• Journal of Soil and Groundwater Environment
• /
• v.21 no.3
• /
• pp.21-34
• /
• 2016

In the present study, we applied various machine learning techniques comparatively for prediction of subsurface structures based on multiple secondary information (i.e., well-logging data). The machine learning techniques employed in this study are Naive Bayes classification (NB), artificial neural network (ANN), support vector machine (SVM) and logistic regression classification (LR). As an alternative model, conventional hidden Markov model (HMM) and modified hidden Markov model (mHMM) are used where additional information of transition probability between primary properties is incorporated in the predictions. In the comparisons, 16 boreholes consisted with four different materials are synthesized, which show directional non-stationarity in upward and downward directions. Futhermore, two types of the secondary information that is statistically related to each material are generated. From the comparative analysis with various case studies, the accuracies of the techniques become degenerated with inclusion of additive errors and small amount of the training data. For HMM predictions, the conventional HMM shows the similar accuracies with the models that does not relies on transition probability. However, the mHMM consistently shows the highest prediction accuracy among the test cases, which can be attributed to the consideration of geological nature in the training of the model.

• Proceedings of the Korean Radioactive Waste Society Conference
• /
• 2009.06a
• /
• pp.186-186
• /
• 2009

The characterization strategy of fracture networks are classified into a deterministic or statistical characterization according to the type of required information. A deterministic characterization is most efficient for a sparsely fractured system, while the statistics are sufficient for densely fractured rock. In this study, the ensemble mean and variability of the effective connectivity is systematically analyzed with various density values for different network structures of a power law size distribution. The results of high resolution Monte Carlo analyses show that statistical characteristics can be a necessary information to determine the transport properties of a fracture system when fracture density is greater than a percolation threshold. When the percolation probability (II) approaches unity with increasing fracture density, the effective connectivity of the network can be safely estimated using statistics only (sufficient condition). It is inferred from conditional simulations that deterministic information for main pathways can reduce the uncertainty in estimation of system properties when the network becomes denser. Overall results imply that most pathways need to be identified when II < 0.5 statistics are sufficient when II $\rightarrow$ 1 and statistics are necessary and the identification of main pathways can significantly reduce the uncertainty in estimation of transport properties when 0.5$\ll$1. It is suggested that the proper estimation of the percolation probability of a fracture network is a prerequisite for an appropriate conceptualization and further characterization.

• 한국지구물리탐사학회:학술대회논문집
• /
• 2003.11a
• /
• pp.676-683
• /
• 2003

A laboratory experiment using artificial fracture rocks was used to understand the 3-dimensional dispersion of a tracer and the mixing process in a fractured network. In this experiment, 12cm polystyrene foam cubes with two electrodes for monitoring electric conductivity (EC) were used as artificial fractured rocks. Distilled water with 0.5mS/m was used as a tracer in water with 35mS/m and the difference of EC between the tracer and the water was monitored by a multipoint simultaneous measurement system of electrical resistance. The results showed that even if the fracture arrangement pattern was not straight in the direction of the flow, the tracer did not diffuse along individual fractures and an oval tracer plume, which was the distribution of tracer concentrations, tended to be form in the direction of the flow. The vertical cross section of the tracer distribution showed small diffusivity in the vertical direction. The calculated total tracer volume passing through each measurement point in the horizontal cross section showed while that the solute passed through measurement points near the direction of hydraulic gradient and in other directions, the passed tracer volumes were small. Using Peclet number as a criterion, it was found that the mass distribution at the fracture intersection was controlled in the stage of transition between the complete mixing model and the streamline routing model.

• Geosciences Journal
• /
• v.22 no.6
• /
• pp.1041-1052
• /
• 2018

A single-well push-drift-pull tracer test using two different tracers ($SF_6$ and salt) was performed at the Environmental Impact Test (EIT) site to determine suitable locations for monitoring wells and arrange them prior to artificial $CO_2$ injection and leak tests. Local-scale estimates of hydraulic properties (linear groundwater velocity and effective porosity) were obtained at the study site by the tracer test with two tracers. The mass recovery percentage of the volatile tracer ($SF_6$) was lower than that of the non-volatile tracer (salt) and increased drift time may make degassing of $SF_6$ intensified. The $CO_2$ leakage monitoring results for both unsaturated and saturated zones suggest that the $CO_2$ monitoring points should be located near points at which a high concentration gradient is expected. Based on the estimated hydraulic properties and tracer mass recovery rates, an optimal $CO_2$ monitoring network including boreholes for monitoring the unsaturated zone was constructed at the study site.

• The Journal of Engineering Geology
• /
• v.24 no.4
• /
• pp.631-641
• /
• 2014

Characteristics of water-level changes in the Temporary Gulpocheon Discharge Channel were identified by observing and analyzing changes in the subterranean water level induced by hydrological stresses the underground aquifer. The subterranean water level refers to the level at which the pressure of subterranean water passing through the corresponding position has an equipotential value that is in equilibrium with the atmospheric pressure at that location. This water level is not fixed but changes in response to hydrological stress. It can be identified by repeatedly measuring the distance from the observation point to the surface of the subterranean water. The subterranean water-level change equation and the variance range of the hydrological curve of subterranean water over 24 hours at the Gimpo-Gimpo National Groundwater Monitoring Network (NGMN) were used as assessment factors. The variance characteristics of the subterranean water at the 18 monitoring system locations were classified into three impact, observational wish, and non-impact. The impact type accounted for 50% of the subterranean water of and accurately reflected the water-level changes due to hydrological stress, showing that distance is the major controlling factor. The observational wish type accounted for 27.8% of the subterranean water, and one of the two assessment factors did not meet the assessment factors. The nonimpact type accounted for 22.2% of the subterranean water. This type satisfied the two assessment factors and represents subterranean water-level changes response to precipitation.

• Journal of Soil and Groundwater Environment
• /
• v.23 no.1
• /
• pp.1-13
• /
• 2018

A series of experiments using transparent micromodels with an artificial pore network etched on glass plates was performed to investigate the effects of flow rate on the migration and distribution of resident wetting porewater (deionized water) and injecting non-wetting fluid (n-hexane). Multicolored images transformed from real RGB images were used to distinguish n-hexane from porewater and pore structure. Hexane flooding followed by immiscible displacement with porewater, migration through capillary fingering, preferential flow and bypassing were observed during injection experiments. The areal displacement efficiency increases as the injection of n-hexane continues until the equilibrium reaches. Experimental results showed that the areal displacement efficiency at equilibrium increases as the flow rate increases. Close observation reveals that preferential flowpaths through larger pore bodies and throats and clusters of entrapped porewater were frequently created at lower flow rate. At higher flow rate, randomly oriented forward and lateral flowpaths of n-hexane displaces more porewater at an efficiency close to stable displacement. It may resulted from that the pore pressure of n-hexane, at higher flow rate, increases fast enough to overcome capillary pressure acting on smaller pore throats as well larger ones. Experimental results in this study may provide fundamental information on migration and distribution of immiscible fluids in subsurface porous media.

• The Journal of Engineering Geology
• /
• v.21 no.1
• /
• pp.79-85
• /
• 2011

We investigated the flow properties of groundwater in areas of carbonate rocks at Yeongwol and Jeongseon, Gangwon Province, based on measurements of hydraulic conductivity. Existing hydraulic conductivity data were compiled from 46 wells in the study area. These wells were sunk close to Golji stream and the Joyang and Dong rivers, which flow through the study area. The hydraulic conductivities range from 0.004 to 1.1 m/day, and show a gradually decreasing trend with decreasing well depth (y=-0.003x - 0.927, $r^2$=0.129). The study area was classified into zone A (< 0.1 m/day), zone B (0.1-1.0 m/day), and zone C (> 1 m/day) according to hydraulic conductivity. Zones A, B, and C make up 87% (n = 40), 11 % (n = 5), and 2% (n = 2) of the surface of the study area, respectively. Among the three zones, zone A contains few fractures whereas zone C contains many fractures. These results indicate that groundwater flow in carbonate regions is strongly influenced by the fracture network.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.14 no.6
• /
• pp.1235-1240
• /
• 2019

It is well known that the ground water level changes rapidly before and after the earthquake, and the variation of ground water level prediction is used to predict the earthquake. In this paper, we predict the ground water level in Miryang City using ANFIS algorithm for earthquake prediction. For this purpose, this paper used precipitation and temperature acquired from National Weather Service and data of underground water level from Rural Groundwater Observation Network of Korea Rural Community Corporation which is installed in Miryang city, Gyeongsangnam-do. We measure the prediction accuracy using RMSE and MAPE calculation methods. As a result of the prediction, the periodic pattern was predicted by natural factors, but the change value of ground water level was changed by other variables such as artificial factors that was not detected. To solve this problem, it is necessary to digitize the ground water level by numerically quantifying artificial variables, and to measure the precipitation and pressure according to the exact location of the observation ball measuring the ground water level.

• Journal of Korea Water Resources Association
• /
• v.55 no.11
• /
• pp.903-911
• /
• 2022

Groundwater, one of the resources for supplying water, fluctuates in water level due to various natural factors. Recently, research has been conducted to predict fluctuations in groundwater levels using Artificial Neural Network (ANN). Previously, among operators in ANN, Gradient Descent (GD)-based Optimizers were used as Optimizer that affect learning. GD-based Optimizers have disadvantages of initial correlation dependence and absence of solution comparison and storage structure. This study developed Gradient Descent combined with Harmony Search (GDHS), a new Optimizer that combined GD and Harmony Search (HS) to improve the shortcomings of GD-based Optimizers. To evaluate the performance of GDHS, groundwater level at Icheon Yullhyeon observation station were learned and predicted using Multi Layer Perceptron (MLP). Mean Squared Error (MSE) and Mean Absolute Error (MAE) were used to compare the performance of MLP using GD and GDHS. Comparing the learning results, GDHS had lower maximum, minimum, average and Standard Deviation (SD) of MSE than GD. Comparing the prediction results, GDHS was evaluated to have a lower error in all of the evaluation index than GD.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2007.05a
• /
• pp.1-7
• /
• 2007

New Zealand is a hydrologically diverse and active country. This paper presents an overview of the major hydrological issues and problems facing New Zealand and provides examples of some the research being undertaken to solve the problems. Fundamental to any environmental decision making is the provision of good quality hydrometric data. Reduced funding for the national hydrometric network has meant a reduction in the number of monitoring sites, the decision on how to redesign the network was made using information on geographic coverage and importance of each site. New Zealand faces a major problem in understanding the impacts of rapid land use change on water quantity and quality. On top of the land use change is overlain the issue of agricultural intensification. The transfer of knowledge about impacts of change at the small watershed scale to much larger, more complex watersheds is one that is attracting considerable research attention. There is a large amount of research currently being undertaken to understand the processes of water and nutrient movement through the vadose zone into groundwater and therefore understanding the time taken for leached nutrients to reach receiving water bodies. The largest water management issue of the past 5 years has been based around fair and equitable water allocation when there is increasing demand for irrigation water. Apart from policy research into market trading for water there has been research into water storage and transfer options and improving irrigation efficiency. The final water management issue discussed concerns the impacts of hydrological extremes (floods and droughts). This is of particular concern with predictions of climate change for New Zealand suggesting increased hydrological extremes. Research work has concentrated on producing predictive models. These have been both detailed inundation models using high quality LIDAR data and also flood models for the whole country based on a newly interpolated grid network of rainfall.