• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2004.04a
• /
• pp.288-291
• /
• 2004

기존의 SOFPNN은 데이터 수가 적고 비선형 요소가 많은 시스템에 대한 체계적이고 효율적인 최적 모델 을 구축할 수 있었으며 각 층 노드의 선택 입력을 변화시킴으로써 네트워크 구조 전체의 적응능력을 향상 시켰다. SOFPNN의 구조는 퍼지 다항식 뉴론(FPN)들로 구성되어 있으며, 층이 진행하는 동안 모델 스스로 노드의 선택과 제거를 통해 최적의 네트워크 구조를 생성할 수 있는 유연성을 가지고 있다. 그러나, 노드의 입력변수의 수와 규칙 후반부 다항식 차수 그리고 입력변수는 설계자의 경험 또는 반복적인 학습을 통해 선호된 네트워크 구조를 선택하였으나, 최적의 네트워크 구조를 구축하는데는 어려옴이 내재되어 있었다. 본 논문에서는 자기구성 퍼지 다항식 뉴럴네트워크(Self-Organizing Fuzzy Polynomial Neural Networks： SOFPNN)을 최적화시키기 위해 유전자 알고리즘을 이용하여 자기구성 퍼지 다항식 뉴럴 네트워크의 입력변수의 수와 이에 해당되는 입력변수 그리고 규칙 후반부 다항식의 차수를 탐색하여 최적 의 자기구성 퍼지 다항식 뉴럴 네트워크를 구축한다. 따라서 모델 구축에 있어서 유연성과 정확성을 가지며 객관적이고 좀 더 정확한 예측 능력을 가진 SOFPNN 모델 구조를 구축할 수가 있다.

• The Journal of Engineering Geology
• /
• v.15 no.2 s.42
• /
• pp.123-132
• /
• 2005

This study presents the results of a laboratory investigation performed to study physical properties of soil-bentonite mixtures through the vertical permeation test and dielectric measurement test using Frequency Domain Reflectometry system for the liner of waste landfill. For the laboratory experiments, geotechnical testing was conducted on pre-mixed soil-bentonite which is consisted of standard sand, weathered granite soil and bentonite for estimating physical parameters such as a volumetric water content, void ratio and dielectric constant. In experiment results, initial soil-bentonite mixing rate has an effect of change of volumetric water content. Also change of volumetric water content of a soil-bentonite mixture is clearly detected to measure a response of dielectric constant. In order to estimate an unsaturated permeable property of soil-bentonite mixtures, equations between volumetric water content and dielectric constant were derived from this study.

• Journal of the Korean Society of Groundwater Environment
• /
• v.7 no.1
• /
• pp.1-7
• /
• 2000

In order to reduce the groundwater contamination around the Kimpo landfill in Korea by the leakage of the leachate within it, The method establishing 5 sets of Radial Collector Well Laterals(RCWLs) within the landfill, the method utilizing the wells dewatering the condensed water, the method establishing an interception wall to the 2nd layer at the circumference of the landfill and 22 sets of RCWLs within it, and the method establishing an interception wall to the 3rd layer and 40 sets of RCWLs were studied. Hydraulic parameters were measured for this study and then the groundwater flow and contaminant transport systems around the Kimpo landfill were analyzed with the MODFLOW and MT3D models. Conclusively, the method establishing an interception wall to the 2nd layer and 22 sets of RCWLs was evaluated as the most stable and economical option to reduce groundwater contamination concentration below drinking water standards.

• The Journal of the Acoustical Society of Korea
• /
• v.20 no.8
• /
• pp.12-18
• /
• 2001

In this paper, we propose the Recurrent Neural Predictive HMM (RNPHMM). The RNPHMM is the hybrid network of the recurrent neural network and HMM. The predictive recurrent neural network trained to predict the future vector based on several last feature vectors, and defined every state of HMM. This method uses the prediction value from the predictive recurrent neural network, which is dynamically changing due to the effects of the previous feature vectors instead of the stable average vectors. The models of the RNPHMM are Elman network prediction HMM and Jordan network prediction HMM. In the experiment, we compared the recognition abilities of the RNPHMM as we increased the state number, prediction order, and number of hidden nodes for the isolated digits. As a result of the experiments, Elman network prediction HMM and Jordan network prediction HMM have good recognition ability as 98.5% for test data, respectively.

• Journal of the Korean Geotechnical Society
• /
• v.20 no.3
• /
• pp.5-11
• /
• 2004

Characteristics of advective-diffusive transport of inorganic chemicals in clayey soils as well as in two hardened barrier materials of silica and lime are analyzed from the laboratory column test and compared with those of pure diffusive column test. The results show that the average dispersion coefficients of three materials are $4.39\times l0^{-10}\textrm{m}^2 /s,\; 1.98\times l0^{-10}\textrm{m}^2 /s,\; 1.99\times l0^{-10}\textrm{m}^2 /s$, respectively, and the value of clay is higher than that of hardened barrier materials. There was no significant difference between the dispersion coefficients of advective-diffusive column tests and the effective diffusion coefficients from the pure diffusive column tests, if advective velocity was lower than l0$^{-7}$$m^2$/s. The range of dispersion coefficients of advective-diffusive column tests was narrower than that of diffusion coefficients of pure diffusion tests.

• Journal of Korea Water Resources Association
• /
• v.46 no.12
• /
• pp.1221-1234
• /
• 2013

The purpose of this study is to suggest the methodology for the computation of uplift pressure and discharge of the seepage flow under gravity dam. A 3-dimensional FDM model is developed for this purpose and this model can simulate the saturated Darcian flow in heterogeneous media. For the verification of the numeric model, test simulation has been executed and the mass balance has been checked. The error does not exceed 3%. Using the developed model, The uplift pressure and seepage flow discharge under gravity dam has been calculated. The uplift pressure shows the similar pattern, comparing with the result of flow-net method. As the length of grout curtain increases, the uplift pressure decreases linearly, but the seepage flow discharge shows the non-linear decreasing pattern. The coefficients of the formulas in the dam-design criteria have been analysed, and ${\alpha}=1/3$ corresponds to the value when the length of curtain grout is 70% of the aquifer height. The uplift pressure near the pressure relief drain has the big curvature vertically and horizontally. The developed model in this study can be used for the evaluation of the effects of seepage flow under gravity dam.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.3 no.2
• /
• pp.313-319
• /
• 1999

We devise the 3 models of Multilayered Recurrent Prediction Neural Network(MLRPNN), which are obtained by modifying the Multilayered Perceptron(MLP) with 4 layers. We experimentally study the speech recognition performance of 3 models by a comparative method, according to the variation of the prediction order, the number of neurons in two hidden layers, initial values of connecting weights and transfer function, respectively. By the experiment, the recognition performance of each MLRPNN is better than that of MLP. At the model that returns the output of the upper hidden layer to the lower hidden layer, the recognition performance shows the best value. All MLRPNNs, which have 10 or 15 neurons in the upper and lower hidden layer and is predicted by 3rd or 4th order, show the improved speech recognition rate. On learning, these MLRPNNs have a better recognition rate when we set the initial weights between -0.5 and 0.5, and use the unipolar sigmoid transfer function in the lower hidden layer.

• Journal of Soil and Groundwater Environment
• /
• v.6 no.2
• /
• pp.57-67
• /
• 2001

The purpose of this study is to predict appropriate leachate rates and leachate transport velocity through weathered zone and basement rock on the transient condition at Nanji waste landfill. The leachate transport in the Nanji waste landfill is analyzed using MODFLOW(A Modular 3-D Finite Different Groundwater Flow Model) model which simulates three dimension groundwater flow and MT3D(A Modular Three Dimentional Transport Model) model which describes three dimensional transport for advection, dispersion and chemical reaction of dissolved constituents in groundwater system on the transient condition. Leachate production rates are estimated by HELP(Hydraulical Evaluation of Landfill Performance) model and used weather records for recent 10 years. Leachate transport is predicted by a change of leachate level to after/before established HDPE, established slurry wall and wells.

• Journal of the Korean Geosynthetics Society
• /
• v.15 no.3
• /
• pp.39-47
• /
• 2016

In order to prevent leakage of leachate in offshore landfill, hydraulic barrier is indispensable. In the case of rubble mound revetment at offshore landfill, HDPE sheet in slope should be installed. In this study, seepage analysis were performed to evaluate seepage and flow in the seabed and revetment according to extended length of HDPE sheet by using SEEP/W. Results show that, in the case of low permeability layer is distributed where extended length of HDPE sheet was more than 1m, seepage flux in seabed and revetment was greatly reduced. In order to prevent seepage in seabed and revetment in the case of sand layer of high permeability is existed above low permeable layer, extended length of HDPE sheet and impermeable improvement width of permeable layer should be more than 1m at seabed.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1998.06e
• /
• pp.357-360
• /
• 1998

4층구조인 다층퍼셉트론으로부터 입력층을 제외한 각 측의 출력성분을 하위은닉층으로 귀환하는 3모델의 다층회귀신경망을 구성하고, 각 모델별 망의 크기에 따른 음성인식성능을 분석 비교한다. 과거의 입력신호를 출력층에서 예측하여 오차신호를 계산하고, 이 오차신호가 최소화하는 방향으로 연결세기를 조정한다. 실험결과 3회귀모델중 상위은닉층의 회귀연결방식이 가장 양호한 인식율을 나타내었으며, 각 망 공히 상, 히위은닉층의 뉴런수 10, 15개, 예측차수 3, 4차 일 때 인식성능이 양호하였다. 그리고 회귀신경망이 비회귀신경망에 비해 인식율이 크게 향상된다는 것을 확인 할 수 있었다.