• Journal of the Korean Society of Hazard Mitigation
• /
• v.10 no.1
• /
• pp.73-80
• /
• 2010

The optimal design of water distribution system have started with the least cost design of single objective function using fixed hydraulic variables, eg. fixed water demand and pipe roughness. However, more adequate design is accomplished with considering uncertainties laid on water distribution system such as uncertain future water demands, resulting in successful estimation of real network's behaviors. So, many researchers have suggested a variety of approaches to consider uncertainties in water distribution system using uncertainties quantification methods and the optimal design of multi-objective function is also studied. This paper suggests the new approach of a multi-objective optimization seeking the minimum cost and maximum robustness of the network based on two uncertain variables, nodal demands and pipe roughness uncertainties. Total design procedure consists of two folds: least cost design and final optimal design under uncertainties. The uncertainties of demands and roughness are considered with Latin Hypercube sampling technique with beta probability density functions and multi-objective genetic algorithms (MOGA) is used for the optimization process. The suggested approach is tested in a case study of real network named the New York Tunnels and the applicability of new approach is checked. As the computation time passes, we can check that initial populations, one solution of solutions of multi-objective genetic algorithm, spread to lower right section on the solution space and yield Pareto Optimum solutions building Pareto Front.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2015.05a
• /
• pp.466-466
• /
• 2015

지역빈도해석은 수문학에서 오랜 역사를 갖고 있으며, 수년에 걸쳐 수문학적 변량의 정량적 추정을 위해 다양한 접근방법들이 제안되어 왔다. 그러나 제안된 방법들의 가설설정 수준이 높기 때문에 실제 적용에 제약이 많고, 적용 시에도 예측에 대한 불확실성이 높은 문제점이 있다. 본 연구에서는 이러한 문제점을 개선하기 위한 방법으로 계층적 베이지안 모델을 이용한 지역빈도해석 모형을 제안하고자 한다. 본 모형은 2개의 계층적 구조로 구성된다. 첫번째 계층은 재현기간별 GEV 분포의 매개변수를 정규화하여 주변분포로 설정하고, Kriging 기법을 이용하여 지형학적, 기상학적 정보들과 극치강수량 효과를 적합시켜 공간적 이질성과 미계측 유역에 대한 효과적인 보간을 가능하게 한다. 두번째 계층은 지점의 특성을 나타내는 매개변수들간의 공분산을 Bayesian 모델에 연계하여 매개변수들의 공간적 변동성을 나타낸다. 2개 계층의 결합확률분포는 MCMC 기법을 이용하여 예측값에 대한 불확실성을 정량적으로 분석하게 된다. 본 모형을 통해 홍수량 추정 시 필요한 시간 단위 극치강수량의 공간적 분포를 효과적으로 추정할 수 있을 것으로 판단된다.

• Journal of Korean Society for Geospatial Information Science
• /
• v.22 no.4
• /
• pp.47-52
• /
• 2014

Species distribution models have been widely applied in order to assess biodiversity, design reserve, manage habitat and predict climate change. However, SDMs has been used restrictively to the public and policy sectors owing to model uncertainty. Recent studies on ensemble and consensus models have been increased to reduce model uncertainty. This paper was carried out single model and multi model for Corylopsis coreana and compares two models. First, model evaluation was used AUC, kappa and TSS. TSS was the most effective method because it was easy to compare several models and convert binary maps. Second, both single and ensemble model show good performance and RF, Maxent and GBM was evaluated higher, GAM and SRE was evaluated lower relatively. Third, ensemble model tended to overestimate over single model. This problem can be solved by the suitable model selection and weighting through collaboration between field experts and modeler. Finally, we should identify causes and magnitude of model uncertainty and improve data quality and model methods in order to apply special decision-making support system and conservation planning, and when we make policy decisions using SDMs, we should recognize uncertainty and risk.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2023.05a
• /
• pp.347-347
• /
• 2023

홍수 예·경보 시에는 시간-단위 또는 그 이하의 시간 척도에서 작용하는 강우에 대한 고도의 영향이 중요하게 되며, 특히 상대적으로 더 드문 관측 밀도가 있는 산악지역에서 강우의 공간분포에 대한 산악 효과의 중요도가 더 높아지게 된다. 일반적으로 1시간 시간스케일에서 강우-고도의 관계를 살펴보기 위해서는 대략 5km 내외의 관측 밀도를 가져야 하는 것으로 알려져 있으나 이러한 지역은 매우 드물다. 최근 기상 예측 수치모델로부터 모의된 강우량의 품질이 눈에 띄게 향상됨에 따라 국내에도 다양한 연구가 수행된 바 있다. 본 연구에서는 WRF를 이용하여 남강댐 지역의 과거 호우 사상을 재현한 후, 이로부터 생산된 공간적인 강우장을 이용하여 시간-단위의 시간 척도에서 강우량과 고도 사이의 관계를 고려할 수 있는 WREPN(WRF Rainfall-Elevation Parameterized Nowcasting) 모형을 제안한다. 홍수량 분석을 위해 WREPN 모형을 이용하였으며, 비교군으로 실무적으로 많이 사용되는 IDW, Kriging 기반 격자강우가 사용되었다. 격자기반 분포형 강우자료로부터 홍수량을 분석하기 위해 Mod-Clark 모형이 적용되었으며, 입력된 강우자료별매개변수의 불확실성을 분석하기 위해 베이지안 기법이 적용되었다. 매개변수의 불확실성 분석으로부터 강우-고도 관계가 고려된 WREPN 모형의 강우자료가 상대적으로 불확실성이 낮다는 것을 확인할 수 있었다.

• Journal of Korea Spatial Information System Society
• /
• v.8 no.3
• /
• pp.117-130
• /
• 2006

With the rapid development of technologies related to Ubiquitous Sensor Network (USN), sensors are being utilized in various application areas. In general, the data sensed by each sensor node on ubiquitous sensor networks are stored into the central server for efficient search. Because update is delayed to reduce the cost of update in this environment, uncertain data can be stored in the central server. In addition, Uncertain data make query processing produce wrong results in the central server. Thus, this paper examines how to process uncertain data in ubiquitous sensor networks and suggests a new index for efficient processing of uncertain data. The index reduces the cost of update by delaying update in uncertainty areas. Uncertainty areas are areas where uncertain data are likely to exist. In addition, it solves the problem of low accuracy in search resulting from update delay by delaying update only for specific update areas. Lastly, we analyze the performance of the index and prove the superiority of its performance by comparing its performance evaluation.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.9 no.4
• /
• pp.215-223
• /
• 2006

The Total Maximum Daily Load(TMDL) Act just implemented as a new tool of watershed based water quality management, in Korea. Thus, there are a number of pending questions to resolve for successful settlement of the TMDL. The allocation of pollution source is a exceedingly sensitive issue on local development planning. The simple area-based allocation (SAA) is conventional method to allocate the administrational pollution information to watershed based information. The SAA has a limitation that it can't consider the characteristics of spatial distribution of pollution source and it has caused more uncertainty of TMDL. This study was performed to reduce the uncertainty of watershed-based pollution information using the spatial distribution-based allocation(SDA). In the specific area where pollution source is concentrated such as urbanized region, it has been certified that SDA could reduce a tolerance of pollution information dramatically. As a result of study, SDA is expected a effective tool for TMDL and to solve the conflict between development and protection.

• Proceedings of the Korean Statistical Society Conference
• /
• 2004.11a
• /
• pp.117-121
• /
• 2004

분석용 정밀 워게임 시뮬레이션 모형에서는 '모형운영 결과와 실제(또는 실험) 결과를 비교' 하는 통상적인 타당성 척도의 적용이 불가능함에 따라 워게임모형 운영환경에 적합한 새로운 개념의 타당성 척도로서 VEA(Validity for Exploratory Analysis), VSA(Validity subject to Assumption) 등의 개념을 도입하고 이를 탐색적으로 점검하는 방안을 제시한다. 분석용 워게임모형 활용에 있어 또 하나의 걸림돌은 1)시나리오 및 상황의 가변성, 2)무기체계 및 장비 성능에 대한 불확실성, 3)묘사범위 제한 및 논리의 부정확성으로 인한 오류 등으로 엄청난 불확실성(uncertainty)을 기본적으로 내포함에 따라 구체적 의사결정을 위한 종합적 결론 도출이 어렵다는 점이다. 본 연구에서는 이를 메타모델(Meta model) 즉 워게임모형 입출력 자료의 관계를 묘사한 통계적 모형을 구축하고 이를 기반으로 다양한 불확실성 하에서 관심변수간의 관계를 종합적으로 도출하고자 하는 '관련공간모의(Relevant Simulation)' 방안을 제시한다. 이와 같은 방안들은 SVAP(Statistical Validation and Aggregation Procedure)라는 하나의 종합된 절차로서 제시된다.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.8 no.3
• /
• pp.249-257
• /
• 2006

The spatial distribution of design parameters greatly affects tunnel behavior during and after construction, as well as in the long-term temporal responses. However, the tunnel design parameters commonly used in numerical modeling tend to be representative or average values of global-scale properties. Furthermore, the uncertainty and spatial variation of the design parameters increase as the tunnel scale increases. Consequently, the probability of failure also increases. In order to achieve structural stability in large-section tunnels, the design framework must take into consideration the quantitative effect of design parameter variations on tunnel behavior. Therefore, this paper suggests a statistical approach to numerical modeling to explore the effect of spatially distributed design parameters in a circular tunnel. Also, the effect of spatial variation in the lining strength is studied in this paper. The numerical results suggest that the deformation around the tunnel increases with an increase in the variation of the design parameters.

• Journal of the Korean Geotechnical Society
• /
• v.36 no.4
• /
• pp.31-39
• /
• 2020

Shear wave velocity (or shear modulus) is very important in the evaluation of seismic performance of a fill dam under an earthquake. A shear wave velocity profile can be determined by surface wave method such as HWAW and SASW methods but this profile has uncertainty caused by spatial variation of material property in a fill dam. This uncertainty in shear wave velocity profile could be considered using a coefficient of variation of material property in the reliability based analysis. In this paper, the possible 600 shear wave velocity profiles in the core and rockfill zone of fill dam were generated by the random shear wave velocity profile generation method, proposed by Hwang and Park, based on the field shear wave velocity profiles determined by the HWAW and SASW methods. And, through the statistical analysis of generated shear wave velocity profiles in the fill dam, the coefficient of variation (COV) of shear wave velocity with depth were evaluated for the core and rock filled zone of fill dam in Korea.

• Journal of Korea Water Resources Association
• /
• v.47 no.5
• /
• pp.469-482
• /
• 2014

This study developed a Bayesian spatial regional frequency analysis, which aimed to analyze spatial patterns of design rainfall by incorporating geographical information (e.g. latitude, longitude and altitude) and climate characteristics (e.g. annual maximum series) within a Bayesian framework. There are disadvantages to considering geographical characteristics and to increasing uncertainties associated with areal rainfall estimation on the existing regional frequency analysis. In this sense, this study estimated the parameters of Gumbel distribution which is a function of geographical and climate characteristics, and the estimated parameters were spatially interpolated to derive design rainfall over the entire Han-river watershed. The proposed Bayesian spatial regional frequency analysis model showed similar results compared to L-moment based regional frequency analysis, and even better performance in terms of quantifying uncertainty of design rainfall and considering geographical information as a predictor.