• The Korean Journal of Rheology
• /
• v.10 no.2
• /
• pp.74-81
• /
• 1998

사출성형품의 설계는 그 내부의 기계적 물성 변화보다는 전통적으로 용도에 부합하 는 형상을 위주로 하여 이루어져 왔기 때문에 설계조건의 개선을 통하여 성능이 우수한 제 품을 얻기까지 많은 시행착오가 요구되고 있다. 그런데 사출성형 실험이나 물성평가 시험을 하기 전에 성형품의 부위별 기계적 물성을 알수있다면 제품의 설계나 금형 설계에 많은 도 움이 될 수 있으므로 최근에 물성 예측을 위한 방법론들의 개발이 다양하게 시도되고 있다. 따라서 본 연구에서는 학습시스템, 사출성형 수치모사와 기계적 물성과의 상관관계를 밝히 는 방법을 만들어 사출물이 제작되기 전에 그들의 기계적 물성을 사출성형 수치모사에서 얻 어진 열적·기계적 이력으로부터 예측하고자 하였다. 이때 성형품의 기계적 물성과 열적· 기계적 이력 사이에는 매우 복잡하고 비선형적인 상관관계를 보이기 때문에 이들 사이를 비 매개변수적으로 연관짓기 위하여 역전파 인공신경망 알고리듬을 사용하였으며 열적·기계적 이력은 사출성형용 수치모사 소프트웨어를 이용하여 구하였다. 학습과정에서 전역최소값에 도달하지 못하는 인공신경망의 문제점을 해결하기 위하여 모멘텀변수와 잡음지수를 포함하 는 일련의 항을 첨가하여 연결가중치를 보정하였다. 그 결과 어떤 초기값에 의하여 학습이 되더라도 전역최소값에 도달하는 것을 확인하였으며 이를 이용하여 다른 사출조건에서 사출 물의 기계적 물성을 잘 예측할수 있었다.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.16 no.5
• /
• pp.562-567
• /
• 2006

Model Predictive Control(MPC) system uses the current input which minimizes the difference between the desired output and the estimated output in the receding horizon scheme. In many cases (for example, system with constraints or nonlinear system), however, it is not easy to find the optimal solution to the above problem. In this paper, we show that real number coding genetic algorithm can be used to solve the optimal problem for MPC effectively. Also, we show by simulation that the real coding algorithm is mote natural and advantageous than the digital coding one.

• The Korean Journal of Applied Statistics
• /
• v.35 no.4
• /
• pp.517-525
• /
• 2022

In the era of big data, various data mining techniques have been proposed as major analysis methodologies. As complex and diverse data is mass-produced, data mining techniques have attracted attention as a method that forms the foundation of data science. In this paper, we focused on the decision tree, which is frequently used in practice and easy to understand as one of representative data mining methods. Specifically, we analyzed the effect of the splitting method of decision trees on the model performance. We compared the prediction power and structures of decision tree models with different split methods based on various simulated data. The results show that the linear combination split method can improve the prediction accuracy of decision trees in the case of data simulated from nonlinear models with complex structure.

• Korean Journal of Construction Engineering and Management
• /
• v.10 no.4
• /
• pp.87-100
• /
• 2009

The quality of early cost estimates is critical to the feasibility analysis and budget allocation decisions for public capital projects. Various researches have been attempted to develop cost prediction models in the early stage of a construction project. However, existing studies are limited on its applicability to actual projects because they focus primarily on a specific phase as well as utilize restricted information while the amount of information collectable differs from one another along with the project stages. This research aims to develop two-staged cost estimation model for the schematic planning and preliminary design process of a construction projects, considering the available information of each phase. In the schematic planning stage where outlined information of a project is only available, the Case-Based Reasoning model is used for easy and rapid elicitation of a project cost based on the extensive database of more than 90 actual highway construction projects. Then, the representing quantity-based model is proposed for the preliminary design stage where more information on the quantities and unit costs are collectable based on the alternative routes and cross-sections of a highway project. Real case studies are used to demonstrate and validate the benefits of the proposed approach. Through the two-stage cost estimation system, users are able to hold a timely prospect to presume the final cost within the budge such that feasibility study as well as budget allocation decisions are made on effectively and competitively.

• Journal of the Korea institute for structural maintenance and inspection
• /
• v.13 no.5 s.57
• /
• pp.149-159
• /
• 2009

The main purpose of this study is to evaluate the seismic safety of Korean R/C school buildings built in the 1980s, based on "the Japanese Standard for Evaluation of Seismic Capacity of Existing R/C Buildings", the nonlinear static and the nonlinear dynamic analyses. The evaluation result of the Japanese Standard showed that R/C school buildings built in the 1980s have 0.2 through 0.4 of seismic indices($I_S$). This result indicates that more than medium damage could be potentially occurred under a medium intensity level of ground motion(150g). The results of the nonlinear analyses and the post-earthquake damage evaluation method showed that Korean R/C school buildings can be suffered moderate and severe damages under a 150gal and a 200gal intensity levels of ground motions, respectively. These results reveal that R/C school buildings should be urgently required a actual earthquake preparedness measures including seismic strengthening for future earthquake.

• Journal of Korean Society of Environmental Engineers
• /
• v.28 no.11
• /
• pp.1198-1206
• /
• 2006

We have carried out highway traffic noise prediction and measurement for 10 sites with representative road shapes and structures. A road traffic noise prediction model(NIER('99)) has been developed for environmental impact assessment in Korea. With the fitted regression analysis, the distribution ratio($R^2$) and Pearson correction coefficient(r) was 92.4% and 0.96 in $1^{st}$ floor, 38.7% and 0.66 in $3^{rd}$ floor, 42% and 0.65 in $5^{th}$ floor, 7.5% and 0.27 in $7^{th}$ floor, 28.4% and 0.53 in 10th floor, 35.6% and 0.60 in $13^{th}$ floor, 52.7% and 0.73 in $15^{th}$ floor, respectively. The measured values of the noise level except the 1st floor did not show a good agreement with the predicted noise level in the NIER('99) formula. Also, the NIER('99) formula demonstrated that the measured values weren't reasonably close to the predicted values, indicating the validity and adequacy of the predicted models with the fitted vs residual analysis in the 95% of confidence interval and 95% of predict interval. Using the equal variation on the basis of the residual vs fitted value, there was the significant difference for variation between $3^{rd}$ floor and $15^{th}$ floor except $1^{st}$ floor. The results suggested that the NIER('99) model obtained by the results according to the apartment floor must be improved and developed on the road traffic noise.

• Journal of Soil and Groundwater Environment
• /
• v.12 no.4
• /
• pp.54-59
• /
• 2007

In this study, a semi-analytical model to address groundwater level fluctuations in response to precipitations and its infiltration is developed through mathematical modeling based on water balance equation. The developed model is applied to a prediction of groundwater level fluctuations in Hongcheon area. The developed model is calibrated through a nonlinear parameter estimator by using daily precipitation rates and groundwater fluctuations data of a same year 2003. The calibrated input parameters are directly applied to the prediction of groundwater fluctuations of year 2004 and the simulated curve successfully mimics the observed. The developed model is also applied to practical problems such as a prediction of a effect of reduced recharge due to surface coverage change and a induced water level reduction. Through this study, we found that recharge to precipitation ratio is not a constant and may be a function of a precipitation pattern.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.10
• /
• pp.179-184
• /
• 2021

In this paper, we propose the deep learning-based neural network model to predict bunker price. In the shipping industry, since fuel oil accounts for the largest portion of ship operation costs and its price is highly volatile, so companies can secure market competitiveness by making fuel oil purchasing decisions based on rational and scientific method. In this paper, short-term predictive analysis of HSFO 380CST in Singapore is conducted by using three recurrent neural network models like RNN, LSTM, and GRU. As a result, first, the forecasting performance of RNN models is better than LSTM and GRUs using long-term memory, and thus the predictive contribution of long-term information is low. Second, since the predictive performance of recurrent neural network models is superior to the previous studies using econometric models, it is confirmed that the recurrent neural network models should consider nonlinear properties of bunker price. The result of this paper will be helpful to improve the decision quality of bunker purchasing.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2021.06a
• /
• pp.333-333
• /
• 2021

최근 전 지구적인 기후변화 및 온난화의 영향으로 태풍 및 집중호우가 빈번하게 일어나고 있으며, 이로 인한 한천범람 등 홍수재해로 인명 및 재산 피해가 크게 증가하고 있다. 우리나라에서도 태풍 및 집중호우로 인한 호수피해는 매년 발생하고 있으며, 피해 빈도와 강도가 증가하고 있는 실정이다. 이러한 현실을 고려하였을 때에 하천 인근 주민의 생명과 재산을 보호하기 위하여 실시간으로 홍수위 예측을 수행하는 것은 매우 중요하다 할 수 있다. 국내에서 수위예측을 위하여 대표적으로 저류함수모형(Storage Function Model, SFM)을 채택하고 있지만, 유역면적이 작아 홍수 도달시간이 짧은 중소하천에서는 충분한 선행시간과 정확도를 확보하기 어려운 문제점이 있다. 이는 유역면적이 작은 중소하천에서는 유역 및 기상 특성과 관련된 여러 인자 사이의 비선형성이 대하천 유역에 비해 커지는 문제점이 있기 때문이다. 본 연구에서는 위와같은 문제를 해결할 수 있도록, 수문자료와 딥러닝 기법을 적용하여 실시간으로 홍수위를 예측할 수 있는 방법론을 제시하였다. 지난 태풍 및 집중호우로 인하여 급격한 수위상승이 있던 낙동강 지류하천에 대하여 LSTM(Long-Short Term Memory) 모형 기반 실시간 수위예측 모형을 개발하였으며, 선행시간 30~180분 별로 홍수위를 예측하고 관측 수위와 비교함으로써 모형의 적용성을 검증하였다. 선행시간 180분 기준으로 영강 유역 수위예측 결과와 실제 관측치의 평균제곱근 오차는 0.29m, 상관계수는 0.92로 나타났으며, 밀양강 유역의 경우 각각 0.30m, 0.94로 나타났다. 본 연구에서 제시된 딥러닝 기반모형에 10분 단위 실시간 수문자료가 입력된다면, 다음 관측자료가 입력되기 전 홍수예측 결과가 산출되므로 실질적인 홍수예경보체계에 유용하게 사용될 수 있을 것이라 보인다. 모형에 적용할 수 있는 더욱 다양한 수문자료와 매개변수 조정을 통하여 예측결과에 대한 신뢰성을 더욱 높일 수 있다면, 기존의 저류함수모형과 연계하여 홍수대응 능력을 향상시키는데 도움이 될 수 있다.

• Tunnel and Underground Space
• /
• v.23 no.1
• /
• pp.54-65
• /
• 2013

The LCM (Linear Cutting Machine) test is one of the most powerful and reliable methods for designing the disc cutter and for predicting the TBM (Tunnel Boring Machine) performance. It has an advantage to predict the actual load on disc cutter from the laboratory test on the real-size large rock samples, however, it also has a disadvantage to transport and/or prepare the large rock samples and to need an extra cost for experiment. In order to overcome this problem, lots of numerical studies have been performed. In this study, the PFC3D (Particle Flow Code in 3 Dimension) has been adopted for numerical analysis on optimum cutter spacing and failure aspects of Busan Tuff. The optimum cutting condition with s/p ratio of 16 and minimum specific energy of $14MJ/m^3$ was derived from numerical analyses. The cutter spacing for Busan Tuff had the good agreements with those of LCM test and numerical analysis by finite element method.