• Proceedings of the Korea Water Resources Association Conference
• /
• 2015.05a
• /
• pp.422-422
• /
• 2015

수문학적 모델링은 수자원계획에 있어 가장 핵심적인 도구 중에 하나이다. 강우-유출모형의 매개변수 추정시 장기간의 자료를 활용하는데 초점이 맞추어져 있으며, 일반적으로 5년 이상의 자료를 활용하여 매개변수를 추정하는 경년변동(inter-annual variability) 매개변수 추정 방법이 추천되고 있다. 수문학적 변동성 측면에서 볼 때 강우, 온도, 유역의 조건 등의 연내변동성(intra-annual variability)이 경년보다 크게 나타나고 있으나, 이러한 특성을 고려한 수문모형의 매개변수 추정은 이루어지고 있지 않다. 이러한 점에서 연내변동성으로 기인하는 비정상성을 고려한 매개변수 추정 방법의 도입이 필요할 것으로 판단되며, 본 연구에서는 계측유역을 대상으로 다양한 시간규모에서 매개변수 추정을 수행하고 최적의 시간규모를 도출하고자 한다. 이를 위해서 DDS(dynamically dimensioned search) 알고리즘을 도입하여 최적화를 수행하였으며, 다양한 시간 규모에서 모형의 적합특성을 평가하였다. 교차검증을 통하여 매개변수의 통계적 유의성을 확보하였으며, 전통적인 매개변수 추정 절차와 비교 검토를 수행하였다.

• Journal of Wetlands Research
• /
• v.21 no.1
• /
• pp.1-8
• /
• 2019

In this study, the multi-objective optimization method is attemped to optimize the hydrological model to estimate the runoff through two hydrological processes. HL-RDHM, a distributed hydrological model that can simultaneously estimate the amount of snowfall and runoff, was used as the distributed hydrological model. The Durango River basin in Colorado, USA, was selected as the watershed. MOSCEM was used as a multi-objective optimization method and parameter calibration and hydrologic model optimization were tried by selecting 5 parameters related to snow melting and 13 parameters related to runoff. Data from 2004 to 2005 were used to optimize the model and verified using data from 2001 to 2004. By optimizing both the amount of snow and the amount of runoff, the RMSE error can be reduced from 7% to 40% of the simulation value based on the initial solution at three SNOTEL points based on the RMSE. The USGS observation point of the outflow is improved about 40%.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2018.05a
• /
• pp.9-9
• /
• 2018

추계학적 강우생성모형 중 포아송 클러스터(Poisson Cluster) 모형은 단일지점에 대하여 시간강우량의 관측연한 문제점을 해결하기 위한 강우모형으로 강우 단계별 계층적 구조를 이해하는데 유용한 모형이다. 특히 강우 특성을 계절, 지역 등과 같이 비교하는 기준에 따라 5~6개의 비교적 적은 매개변수들로 모의 강우시계열을 생성할 수 있다는 점에서 장기간 강우분석에 필요한 관측연한 문제를 보완할 수 있다. 그러나 매개변수 최적해가 수렴되지 않는 사례가 많고, 매개변수들이 강우의 물리적 특성을 반영하는 것에 비해 내포된 불확실성에 관한 연구는 미흡하다. 본 연구에서는 포아송 클러스터 강우생성모형 중 Neyman-Scott Rectangular Pulse(NSRP) 모형을 Bayesian 모형과 연계한 Bayesian NSRP 모형을 개발하여 매개변수간 물리적 상관성을 고려한 최적화 기법을 개발하였다. Bayesian 모형은 물리적 범위가 다른 매개변수간의 결합확률분포를 산정하여 사후분포(posterior)를 추정하므로 매개변수 최적화와 불확실성 정량화 문제를 동시에 해결할 수 있다. 최종적으로 Bayesian NSRP 모형에 기후변화 시나리오의 통계적 특성을 고려한 시간단위 강우시계열 생성 모의 기법의 활용 가능성을 평가하고자 한다.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2004.05b
• /
• pp.1162-1166
• /
• 2004

본 연구에서는 미국의 Environmental Modeling Research Laboratory(EMRL)에서 개발한 지리정보시스템과 수문유출모형이 접목된 WMS(Watershed Modeling System) Ver.6.1 모형을 HEC-1모형과 연계하여 한강수계 안성천의 제1지류인 한천의 유출 해석을 실시하였다. 이를 위해 유역내의 지형특성인자 추출 및 하천망 구성(WMS), 재현기간별 지속시간별 확률강우량 산정(FARD2002), Huff분포법을 이용한 시간분포, ARF(Area Reduction Factor)적용, HEC-1내의 SCS단위도법, Snyder 단위도법, Clark의 유역추적법에 포함된 각각의 매개변수의 최적화를 시도하여 분석하고, 설계 홍수량 산정시 이용될 수 있는 지침 마련을 목적으로 하였다.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.20 no.2
• /
• pp.412-424
• /
• 2019

In this study, the changes of the streamflow characteristics of the watershed were analysed depending on the infiltration methods of CAT. The study area, Boryeong-dam watershed located in Chungcheongnam-do area, has been suffered from severe drought in recent years and stabilized regarding on the storage rate through efforts such as constructing a channel connecting the upstream of Boryeong-dam from the downstream of the Geum river. In this study, the effects of soil infiltration parameters on the watershed streamflow characteristics were analyzed by the infiltration methods of CAT such as Rainfall Excess, Green&Ampt and Horton. And the parameter calibrations were conducted by SCEUA-P, a global optimization technique module of the PEST, the package for parameter optimization and uncertainty analysis, to compare the yearly variations of soil parameters for infiltration methods of CAT. In addition, the streamflow characteristics were analyzed for three infiltration methods by applying three different scenarios, such as applying calibrated parameters for every years to simulate the model for each years, applying calibrated parameters for the entire period to simulate the model for entire period, and applying the average value of yearly calibrated parameters to simulate the model for entire period.

• Journal of Wetlands Research
• /
• v.19 no.4
• /
• pp.423-432
• /
• 2017

This study used SSARR model to predict the flood hydrograph for the Miho stream in the Geum river basin. First, we performed the sensitivity analysis on the parameters of SSARR model to know the characteristics of the parameters and set the range. For the parameter calibration, optimization methods such as genetic algorithm, pattern search and SCE-UA were used. WSSR and SSR were applied as objective functions, and the results of optimization method and objective function were compared and analyzed. As a result of this study, flood prediction was most accurate when using pattern search as an optimization method and WSSR as an objective function. If the parameters are optimized based on the results of this study, it can be helpful for decision making such as flood prediction and flood warning.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2015.10a
• /
• pp.1285-1288
• /
• 2015

본 논문에서는 영상인식을 위한 딥 러닝에서 사용되는 매개변수 최적화 방법을 제안한다. 학습 성능에 영향을 미치는 매개변수 중 이미지 배치 사이즈 값, 초기 학습율, 최대 학습 반복 횟수에 대해 상호간의 관계를 분석하고 성능을 개선시키기 위해 값을 최적화하는 방법을 연구한다. 제안된 방법을 통한 개선 정도를 분석하기 위해 매개변수의 변화에 따른 학습 소요 시간, 정확도 향상 추이, 메모리 사용량의 변화를 측정한다. 측정된 학습 소요 시간, 정확도 향상 추이, 메모리 사용량의 변화를 분석한 결과 배치 사이즈와 초기 학습 율은 같은 비율로 반비례하게 값을 적용할 때가 이상적 이였으며 서로 다른 환경에서 각각의 학습 소요시간을 측정하는 것으로 배치 사이즈 값과 초기 학습 율에 따른 최적의 최대 학습 반복 횟수를 획득할 수 있었다.

• Journal of Wetlands Research
• /
• v.22 no.1
• /
• pp.15-23
• /
• 2020

The purpose of this study is to investigate how the degree of distribution influences the calibration of snow and runoff in distributed hydrological models using a multi-criteria calibration method. The Hydrology Laboratory-Research Distributed Hydrologic Model (HL-RDHM) developed by NOAA-National Weather Service (NWS) is employed to estimate optimized parameter sets. We have 3 scenarios depended on the model complexity for estimating best parameter sets: Lumped, Semi-Distributed, and Fully-Distributed. For the case study, the Durango River Basin, Colorado is selected as a study basin to consider both snow and water balance components. This study basin is in the mountainous western U.S. area and consists of 108 Hydrologic Rainfall Analysis Project (HRAP) grid cells. 5 and 13 parameters of snow and water balance models are calibrated with the Multi-Objective Shuffled Complex Evolution Metropolis (MOSCEM) algorithm. Model calibration and validation are conducted on 4km HRAP grids with 5 years (2001-2005) meteorological data and observations. Through case study, we show that snow and streamflow simulations are improved with multiple criteria calibrations without considering model complexity. In particular, we confirm that semi- and fully distributed models are better performances than those of lumped model. In case of lumped model, the Root Mean Square Error (RMSE) values improve by 35% on snow average and 42% on runoff from a priori parameter set through multi-criteria calibrations. On the other hand, the RMSE values are improved by 40% and 43% for snow and runoff on semi- and fully-distributed models.

• Journal of the Korean Society of Hazard Mitigation
• /
• v.1 no.3 s.3
• /
• pp.73-82
• /
• 2001

This study examines the applicability of the Nash model and the Diskin model, which are linear and nonlinear runoff models, respectively, by applying optimization techniques to the parameter calibration of the two models. Nonlinear programming which is one of traditional optimization techniques and Genetic Algorithm which has been actively applied recently are used in this study. The Nash and Diskin models which use the calibrated parameter with a flood events are applied to a different flood event in Soyang Dam basin. The results obtained from the parameter calibration show slight discrepancy depending upon the flood events. It has been found in the comparion between the observed hydrograph and the hydrographs obtained from the parameter calibration that the Diskin model can better simulate the observed hydrograph than the Nash model can, especially, for the peak flow. This can be analyzed that the Diskin model which is a nonlinear runoff model is better off in simulating the nonlinear characteristic of the rainfall-runoff process.

• Journal of Korea Water Resources Association
• /
• v.52 no.10
• /
• pp.681-695
• /
• 2019

A Transient Storage Model (TSM) is one of the most widely used model accounting for complex solute transport in natural river to understanding natural river properties with four TSM key parameters. The TSM parameters are estimated via inverse modeling. Parameter estimation of the TSM is carried out by solving optimization problem about finding best fitted simulation curve with measured curve obtained from tracer test. Several studies have reported uncertainty in parameter estimation from non-convexity of the problem. In this study, we assessed best combination of optimization method and objective function for TSM parameter estimation using Cheong-mi Creek tracer test data. In order to find best optimization setting guaranteeing convergence and speed, Evolutionary Algorithm (EA) based global optimization methods, such as CCE of SCE-UA and MCCE of SP-UCI, and error based objective functions were compared, using Shuffled Complex-Self Adaptive Hybrid EvoLution (SC-SAHEL). Overall results showed that multi-EA SC-SAHEL with Percent Mean Squared Error (PMSE) objective function is the best optimization setting which is fastest and stable method in convergence.