최근 급격한 기후 변화로 인해 도로 교통사고의 발생 빈도가 증가하고 있으며, 특히 겨울철에 자주 발생하는 도로 살얼음(블랙아이 스) 현상이 주요 원인 중 하나로 지목되고 있다. 도로살얼음의 형성 메커니즘은 다양한 요인에 따라 복합적으로 작용하며, 당시의 도 로 기상 조건과 도로의 기하학적 구조에 따라 얼음의 형태 및 강도가 결정된다. 그중에서도 도로 노면 온도는 도로살얼음 형성에 중 요한 요소로, 여러 나라에서 겨울철 교통안전 평가를 위한 주요 지표로 사용되고 있다. 그러나 현재 도로 노면 온도에 대한 명확한 정 의가 부족할 뿐만 아니라, 측정 방법에 따라 계측 편차와 온도 손실 등 여러 한계가 존재해 정확한 온도 측정이 어려운 실정이다. 이 에 본 연구는 지중 깊이에 따른 온도 데이터와 도로 기상 데이터를 결합하여 보다 정밀한 도로 노면 온도 예측 방법을 제시하는 것을 목적으로 한다. 연구를 위해 지중 깊이 2cm, 3cm, 4cm, 5cm, 7cm, 9cm, 15cm, 20cm에 각각 온도 센서를 설치하였으며, 기상 데이터는 해당 지점에서 2m 떨어진 AWS(Automatic Weather System)를 통해 대기 온도, 습도, 강수량, 일사량 등의 정보를 수집하였다. 이를 바 탕으로 지중 온도와 기상 조건의 상관관계를 활용하여 노면 온도를 예측하는 방법론을 도출하였다. 본 연구의 결과는 도로 노면 온도 예측의 정확성을 향상시킬 뿐만 아니라, 새로운 접근 방식을 통해 노면 온도의 정의를 재정립하는 데 기여할 것으로 기대된다.

도로포장의 노면 마찰력은 노면 조직 특성에 큰 영향을 받으며, 이를 예측하기 위한 인자로써 MTD(Mean Texture Depth, 평균 노면 조직 깊이)를 주로 사용한다. 그러나 MTD는 노면 특성 중 노면 조직의 깊이만을 나타내므로 여러 요인이 복합적으로 구성되어 있는 노면 조직 특성을 포괄적으로 설명할 수 없다. 이에 선행 연구에서는 다양한 노면조직 특성을 반영하여 보다 적합한 마찰력 예측 식 을 제안하고자 하였다. 노면 마찰력에 영향을 미치는 노면 조직 특성을 정량화하기 위하여 3D 프린팅 시편을 제작한 후 BPT(British Pendulum Tester)를 이용해 노면 마찰력(BPN; British Pendulum Number)을 측정하였다. 선행 연구를 통하여 노면 마찰력에 영향을 미치 는 노면 조직을 MTD, EAN(Exposed Aggregate Number, 골재노출도) 및 골재 형상으로 선정하였으며, 이를 포함한 노면 마찰력 예측 식 을 제안하였다. 그러나 3D 프린팅 시편을 사용하여 제안한 노면 마찰력 예측식의 경우 이상적인 노면조직 특성에 기반하여 제안된 것 으로 실제 현장에서의 노면 조직 특성과 비교 및 검증이 이루어져야 한다. 이에 3D 프린팅 시편을 기반으로 개발된 노면 마찰력 예측 식의 현장 적용성 평가를 위하여 EACP(Exposed Aggregate Concrete Pavement), 밀입도 및 개립도 아스팔트 콘크리트 포장에서 188개의 노면 조직 데이터를 측정하였다. 현장 측정 데이터와 3D 프린팅 시편을 기반으로 개발된 노면 마찰력 예측 식을 비교 검토한 결과 MTD, EAN 및 골재 형상은 노면 마찰력 예측에 있어서 유의미한 지표로 사용될 수 있는 것으로 확인하였다.

기존 신호제어기법은 과거 주기에 파악된 교통상황을 바탕으로 다음 주기의 교통신호시간을 설계하는 방식으로 신호시간을 설계하기 위해 관측할 때의 교통상황과 신호시간을 제공받는 교통상황 간의 간극이 존재하였다. 또한, 설정된 주기길이 동안 차량이 교차로에 일정하게 도착하는 균일분포를 가정하지만, 실제 교차로에 도착하는 교통량의 행태는 비 균일분포로 실제 교통수요에 대응하기 어렵 다는 한계가 존재한다. 본 연구는 이러한 한계를 극복하기 위해 교차로로 진입하는 상류 교차로의 교통정보를 활용하여 단기 미래 도 착 교통량 예측모델 개발을 통해 관측 시점과 제공 시점 간의 간극을 최소화한다. 또한, 기존 주기길이 동안의 교통량 도착분포를 비 균일분포로 가정하여 주기길이가 고정되지 않는 방식(Acyclic)의 적응식 신호제어 기법(ATC) 개발한다. 제안된 단기 미래 도착 교통 량 예측모델은 실제 스마트교차로 자료를 가공하여 시뮬레이션을 통하여 학습데이터를 구축하여 장단기 메모리(LSTM) 모형과 시간 분산(TimeDistributed) 모형을 적용하여 딥러닝 모델을 개발하였다. 적응식 교통신호제어 기법은 실시간 예측 교통량을 활용하여 교통 류별 예측 지체 산출을 통하여 지체가 최소화되는 현시 종료 지점에서 현시를 종료하고 다음 시간 단계에서 예측된 교통량을 통해 최 적 현시를 재산출하는 롤링 호라이즌(Rolling Horizon)을 수행한다. 제안 신호제어 기법의 평가를 위해 미시적 교통 시뮬레이션을 활 용하여 기존 신호제어 기법인 TOD 신호제어 기법과 제안기법 간의 평가를 수행하였다.

This study aims to estimate the scope of damage impact with a real-life explosion case and a damage prediction program (ALOHA) and suggest measures to reduce risk by comparing and analyzing the results using a Probit model. After applying it to the ALOHA program, the toxicity, overpressure, and radiant heat damage of 5 tons of storage scopes between 66 to 413 meters, and the real-life case also demonstrated that most of the damage took place within 300 meters of the LPG gas station. In the Probit analysis, the damages due to radiant heat were estimated as first-degree burns (13-50%), while structural damage (0-75%) and glass window breakage (94-100%) were expected from overpressure, depending on the storage volume. After comparing the real-life case and the damage prediction program, this study concluded that the ALOHA program could be used as the scope of damage impacts is nearly the same as the actual case; it also concluded that the analysis using the Probit model could reduce risks by applying calculated results and predicting the probability of human casualties and structural damages.

Dynamic responses of nuclear power plant structure subjected to earthquake loads should be carefully investigated for safety. Because nuclear power plant structure are usually constructed by material of reinforced concrete, the aging deterioration of R.C. have no small effect on structural behavior of nuclear power plant structure. Therefore, aging deterioration of R.C. nuclear power plant structure should be considered for exact prediction of seismic responses of the structure. In this study, a machine learning model for seismic response prediction of nuclear power plant structure was developed by considering aging deterioration. The OPR-1000 was selected as an example structure for numerical simulation. The OPR-1000 was originally designated as the Korean Standard Nuclear Power Plant (KSNP), and was re-designated as the OPR-1000 in 2005 for foreign sales. 500 artificial ground motions were generated based on site characteristics of Korea. Elastic modulus, damping ratio, poisson’s ratio and density were selected to consider material property variation due to aging deterioration. Six machine learning algorithms such as, Decision Tree (DT), Random Forest (RF), Support Vector Machine (SVM), K-Nearest Neighbor (KNN), Artificial Neural Networks (ANN), eXtreme Gradient Boosting (XGBoost), were used t o construct seispic response prediction model. 13 intensity measures and 4 material properties were used input parameters of the training database. Performance evaluation was performed using metrics like root mean square error, mean square error, mean absolute error, and coefficient of determination. The optimization of hyperparameters was achieved through k-fold cross-validation and grid search techniques. The analysis results show that neural networks present good prediction performance considering aging deterioration.

This study aims to develop a regression model using data from the Ammunition Stockpile Reliability Program (ASRP) to predict the shelf life of 81mm mortar high-explosive shells. Ammunition is a single-use item that is discarded after use, and its quality is managed through sampling inspections. In particular, shelf life is closely related to the performance of the propellant. This research seeks to predict the shelf life of ammunition using a regression model. The experiment was conducted using 107 ASRP data points. The dependent variable was 'Storage Period', while the independent variables were 'Mean Ammunition Velocity,' 'Standard Deviation of Mean Ammunition Velocity,' and 'Stabilizer'. The explanatory power of the regression model was an R-squared value of 0.662. The results indicated that it takes approximately 55 years for the storage grade to change from A to C and about 62 years to change from C to D. The proposed model enhances the reliability of ammunition management, prevents unnecessary disposal, and contributes to the efficient use of defense resources. However, the model's explanatory power is somewhat limited due to the small dataset. Future research is expected to improve the model with additional data collection. Expanding the research to other types of ammunition may further aid in improving the military's ammunition management system.

The ocean is linked to long-term climate variability, but there are very few methods to assess the short-term performance of forecast models. This study analyzes the short-term prediction performance regarding ocean temperature and salinity of the Global Seasonal prediction system version 5 (GloSea5). GloSea5 is a historical climate re-creation (2001-2010) performed on the 1st, 9th, 17th, and 25th of each month. It comprises three ensembles. High-resolution hindcasts from the three ensembles were compared with the Array for Real-Time Geostrophic Oceanography (ARGO) float data for the period 2001-2010. The horizontal position was preprocessed to match the ARGO float data and the vertical layer to the GloSea5 data. The root mean square error (RMSE), Brier Score (BS), and Brier Skill Score (BSS) were calculated for short-term forecast periods with a lead-time of 10 days. The results show that sea surface temperature (SST) has a large RMSE in the western boundary current region in Pacific and Atlantic Oceans and Antarctic Circumpolar Current region, and sea surface salinity (SSS) has significant errors in the tropics with high precipitation, with both variables having the largest errors in the Atlantic. SST and SSS had larger errors during the fall for the NINO3.4 region and during the summer for the East Sea. Computing the BS and BSS for ocean temperature and salinity in the NINO3.4 region revealed that forecast skill decreases with increasing lead-time for SST, but not for SSS. The preprocessing of GloSea5 forecasts to match the ARGO float data applied in this study, and the evaluation methods for forecast models using the BS and BSS, could be applied to evaluate other forecast models and/or variables.

본 연구는 돼지 간 거리(PD), 돈사 내 상대 습도(RRH), 돈사 내 이산화탄소(RCO2) 세 가지 변수를 사용하여, 네 개의 데이터 세트를 구성하고, 이를 다중 선형 회귀(MLR), 서포트 벡터 회귀(SVR) 및 랜덤 포레스트 회귀(RFR) 세 가지 모델 기계학습(ML)에 적용하여, 돈사 내 온도(RT)를 예측하고자 한다. 2022년 10월 5일부터 11월 19일까지 실험을 진행하였다. Hik-vision 2D카메라를 사용하여, 돈사 내 영상을 기록하였다. 이후 ArcMap 프로그램을 사용하여, 돈사 내 영상에서 추출한 이미지 안 돼지의 PD를 계산하였다. 축산환경관리시스템(LEMS) 센서를 사용하여, RT, RRH 및 RCO2를 측정하였다. 연구 결과 각 변수 간 상관분석 시 RT와 PD 간의 강한 양의 상관관계가 나타났다(r > 0.75). 네 가지 데이터 세트 중 데이터 세트 3을 사용한 ML 모델이 높은 정확도가 나타났으며, 세 가지 회귀 모델 중에서 RFR 모델이 가장 우수한 성능을 보였다.

In this study, the magnetocaloric effect and transition temperature of bulk metallic glass, an amorphous material, were predicted through machine learning based on the composition features. From the Python module ‘Matminer’, 174 compositional features were obtained, and prediction performance was compared while reducing the composition features to prevent overfitting. After optimization using RandomForest, an ensemble model, changes in prediction performance were analyzed according to the number of compositional features. The R2 score was used as a performance metric in the regression prediction, and the best prediction performance was found using only 90 features predicting transition temperature, and 20 features predicting magnetocaloric effects. The most important feature when predicting magnetocaloric effects was the ‘Fe’ compositional ratio. The feature importance method provided by ‘scikit-learn’ was applied to sort compositional features. The feature importance method was found to be appropriate by comparing the prediction performance of the Fe-contained dataset with the full dataset.

본 논문에서는 다목적 구조물인 다중연결 해양부유체를 대상으로 변형 기반 모드 차수축소법을 적용하고 차수축소모델의 구조응 답 예측 성능을 향상시키기 위해 유전 알고리즘 기반의 센서 배치 최적화를 수행하였다. 다중연결 해양부유체의 차수축소모델 생성 에 필요한 변형 기반 모드 데이터를 얻기 위해 다양한 규칙파랑하중조건에 대한 유체-구조 연성 수치해석을 수행하고 변형 기반 모드 의 직교성, 자기상관계수를 이용하여 주요 변형 기반 모드를 선정하였다. 다중연결 해양부유체의 경우 차수축소모델의 구조응답 예 측 성능이 계측 및 예측 구조응답 위치에 따라 민감하기 때문에 유전 알고리즘 기반의 최적화를 수행하여 최적의 센서 배치를 도출하 였다. 최적화 결과, 모든 센서 배치 조합에 대한 차수축소모델 생성 및 예측 성능 평가 대비 약 8배의 계산 비용을 절감하였으며, 예측 성능 평가 지표인 평균 제곱근 오차가 초기 센서 배치보다 84% 감소하였다. 또한, 다중연결 해양부유체 모형시험 결과를 이용하여 불 규칙파랑하중에 대한 최적화된 센서 배치의 차수축소모델의 구조응답 예측 성능을 평가 및 검증하였다.

콩과 같은 밭작물은 주로 토양으로부터 수분을 공급받으며 토양 수분 조건에 따라 생육 반응이 민감하게 반응한다. 작물의 생육과 재배 지역의 토양 조건, 기상 등에 따라 적정 토양 수분을 유지하는 것은 작물 생산량의 증가를 위해 중요하다. 따라서, 본 연구에서는 머신러닝 기법을 이용하여 토양 수분 함량 예측 모델을 개발하였다. 깊이에 따른 토양 수분과 외기, 강수량 등 기상 변수와의 상관 관계를 구명하고, 깊이별 토양 수분예측을 위한 부분최소제곱회귀(PLSR) 모델을 알고리즘을 개발하였다. 콩 재배포장의 10cm, 20cm, 30cm 깊이의 토양수분은 FDR 방식의 센서로 측정하였 고, 콩 작물 주변 환경인자(재배환경의 기온, 상대습도, 풍속, 일사량, 일조시간)는 주변의 기상관측소에서 측정된 데이터를 이용하였다. 이를 이용하여 깊이별 미래의 토양수분함량 예측 모델을 개발한 결과, 10cm와 20cm깊이에서 주요 인자는 현재 토양수분함량과 기온이었으며, 30cm 깊이에서의 주요 인자는 현재 토양수분함량과 기온, 풍속으로 나타났다. 토양 깊이가 깊어짐에 따라 토양수분함량 예측 정확도가 향상되었으며, 이는 표면에 가까울수록 토양수분함량이 변화가 크기 때문으로 예상된다. 또한 미래의 토양 수분함량예측시 1시간 후 예측 정확도가 가장 우수하였으며, 이때의 Rv 2와 RMSEV가 10cm 깊이에서 0.993와 1.069%, 20cm 깊이에서 0.994와 0.821% 였으며, 30cm 깊이에서 0.999와 0.149% 였다. 본 연구 결과는 콩 생육환경 진단을 위해 재배 포장의 토양수분함량을 토양층별로 미래의 토양수분함량도 예측이 가능함을 보여준다.

In this paper, machine learning models were applied to predict the seismic response of steel frame structures. Both geometric and material nonlinearities were considered in the structural analysis, and nonlinear inelastic dynamic analysis was performed. The ground acceleration response of the El Centro earthquake was applied to obtain the displacement of the top floor, which was used as the dataset for the machine learning methods. Learning was performed using two methods: Decision Tree and Random Forest, and their efficiency was demonstrated through application to 2-story and 6-story 3-D steel frame structure examples.

Machine learning is widely applied to various engineering fields. In structural engineering area, machine learning is generally used to predict structural responses of building structures. The aging deterioration of reinforced concrete structure affects its structural behavior. Therefore, the aging deterioration of R.C. structure should be consider to exactly predict seismic responses of the structure. In this study, the machine learning based seismic response prediction model was developed. To this end, four machine learning algorithms were employed and prediction performance of each algorithm was compared. A 3-story coupled shear wall structure was selected as an example structure for numerical simulation. Artificial ground motions were generated based on domestic site characteristics. Elastic modulus, damping ratio and density were changed to considering concrete degradation due to chloride penetration and carbonation, etc. Various intensity measures were used input parameters of the training database. Performance evaluation was performed using metrics like root mean square error, mean square error, mean absolute error, and coefficient of determination. The optimization of hyperparameters was achieved through k-fold cross-validation and grid search techniques. The analysis results show that neural networks and extreme gradient boosting algorithms present good prediction performance.

In this study, a new model using artificial neural networks is proposed to improve the thickness error between the plates, which occurs when the rolling conditions change a lot during the thick rolling. The model was developed by using Python, and the input values are the change in the finish rolling temperature between the plates, the change in target tensile strength, the change in target thickness, and the change in rolling force. The new model is 31.76% better than the existing model based on the standard deviation value of the thickness error. This result is expected to reduce quality costs when applied to online models at actual production sites in the future.

Liquefied hydrogen is attracting attention as an energy source of the future due to its hydrogen storage rate and low risk. However, the disadvantage is that the unit price is high due to technical difficulties in production, transportation, and storage. This study was conducted to improve the design accuracy and development period of needle valves, which are important parts with a wide technical application range among liquefied hydrogen equipment. Since the needle valve must discharge an appropriate flow rate of the liquefied fluid, it is important to determine the needle valve design parameters suitable for the target flow rate. Computational Fluid Dynamics and Artificial Neural Network technology used to determine the design variables of fluid flow were applied to improve the setting and analysis time of the parameter. In addition, procedures and methods for applying the design parameter of needle valves to Convolutional Neural Networks were presented. The procedure and appropriate conditions for selecting parameters and functional conditions of the Convolutional Neural Network were presented, and the accuracy of predicting the flow coefficient according to the design parameter was secured 95%. It is judged that this method can be applied to other structures and machines.

PURPOSES : This study aimed to predict the number of future COVID-19 confirmed cases more accurately using public and transportation big data and suggested priorities for introducing major policies by region. METHODS : Prediction analysis was performed using a long short-term memory (LSTM) model with excellent prediction accuracy for time-series data. Random forest (RF) classification analysis was used to derive regional priorities and major influencing factors. RESULTS : Based on the daily number of COVID-19 confirmed cases from January 26 to December 12, 2020, as well as the daily number of confirmed cases in Gyeonggi Province, which was expected to occur on December 24 and 25, depending on social distancing, the accuracy of the LSTM artificial neural network was approximately 95.8%. In addition, as a result of deriving the major influencing factors of COVID-19 through random forest classification analysis, according to the number of people, social distancing stages, and masks worn, Bucheon, Yongin, and Pyeongtaek were identified as regions expected to be at high risk in the future. CONCLUSIONS : The results of this study can help predict pandemics such as COVID-19.

PURPOSES : High temperatures induce excessive expansion in pavements, thus causing the closure of contraction joints between expansion joints. This results in the integration of slabs within the expansion joints into a unified slab. Compressive forces are generated owing to the friction that ensues between the unified slab and lower base layer. As the integrated slab expands and exceeds the allowable width of the expansion joint, the end restraint generates an additional compressive force. The escalating force, which reaches a critical threshold, induces buckling, thus compromising stability and causing blow-up incidents, which poses a significant hazard to road users. The unpredictable nature of blow-up incidents render their accurate prediction challenging because the compressive force within the slab must be predicted and the threshold for blow-up occurrence must be determined. METHODS : In this study, a GWNU blow-up model was developed to predict both the compressive force and period of blow-up incidents in jointed concrete pavements. The climate conditions, pavement structure, materials, and expansion joints were considered in this model. In the first stage of the model, the time at which the integrated slab expanded and surpassed the allowable width of the expansion joint was determined, and the compressive force was calculated. Subsequently, the compressive force within the integrated slab, considering both the end restraints and friction, was predicted. A large-scale blow-up test was performed to measure the blow-up force based on changes in the geometric imperfections. The measured blow-up force was adopted as the blow-up occurrence threshold, and the point at which the predicted compressive force within the slab exceeded the blow-up force was identified as the blow-up occurrence time. RESULTS : Using the GWNU blow-up model, the blow-up occurrence on the Seohean Expressway in Korea is predicted in the presence or absence of the alkali-silica reaction (ASR). Analysis is conducted using the expansion joint spacing and width as variables. As the expansion joint spacing increases, blow-up occurs sooner, and as the width increases, only the expansion joint life decreases. When applying an expansion joint spacing of 300 m and a width of 100 mm under an ASR with 99.9% TTPG reliability, the sum of the expansion joint life and blow-up occurrence time is 16 years. CONCLUSIONS : In the case of jointed concrete pavements where ASR occurred, installing an expansion joint spacing of 300 m and a width of 100 mm does not satisfy the design life of 20 years, and the expansion joint width minimally affect the blow-up occurrence time. To prevent blow-up incidents, a spacing of less than 300 m for the expansion joint is recommended. Based on the analysis results, the blow-up occurrence time and location can be predicted from the characteristics of the installed expansion joint, through which blow-up incidents can be prevented via preliminary maintenance.