSR of 0.682 for GRU and 0.683 for GBDT. The GRU shows better prediction accuracy when the observation frequency is relatively short (i.e., 2, 4, and 8 h) where GBDT shows better prediction accuracy when the observation frequency is relatively long (i.e. 48, 120, 160 h). The results suggest that the characteristics of input data should be considered to develop an appropriate model to predict turbidity.

• Journal of Korea Water Resources Association
• /
• v.41 no.3
• /
• pp.293-303
• /
• 2008

Reservoir turbidity flows degrade the efficiency and sustainability of water supply system in many countries located in monsoon climate region. A decision support system called RTMMS aimed to assist reservoir operations was developed for the real time monitoring, modeling, and management of turbidity flows induced by flood runoffs in Daecheong reservoir. RTMMS consists of a real time data acquisition module that collects and stores field monitoring data, a data assimilation module that assists pre-processing of model input data, a two dimensional numerical model for the simulation of reservoir hydrodynamics and turbidity, and a post-processor that aids the analysis of simulation results and alternative management scenarios. RTMMS was calibrated using field data obtained during the flood season of 2004, and applied to real-time simulations of flood events occurred on July of 2006 for assessing its predictive capability. The system showed fairly satisfactory performance in reproducing the density flow regimes and fate of turbidity plumes in the reservoir with efficient computation time that is a vital requirement for a real time application. The configurations of RTMMS suggested in this study can be adopted in many reservoirs that have similar turbidity issues for better management of water supply utilities and downstream aquatic ecosystem.

• Journal of Korea Water Resources Association
• /
• v.49 no.1
• /
• pp.1-9
• /
• 2016

A correlation analysis between raw water turbidity at two wide-area water treatment plants and hydrological data was conducted for efficient water supply, design and management of water treatment plant. Both correlation analysis and principal component analysis were conducted using hydrological time series data such as inflow discharge, outflow discharge, and rainfall at dam basin of intake station of wide-area water treatment plants. And, forecasting of change in turbidity was conducted using regression equation for turbidity prediction. The raw water turbidity of two water treatment plants was strongly related to time series of discharge. The raw water turbidity of Chungju water treatment plant is strongly related to outflow discharge at Chungju dam (0.708). Whereas, the raw water turbidity of Wabu water treatment plant is strongly related to inflow discharge at Paldang dam (0.805). Similar trends between turbidity forecasting result using regression equation and calculation result using estimation equation on Korea water supply facilities standard were obtained. The result of this study can provide basic data for construction and management of water treatment plant.

• Ecology and Resilient Infrastructure
• /
• v.10 no.4
• /
• pp.107-115
• /
• 2023

High turbidity in source water can have adverse effects on water treatment plant operations and aquatic ecosystems, necessitating turbidity management. Consequently, research aimed at predicting river turbidity continues. This study developed a multi-class classification model for prediction of turbidity using LightGBM (Light Gradient Boosting Machine), a representative ensemble machine learning algorithm. The model utilized data that was classified into four classes ranging from 1 to 4 based on turbidity, from low to high. The number of input data points used for analysis varied among classes, with 945, 763, 95, and 25 data points for classes 1 to 4, respectively. The developed model exhibited precisions of 0.85, 0.71, 0.26, and 0.30, as well as recalls of 0.82, 0.76, 0.19, and 0.60 for classes 1 to 4, respectively. The model tended to perform less effectively in the minority classes due to the limited data available for these classes. To address data imbalance, the SMOTE (Synthetic Minority Over-sampling Technique) algorithm was applied, resulting in improved model performance. For classes 1 to 4, the Precision and Recall of the improved model were 0.88, 0.71, 0.26, 0.25 and 0.79, 0.76, 0.38, 0.60, respectively. This demonstrated that alleviating data imbalance led to a significant enhancement in Recall of the model. Furthermore, to analyze the impact of differences in input data composition addressing the input data imbalance, input data was constructed with various ratios for each class, and the model performances were compared. The results indicate that an appropriate composition ratio for model input data improves the performance of the machine learning model.

• Journal of Korean Society of Forest Science
• /
• v.101 no.3
• /
• pp.381-386
• /
• 2012

This study was carried out to clarify the turbidity change on three stands (Castanea crenata, Pinus densiflora and Plantation Land) by rainfall characteristics in small watershed. The change of turbidity showed in order of plantation land, Castanea stand and Pinus stand. The linear equations models between turbidity and rainfall intensity were able to account for 91% in Castanea stand, 80% in Pinus stand and 71% in plantation land. The linear equations models between turbidity and duration of rainfall were able to account for about 0-1% in three stands. The linear equations models between turbidity and preceding dry days were able to account for about 30% in three stands. The linear equations models between turbidity and accumulative rainfall were able to account for about 6-22% in three stands. The results indicates that soil runoff by land use and development of forest area could be applied to the mitigation measures such as afforestation and erosion check dam for erosion control and water quality management in small watershed.

• Journal of Environmental Impact Assessment
• /
• v.20 no.2
• /
• pp.227-234
• /
• 2011

In the study, CE-QUAL-W2 was used and its examination and correction were conducted targeting 2001 and 2003 when the condition of rainfall was contradicted. Using the proved model in 2003, a scenario was implemented with management of locations for dewatering outlets and actual data for dam management in 1987 when inflow and outflow level were almost same. In case of the scenario which the location of dewatering outlets was 5m higher than usual location, exclusion efficiency for turbid water inflow at the beginning of precipitation was good. In case of the scenario which the location of dewatering outlets was 10m lower than usual location, exclusion efficiency for excluding turbid water remained in a reservoir after the end of precipitation. However, the scenario applying dam management data in 1987, exclusion efficiency was relatively low. In the scenario, power-generating water release spot at EL.57m for first four days after the beginning of precipitation, EL.52m for 5th to 8th and EL.42m from 9th days. An analysis of the scenario reveals that both excessive days exceeded 30 NTU and average turbidity levels were decreased comparing before and after the alteration on outlets. The average turbidity levels were decreased by minimum of 55% to maximum of 70% and 30NTU exceeding days were decreased by 45 days at maximum. Also, since it could exclude most of turbid water in a reservoir before the destatifcation, the risk for turbid water evenly distributed in a reservoir along with turn-over could be decreased as well.