• Proceedings of the Korea Water Resources Association Conference
• /
• 2011.05a
• /
• pp.353-356
• /
• 2011

본 연구에서는 G/R 비의 실시간 결정을 목적으로 Dual Kalman Filter를 이용하였다. Dual Kalman Filter 는 이중추정(dual estimation)을 기반으로 하는 자료동화기법으로 기존 Kalman Filter와 상이한 상태-공간 모형으로 구성된다. 이에 Dual Kalman Filter와 기존 Kalman Filter의 적용성능을 비교 검토하였으며, 다양한 비교를 위하여 강우의 임계치와 누적시간의 고려여부에 따른 결과를 추가적으로 검토하였다. 두 기법의 적용성능 비교결과 Dual Kalman Filter가 우수한 것으로 나타났다. 이는 Dual Kalman Filter 기법이 G/R 비의 큰 변동성과 이상치를 효과적으로 필터링하고, 시계열 모형의 매개변수를 실시간으로 갱신하여 정확한 예측치를 추정하였기 때문인 것으로 판단된다.

• Journal of the Korean Vacuum Society
• /
• v.5 no.4
• /
• pp.292-300
• /
• 1996

The scattering of incident ions and target atoms in the amorphous solid matters are calculated by Monte Carlo simulation method. The experimentally derived universal scattering cross-section of Kalbitzer and Oetzmann is used to describe nuclear scattering. For electronic energy loss, the Lindhard-Scharff and Bethe formula are used. Comparing the ion scattering formulas and ranges with the known results of experiment and other programs, we find our results are good agreement with others.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2010.05a
• /
• pp.1453-1456
• /
• 2010

본 연구에서는 권역별 일강수량에 대한 기후변화 영향을 평가하고자 한다. 이를 위해서 지역기후 모형으로부터 유도된 A2시나리오가 기본자료로서 활용되며 분석에 앞서 기후변화 시나리오의 편의를 보정하였다. 보정된 지역기후 시나리오는 비정상성 Markov Chain 모형을 통해 강수지점별로 상세수문시나리오로 가공되어 분석에 이용되었다. 강수량에 대한 연주기 분포에 대한 특성을 권역별로 시기별로 평가하였다. 연주기의 분포는 관측치와 유사한 거동을 보이고 있으며 강우량이 2075년대에 증가하는 것으로 나타났다. 기후변화에 따른 일강수량에 발생 특성을 평가하기 위해서 Dryspell과 Wetspell을 지속시간분포별로 추정하여 분석하였다. 전체적으로 Dryspell이 지속시간별로 증가하는 것으로 나타나고 있으며 무강수일수의 증가도 전망되고 있어 같은 연강수량이라도 변동성이 크게 나타날 수 있는 개연성이 크다 하겠다. 일강수량, 월강수량, 연강수량에 대한 다양한 분석이 수행되었으며 기후변화에 따른 강수량의 변동양상을 권역별로 평가하였다.

• Annual Conference on Human and Language Technology
• /
• 2022.10a
• /
• pp.608-612
• /
• 2022

딥러닝의 발전 이후, 다양한 분야에서는 딥러닝을 이용해 이전에 어려웠던 작업들을 해결하여 사용자에게 편의성을 제공하고 있다. 하지만 아직 딥러닝을 통해 이상적인 서비스를 제공하는 데는 어려움이 있다. 특히, 음성 인식 작업에서 음성 양식에서 이용 방안에 대하여 다양성을 제공해주는 음성을 텍스트로 전환하는 Speech-To-Text(STT)은 문장 결과가 이상치에 달하지 못해 오류가 나타나게 된다. 본 논문에서는 STT 결과 보정을 문법 교정으로 치환하여 종단에서 올바른 토큰들을 조합하여 성능 향상을 하기 위해 각 토큰 별 품질 평가를 진행하는 모델을 한국어에서 적용하고 성능의 향상을 확인한다.

• Geophysics and Geophysical Exploration
• /
• v.14 no.4
• /
• pp.305-315
• /
• 2011

A total of 864 measurements for P- and S- wave velocity of acrylic and stainless steel core samples have been performed with respect to their lengths and axial load applied. S-wave velocity measurement was much harder than P-wave velocity, so that it showed higher deviation in measured S-wave velocity with respect to repeated measurement, length of the cores, and the axial load applied. Velocity measurements for acrylic cores showed more stable and less than half of the variation between the measurements than the stainless steel cores. This seems to be come from better coupling between the transducers and acrylic cores than stainless cores, and from larger value of the first arrival time in a similar system noise environments. From the analysis of the 864 measurements, it is recommended that the length of the core be 60 ~ 90 mm, axial load between 20 kg (27.7 $N/cm^2$) and 30 kg (41.6 $N/cm^2$) for measurement of wave velocity of the acrylic and stainless steel cores. Especially for measuring S-wave velocity of stainless steel core, core length should be less than 50 mm, otherwise it will be affected by mode conversion or others. These results can be used in measurement and correction for system delay in wave velocity measurement for rock cores.

• Geophysics and Geophysical Exploration
• /
• v.15 no.4
• /
• pp.209-218
• /
• 2012

Repeatability and reproducibility in solid weight and effective porosity measurements have been discussed using 8 core samples with different diameters, lengths, rock types, and effective porosities. Further, the effect of temperature on the effective porosity measurement has been discussed as well. Effective porosity of each sample has been measured 7 times with vacuum saturation method with vacuum pressure of 1 torr and vacuum time of 80 minutes. Firstly, effective porosity of each sample is measured one by one, so that it can provide a reference value. Then for reproducibility check, effective porosity measurements with vacuum saturation of 2, 4, and 8 samples simultaneously have been performed. And finally, repeated measurements for 3 times for each sample are made for repeatability check. Average deviation from the reference set in solid weight showed 0.00 $g/cm^3$, which means perfect repeatability and reproducibility. For effective porosity, average deviations are less than 0.07% and 0.05% in repeatability and reproducibility test sets, respectively, which are in good agreement too. Most of porosities measured in reproducibility test lies within the deviation range in repeatability test sets. Thus, simultaneous vacuum saturation of several samples has little impact on the effective porosity measurement when high vacuum pressure of 1 torr is used. Air temperature can cause errors on submerged weight read and even effective porosity, because it is closely related to the temperature, density, and buoyancy of water. Consequently, for accurate measurement of effective porosity in a laboratory, efforts for maintaining air or water temperature constant during the experiment, or a temperature correction from other information are needed.

• Survey Research
• /
• v.9 no.2
• /
• pp.1-27
• /
• 2008

Problems of incomplete data are pervasive in statistical analysis. In particular, incomplete data have been an important challenge in repeated measures studies. The objective of this study is to give a brief introduction to missing data mechanisms and conventional/recent missing data methods and to assess the performance of various missing data methods under ignorable and non-ignorable missingness mechanisms. Given the inadequate attention to longitudinal studies with missing data, this study applied recent advances in missing data methods to repeated measures models and investigated the performance of various missing data methods, such as FIML (Full Information Maximum Likelihood Estimation) and MICE(Multivariate Imputation by Chained Equations), under MCAR, MAR, and MNAR mechanisms. Overall, the results showed that listwise deletion and mean imputation performed poorly compared to other recommended missing data procedures. The better performance of EM, FIML, and MICE was more noticeable under MAR compared to MCAR. With the non-ignorable missing data, this study showed that missing data methods did not perform well. In particular, this problem was noticeable in slope-related estimates. Therefore, this study suggests that if missing data are suspected to be non-ignorable, developmental research may underestimate true rates of change over the life course. This study also suggests that bias from non-ignorable missing data can be substantially reduced by considering rich information from variables related to missingness.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.22 no.3
• /
• pp.352-361
• /
• 2021

An underwater positioning method that can be applied to structures for underwater construction is being developed at the Korea Institute of Ocean Science and Technology. The method uses an extended Kalman filter (EKF) based on an inertial navigation system for precise and continuous position estimation. The observation matrix was configured to be variable in order to apply asynchronous measured sensor data in the correction step of the EKF. A Doppler velocity logger (DVL) can acquire signals only when attached to the bottom of an underwater structure, and it is difficult to install and recover. Therefore, a complex sensor device for underwater structure attachment was developed without a DVL in consideration of an underwater construction environment, installation location, system operation convenience, etc.. Its performance was verified through a water tank test. The results are the measured underwater position using an ultra-short baseline, the estimated position using only a position vector, and the estimated position using position/velocity vectors. The results were compared and evaluated using the circular error probability (CEP). As a result, the CEP of the USBL alone was 0.02 m, the CEP of the position estimation with only the position vector corrected was 3.76 m, and the CEP of the position estimation with the position and velocity vectors corrected was 0.06 m. Through this research, it was confirmed that stable underwater positioning can be carried out using asynchronous sensors without a DVL.

• Journal of Korea Water Resources Association
• /
• v.51 no.2
• /
• pp.151-163
• /
• 2018

The objective of this study was to estimate LOADEST (LOAD Estimator) coefficients for simulating pollutant loads in ungauged watersheds. Regression models of LOADEST were used to simulate pollutant loads, and the multiple linear regression (MLR) was used for coefficients estimation on watershed characteristics. The fifth and third model of LOADEST were selected to simulate T-N (Total-Nitrogen) and T-P (Total-Phosphorous) loads, respectively. The results and statistics indicated that regression models based on LOADEST simulated pollutant loads reasonably and model coefficients were reliable. However, the results also indicated that LOADEST underestimated pollutant loads and had a bias. For this reason, simulated loads were corrected the bias by a quantile mapping method in this study. Corrected loads indicated that the bias correction was effective. Using multiple regression analysis, a coefficient estimation methods according to the watershed characteristic were developed. Coefficients which calculated by MLR were used in models. The simulated result and statistics indicated that MLR estimated the model coefficients reasonably. Regression models developed in this study would help simulate pollutant loads for ungauged watersheds and be a screen model for policy decision.

• Journal of Korea Water Resources Association
• /
• v.47 no.2
• /
• pp.119-133
• /
• 2014

The study adopted extended Kalman filter technique in an effort to predict Z-R relationship parameter as a stable value in real-time. Toward this end, a parameter estimation model was established based on extended Kalman filter in consideration of non-linearity of Z-R relationship. A state-space model was established based on a study that was conducted by Adamowski and Muir (1989). Two parameters of Z-R relationship were set as state variables of the state-space model. As a result, a stable model where a divergence of Kalman gain and state variables are not generated was established. It is noteworthy that overestimated or underestimated parameters based on a conventional method were filtered and removed. As application of inappropriate parameters might cause physically unrealistic rain rate estimation, it can be more effective in terms of quantitative precipitation estimation. As a result of estimation on radar rainfall based on parameters predicted with the extended Kalman filter, the mean field bias correction factor turned out to be around 1.0 indicating that there was a minor difference from the gauge rain rate without the mean field bias correction. In addition, it turned out that it was possible to conduct more accurate estimation on radar rainfall compared to the conventional method.