Algal blooms are significant environmental problems which threaten the water supply system and ecosystem. To manage the problem, the effective forecasting model is necessary, but it is still challenging to predict the algal bloom due to its uncertainty and complexity. To improve the prediction performance, this study proposed the advanced model based on LSTM networks. Merged-LSTM model contains the three parallel LSTM layers and merged layers which is available to use the additional data from the diverse sources without problem in the training process. To predict the chlorophyll-a of target area, data from an additional monitoring station in upstream and auxiliary environmental data were put into parallel layers as well as data from the target area. The prediction result of the proposed model outperforms the existing models, and also shows a better training process with larger data dimensions. The proposed model and its result also suggest that the possibility of prediction of algal bloom with more advanced models and corresponding data sources.
A genetic algorithm (GA) was combined with artificial neural networks (ANN), designated as neuro-genetic algorithm (NGA) in this study, to determine the effective number of nodes and optimal activated functions (FAs) in an ANN structure. Developed NGA was applied to predict Chlorophyll-a (Chl-a) concentrations in one-month increments in Lakes used as drinking water sources. Correlation analysis was used to setup input parameters. A simulation was conducted for four study sites with the most serious Chl-a problems in South Korea. Results from correlation analysis have indicated that phosphate phosphorus (PO4-P) and electrical conductivity showed high correlation with Chl-a, a factor not often considered in other studies. As the results of prediction of one-month forward Chl-a concentration, NGA showed high accuracy, with averaged determination coefficients of 0.89 and 0.84 in training and testing period, respectively. Double hidden layers showed better performance than a single hidden layer, while a logistic sigmoid function was frequently selected by the genetic algorithm in hidden layers in comparison with linear and hyperbolic tangent function. Practical uses for NGA in proactive water quality management are also discussed in this study.
1 June 2005 in Journal of Environmental Engineering
During crossflow microfiltration, formation of cake layer on the membrane inhibits the filter performance. Previous studies on cake formation have focused on the monodisperse conditions, although most field applications are under the polydisperse conditions. In this study, microfiltration experiments were conducted to investigate the effect of particle size on cake formation, especially, at polydisperse conditions. Kaolin and two kinds of CaC O3 were used as polydisperse particles and a polysulfon hollow fiber microfilter was used as the membrane for the experiment. To describe the effect of particle size on cake formation at polydisperse conditions, effective diameter for shear-induced diffusion (EDSD) was introduced as a representative particle size. Statistical analyses of the experimental results showed that EDSD can be used in quantifying the potential of cake formation at polydisperse conditions, and the larger the EDSD, the thinner the cake layer. EDSD was an imperative indicator which showed the p...
Recent water-related disasters have shown that not all disrupted events are prevented with water infrastructure systems and current water systems are becoming more vulnerable to disruptions due to the high uncertainty of disrupted events. Many scholars in various fields suggest diversification in the system as a way to respond to the uncertainty. In the real world, however, it is difficult to maximize its use, especially with water infrastructure, due to high costs and incomplete assessment methods. Thus this study attempts to develop a method to quantify cost-effectiveness of diversification using a drought case study in Korea. Modern Portfolio Theory is used to find optimal combinations of water resources infrastructures in terms of diversification. First, expected return and risk of individual water resources for water supply are estimated. Then, expected return and risk of individual portfolios of the water resources are evaluated by varying their shares of 0 to 100%. Finally, non-inferior portfolios are identified and an optimal portfolio for an acceptable return or risk is selected as a solution. Consequently, a portfolio is selected as a desirable one to practically enhance diversification in water infrastructure systems against real world uncertainty in consideration of cost and budget.
1 February 2010 in Journal of Hydraulic Engineering
Effects of bubble size, bubble diffusing area, and other parameters in air diffusers on destratification are studied using laboratory- and pilot-scale tanks with two layers. Dimensionless group involving such variable as bubble size, bubble diffusing area, and tank area are used to quantify these effects. Based on the results of experiments, a model is developed to predict destratification efficiency. Bubble diameter and overall tank area are found inversely related to destratification efficiency while bubble diffusing area is directly related to destratification efficiency.