• The Journal of Engineering Research
• /
• v.5 no.1
• /
• pp.89-107
• /
• 2004

This study has tried to develop the modified DRASTIC Model by supplying the parameters, such as structural lineament density and landuse, into conventional DRASTIC model, and to predict the potential of groundwater contamination using GIS in Whanam 2 Area, Gyeonggi Province, Korea. Since the aquifers in Korea is generally through the joints of rock-mass in hydrogeological environment, lineament density affects to the behavior of groundwater and contaminated plumes directly, and land-use reflect the effect of point or non-point source of contamination indirectly. For the statistical analysis, lattice layers of each parameter were generated, and then level of confidence was assessed by analyzing each correlation coefficient. Composite contamination map was achieved as a final result by comparing modified DRASTIC potential and the amount of generation load of several contaminant sources logically. The result could suggest the predictability of the area of contamination potential on the respects of hydrogeological aspect and water quality.

• Journal of Nuclear Fuel Cycle and Waste Technology(JNFCWT)
• /
• v.18 no.2
• /
• pp.133-141
• /
• 2020

A geological repository comprises a natural barrier and an engineered barrier system. Its design components consist of canisters, buffers, backfill, and near-field rock. Among the engineered barrier system components, bentonite buffers minimize the groundwater flow from near-field rock and prevent the release of nuclide. Investigation of the hydraulic conductivity of the buffer to groundwater flow is an important factor in the performance evaluation of the stability and integrity of the engineered barrier of the repository. In this study, saturated hydraulic conductivity tests were performed using Gyeongju bentonite at various dry densities and temperatures, and a hydraulic conductivity prediction model was developed through multiple regression analysis using the 120 result sets of hydraulic conductivity. The test results showed that the hydraulic conductivity tends to decrease as the dry density increases. In addition, the hydraulic conductivity increased with increasing temperature. The multiple regression analysis results showed that the coefficient of determination (R²) of the hydraulic conductivity prediction equation was as high as 0.93. The hydraulic conductivity prediction equation presented in this study could be used for the design of engineered barrier systems.

• Proceedings of the Korean Geotechical Society Conference
• /
• 2010.03a
• /
• pp.1172-1177
• /
• 2010

For the deep excavation in urban area, the braced-cut method is mainly adopted. In this case, inadequate consideration of ground water level may result in wrong prediction of structural behavior. In this study, the effects of hydraulic interaction between wall and grout were investigated using the finite element method. The maximum stress in case of confined ground water condition is obtained at the final excavation stage in the range of 70~80% of excavation depth. The stress of impermeable case is about 50% larger than that of permeable case. When the relative permeabililty of wall-grout become smaller, the stress is getting bigger. And the stress tends to converge in case of 1/100 or less of the relative permeability.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2016.05a
• /
• pp.562-562
• /
• 2016

신경망 모형에서 학습이란 주어진 입출력시스템에 대하여 원하는 동작을 수행할 수 있도록 연결 강도를 최적의 상태로 적응(adaptation)시키는 과정을 의미한다. 따라서 강수와 지하수위의 관계를 연계시킨 인공신경망기법은 선택적으로 예측 지하수위에 영향을 미치는 변수들을 학습에 의하여 택함으로써 예측모형을 구성할 수 있다. 즉, 예측 지하수위와의 상관관계에 의하여 입력되는 변수와의 연결강도를 조정하여 매개변수 조정 및 모형의 최적화를 자동화할 수 있다. 본 연구에서는 지하수위에 영향을 주는 요소는 지하수위와 강우량이라고 가정하고, 지하수위의 입출력과정을 시계열 분석에 의하여 모형화하였으며 예측지하수위는 강우 및 지하수위의 선행조건과 매우 밀접한 관계를 갖는다. 따라서 선행강우 및 지하수위의 상태에 따라 이를 입력하여 미래의 지하수위를 예측하게 된다. 이 모형을 제주지역의 관측소에 적용한 결과 관측소별로 타당한 예측결과를 도출하였다.

• Advances in environmental research
• /
• v.9 no.2
• /
• pp.123-133
• /
• 2020

A time series dataset was conducted to ascertain the effect of water table on the variability in and emission of CH₄ and CO₂ concentrations at a closed landfill site. An in-situ data of methane/carbon dioxide concentrations and environmental parameters were collected by means of an in-borehole gas monitor, the Gasclam (Ion Science, UK). Linear regression analysis was used to determine the strength of the correlation between ground-gas concentration and water table. The result shows CH₄ and CO₂ concentrations to be variable with strong negative correlations of approximately 0.5 each with water table over the entire monitoring period. The R² was slightly improved by considering their concentration over single periods of increasing and decreasing water table, single periods of increasing water table, and single periods of decreasing water table; their correlations increased significantly at 95% confidence level. The result revealed that fluctuations in groundwater level is the key driving force on the emission of and variability in groundgas concentration and neither barometric pressure nor temperature. This finding further validates the earlier finding that atmospheric pressure - the acclaimed major control on the variability/migration of CH₄ and CO₂ concentrations on contaminated sites, is not always so.

• Journal of the Korean Geophysical Society
• /
• v.4 no.4
• /
• pp.267-285
• /
• 2001

This study has tried to develop the modified DRASTIC Model by supplying the parameters,such as structural lineament density and landuse, into conventional DRASTIC medal, and to predict the potential of groundwater contamination using GIS in Whanam 2 Area, Gyeonggi Province, Korea. Since the aquifers in Korea is generally through the joints of rock-mass in hydrogeological environment, lineament denisity affects to the behavior of goundwater and contaminated plumes directly, and land-use reflect the effect of point or non-point source of contamination indirectly. For the statistical analysis, lattice layers of each parameter were generated, and then level of confidence was assessed by analyzing each correlation coefficient. Composite contamination map was achieved as a final result by comparing modified DRASTIC potential and the amount of generation load of several contaminant sources logically. The result could suggest the predictability of the area of contamination potrntial in the respects of hydrogeological aspect and water quality.

• Environmental Analysis Health and Toxicology
• /
• v.16 no.2
• /
• pp.43-50
• /
• 2001

A daily newspaper in Korea addressed an controversial issue recently that the concentration of radon measured from the groundwater in Taejon was found out a relatively high level. The cancer risk arising from ingestion of such radon should be derived from calculation of the dose absorbed by the tissues at risk. The study performed by the National Research Council in United States confirmed that the use of a PBPK model for the ingested radon could provide the useful information regarding the distribution of radon among the organs of the body. This study presents an approach for the internal dose assessment of ingested radon for this case. At first, the study develops a PBPK model for ingested radon. However, the important issue is how to simulate a more realistic situation using the model associated with repeated oral doses rather than a single oral dose. The simulations are performed for repeated oral exposures per 8-hour interval using the PBPK model for a male adult. The concentration and cumulative value of radon concentration are calculated and analyzed for lung tissue and adipose group, respectively. The results could be used for the realistic prediction of the internal dose of radon in the human body for repeated oral exposures.

• Journal of Nuclear Fuel Cycle and Waste Technology(JNFCWT)
• /
• v.14 no.1
• /
• pp.11-19
• /
• 2016

This paper focuses on the estimation of durability and service-life of reinforced concrete structures in Wolsong Low- and intermediate-level wastes Disposal Center (WLDC) in Korea. There are six disposal silos located in the saturated environment. The silo concrete is degraded due to reactions with groundwater and chemical attacks, and finally it will lose its properties as a transport barrier. The infiltration of sulfate and magnesium, leaching of potassium hydroxide, and chlorine induced corrosion are the most significant factors for degradation of reinforced concrete structure in underground environment. From the result of evaluation of the degradation time for each factor, the degradation rate of the reinforced concrete due to sulfate and magnesium is $1.308{\times}10^{-3}cm/yr$, and it is estimated to take 48,000 years for full degradation while potassium hydroxide is leached in depth of less than 1.5 cm at 1,000 years after the initiation of degradation. In case of chlorine induced corrosion, it takes 1,648 years to initiate corrosion in the main reinforced bar and 2,288 years to reach the lifetime limit of the structural integrity, and thus it is evaluated as the most significant factor.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.206-206
• /
• 2022

제주도는 투수성이 좋은 대수층이 발달한 화산섬으로 지하수가 가장 중요한 수자원이다. 인위적 요인과 기후변화로 인해 제주도의 지하수위가 저하하는 추세를 보이고 있음에 따라 지하수의 적정 관리를 위해 지하수위의 정확하고 장기적인 예측이 매우 중요하다. 다양한 환경적인 요인이 지하수의 함양 및 수위에 영향을 미치는 것으로 알려져 있지만, 제주도의 특징적인 기상인자가 지하수 시스템에 어떻게 영향을 미치는지를 파악하기 위한 연구는 거의 진행되지 않았다. 지하수위측에 있어서 물리적 모델을 이용한 방안은 다양한 조건에 의해 변화하는 지하수위의 정확하고 빠른 예측에 한계가 있는 것으로 알려져 있다. 이에 본 연구에서는 제주도 애월읍과 남원읍에 위치한 지하수위 관측정의 일 수위자료와 강수량, 온도, 강설량, 풍속, VPD의 다양한 기상 자료를 대상으로 인공신경망 알고리즘인 다층 퍼셉트론(MLP)와 Long Short Term Memory(LSTM)에 기반한 표준지하수지수(SGI) 예측 모델을 개발하였다. MLP와 LSTM의 표준지하수지수(SGI) 예측결과가 상당히 유사한 것으로 나타났으며 MLP과 LSTM 예측모델의 결정계수(R²)는 애월읍의 경우 각각 0.98, 남원읍의 경우 각각 0.96으로 높은 값을 보였다. 본 연구에서 개발한 지하수위 예측모델을 통해 효율적인 운영과 정밀한 지하수위 예측이 가능해질 것이며 기후변화 대응을 위한 지속가능한 지하수자원 관리 방안 마련에 도움을 줄 것이라 판단된다.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.186-186
• /
• 2022

강수 및 침투 등으로 발생하는 지하수위의 변동을 예측하는 것은 지하수 자원의 활용 및 관리에 필수적이다. 지하수위의 변동은 지하수 자원의 활용 및 관리뿐만이 아닌 홍수 발생과 지반의 응력상태 등에 직접적인 영향을 미치기 때문에 정확한 예측이 필요하다. 본 연구는 인공신경망 중 다층퍼셉트론(Multi Layer Perceptron, MLP)을 이용한 지하수위 예측성능 향상을 위해 MLP의 구조 중 Optimizer를 개량하였다. MLP는 입력자료와 출력자료간 최적의 상관관계(가중치 및 편향)를 찾는 Optimizer와 출력되는 값을 결정하는 활성화 함수의 연산을 반복하여 학습한다. 특히 Optimizer는 신경망의 출력값과 관측값의 오차가 최소가 되는 상관관계를 찾는 연산자로써 MLP의 학습 및 예측성능에 직접적인 영향을 미친다. 기존의 Optimizer는 경사하강법(Gradient Descent, GD)을 기반으로 하는 Optimizer를 사용했다. 하지만 기존의 Optimizer는 미분을 이용하여 상관관계를 찾기 때문에 지역탐색 위주로 진행되며 기존에 생성된 상관관계를 저장하는 구조가 없어 지역 최적해로 수렴할 가능성이 있다는 단점이 있다. 본 연구에서는 기존 Optimizer의 단점을 개선하기 위해 지역탐색과 전역탐색을 동시에 고려할 수 있으며 기존의 해를 저장하는 구조가 있는 메타휴리스틱 최적화 알고리즘을 이용하였다. 메타휴리스틱 최적화 알고리즘 중 구조가 간단한 화음탐색법(Harmony Search, HS)과 GD의 결합모형(HS-GD)을 MLP의 Optimizer로 사용하여 기존 Optimizer의 단점을 개선하였다. HS-GD를 이용한 MLP의 성능검토를 위해 이천시 지하수위 예측을 실시하였으며 예측 결과를 기존의 Optimizer를 이용한 MLP 및 HS를 이용한 MLP의 예측결과와 비교하였다.