PURPOSES : In this study, a model was developed to estimate the concentrations of particulate matter (PM2.5 and PM10) in expressway tunnel sections. METHODS : A statistical model was constructed by collecting data on particulate matter (PM2.5 and PM10), weather, environment, and traffic volume in the tunnel section. The model was developed after accurately analyzing the factors influencing the PM concentration. RESULTS : A machine learning-based PM concentration estimation model was developed. Three models, namely linear regression, convolutional neural network, and random forest models, were compared, and the random forest model was proposed as the best model. CONCLUSIONS : The evaluation revealed that the random forest model displayed the least error in the concentration estimation model for (PM2.5 and PM10) in all tunnel section cases. In addition, a practical application plan for the model developed in this study is proposed.

PURPOSES : The type and degree of structural conditions and influencing factors distributed across representative sections should be similar to those distributed across entire sections as the representative sections have been predominantly used for developing performance prediction models, which substitute entire sections of road pavement. Therefore, a logic that selects the representative sections with similar distributions of structural conditions and the influencing factors with those of entire expressway asphalt pavement sections requires development. METHODS : The logic developed in this study to select the representative sections of asphalt pavements comprised three steps. First, the data on the structural conditions of the pavement and the influencing climate conditions and pavement materials were collected and organized. Consequently, in the second step, the candidate sections were selected, with the severity of the structural conditions of the pavement distributed widely and evenly. Finally, in addition to the widely and evenly distributed pavement conditions, the representative sections with climatic conditions and pavement materials were selected. RESULTS : A total of 6,352 ordinary asphalt pavement sections and 596 composite asphalt pavement sections were selected as entire expressway asphalt pavement sections and the data were collected and organized according to the logic developed in this study. Three times the representation sections were selected as candidate sections and, finally, 85 sections were selected as representative sections. The distribution of structural conditions and influencing climate conditions and pavement materials in the representative sections were similar to those in the entire sections. In addition, the representative sections were spread evenly across the country. CONCLUSIONS : The sections presenting similar distributions of structural conditions and the influencing factors of entire expressway asphalt pavement sections could be selected in this study. Using the representative sections selected in this study, a remodeling index model will be developed for predicting the asphalt pavement sections that require large-scale repair.

The numerical analysis of two-dimensional transient flow around the obstacle with rotated square cross sections was carried out. The obtained velocity distributions for each time step and each rotation angle were imaged to provide data for CNN(convolutional neural network). Both classification and regression neural networks were used for prediction of rotation angle. As results The classification method incorrectly predicted the rotation angle in only 2 of the 470 images. The regression method predicted the rotation angle errors within except 2 out of 470 images. From these facts, it could be concluded that both methods can be sufficiently applicable to the flow analysis.

PURPOSES : The purpose of this study is to analyze the impact of the level of the light-environment and the driver's visual ability on the change in the driver's perception of a forward curved section at night. The study also aims to identify factors that should be considered to ensure safety while entering curved sections of a road at night. METHODS : Data collected from a virtual driving experiment, conducted by the Korean Institute of Construction Technology (2017), were used. Logistic regression was applied to analyze the effects of changes in the light-environment factors (road surface luminance and glare) and the driver’s visual ability on a driver's perception of the road. Additionally, analysis of the moderated effect of visual ability on light-environment factors indicated that the difference in drivers’ visual abilities impact the influence of light-environment factors on their perception. A driver's ability to perceive, as a response variable, was categorized into 'failure' and 'success' by comparing the perceived distance and minimum reaction sight distance. Covariates were also defined. Road surface luminance levels were categorized into 'unlit road surface luminance' (luminance ≤ 0.1 nt) and 'lit road surface luminance' (luminance > 0.1 nt), based on 0.1 nt, which is the typical level observed on unlit roads. The glare level was categorized as 'with glare' and 'without glare' based on whether the glare was from a high-beam caused by an oncoming vehicle or not. The driver's visual ability level was categorized into 'low visual ability' (age ≥ 50) and 'high visual ability' (age ≤ 49), considering that after the age of 50, the drive’s visual ability sharply declines. RESULTS : The level of road surface luminance, glare, and driver's visual ability were analyzed to be significant factors that impact the driver's ability to perceive curved road sections at night. A driver's perception was found to reduce when the road surface luminance is very low, owing to the lack of road lighting ('unlit road luminance'), when glare is caused by oncoming vehicles ('with glare'), and if the driver's visual ability level is low owing to an older age ('low visual ability'). The driver's ability to perceive a curved section is most affected by the road surface luminance level. The effect is reduced in the order of glare occurrence and the driver's visual ability level. The visual ability was analyzed as a factor that impacts the intensity of the effect of change of the light-environment on the change of the driver's ability to perceive the road. The ability to perceive a curved section deteriorates significantly in 'low visual ability' drivers, aged 50 and above, compared to drivers with 'high visual ability,' under the age of 49, when the light-environment conditions are adverse with regard to the driver’s perception (road surface luminance: 'lit road surface luminance'→'unlit road surface luminance,' glare: 'without glare'→'with glare'). CONCLUSIONS : Supplementation, in terms of road lighting standards that can lead to improvements in the level of light-environment, should be considered first, rather than the implementation of restrictions on the right of movement, such as restricting the passage of low visual ability or aging drivers who are disadvantageous in terms of gaining good perception of the road at night. When establishing alternatives so that safety on roads at night is improved, it is necessary to consider improving drivers' perception by expanding road lighting installation. The road lighting criteria should be modified such that the glare caused by oncoming traffic, which is an influential factor in the linear change in perception, and the level of light-environment thereof are improved.

PURPOSES : The objective of this study is to figure out the trend and characteristics of fine particulate matter (PM2.5) and nitrogen oxide (NOx) concentration in underpass sections. The effect of traffic and meteorological condition on PM2.5 / NOx concentration was analyzed using field monitoring data. METHODS : Based on the literature review, PM2.5 and NOx concentration data were monitored using DustTrak II aerosol monitoring system and Serinus 40 oxides of nitrogen analyzer, respectively. Meteorological and traffic information was collected using automatic weather system and traffic volume counter, respectively. RESULTS : PM2.5 has a positive and negative correlation with relative humidity and wind speed, respectively. Meanwhile, NOx was found to have no correlation with meteorological conditions. The NO/NO2 ratio tends to change with traffic volume, indicating higher correlation between NO and traffic volume; the observed NO2 is mostly a secondary material produced by NO oxidation. CONCLUSIONS : Our study provides clear characteristics of NOx and PM2.5 and correlations with meteorological and traffic information in the underpass sections. It is found from this study that the increase in wind speed causes reduction in the concentration of PM2.5 owing to the diffusion and dispersion phenomena. On the other hand, the meteorological conditions were found to barely have correlations with NOx concentrations in this study. The traffic volume could significantly affect the NOx concentration and NO / NO2 ratio, which is directly correlated to the emissions from vehicles.

국내 하천에서 발생하는 준설 및 보 건설은 하천 연속성 차단과 교란을 유도하여 수서 생물서식환경에 변화를 가져 온다. 본 연구에서는 4대강 보 (이포보, IP; 세종보, SJ; 죽산보, JS; 강정고령보, GG; 달성보, DS)에 서식하는 깔따구 군집 분포를 조사하고 서식환경에 영향을 주는 여러 환경인자를 측정하였다. 조사 지역 중 IP, SJ은 다른 조사 지역에 비해 WT, pH, TOC, Chl-a가 낮은 수준을 보였으며, 깔따구 개체수 결과에서는 Chironominae, Orthocladinae, Tanypodinae 가 비교적 균등한 수준으로 관찰되었다. 반면, JS, GG, DS는 Chironominae가 높은 비율로 우점하며, TOC와 Chl-a의 농도가 높게 나타났다. 각 조사 정점에 대한 깔따구 군집 조성의 특징과 환경요인을 반영한 집괴분석 결과 4대강 보는 3개의 그룹으로 구분되었으며, 이는 정점별 환경 차이와 깔따구의 대악 및 하순기절의 구조에 따른 먹이원 선호도 차이와 일치하였다. 따라서 본 연구에서는 연구 정점 간 먹이원의 차이에 의해 깔따구의 군집 구조의 차이가 나타나는 것을 확인 하였으며, 향후 각 깔따구 과 별 주 먹이원에 대한 연구에 대한 필요성을 제시한다.

본 논문에서는 AISC 표준 단면을 설계 변수로 하는 캔틸레버 타입 헬리데크 모델의 유전 알고리즘 최적설계를 소개한다. AISC 표준 단면을 단면 형상별로 분류하고 단면적 순으로 정렬한 후 정수 단면 번호를 부여하여 설계 변수로 최적설계를 수행하였다. 이 과정을 통하여 이산화된 설계 변수를 가지는 최적설계 문제를 해결하기 위해 유전 알고리즘을 적용하였다. 또한, 제약조건으로 허용응력 및 허용응력비 검사 조건을 모두 고려하여 구조물의 구조 안정성을 고려한 설계를 수행하였다. 최적설계 과정중 매 반복계산 마다 수행되는 구조해석 시간을 단축시키기 위해 선형 중첩법을 사용하였고, 이를 통해 구조 해석 시간을 약 75% 감소시킬 수 있었다. 또한 헬리데크 최적설계의 경량 효과를 높이기 위해 부재 그룹 세분화를 하였고, 그 결과를 선행 연구 모델, 기존의 부재 그룹 모델과 비교하였다. 그 결과 선행연구 대비 약 30톤의 부재를 절감할 수 있었으며, 구조적으로도 보다 안전한 헬리데크 설계를 얻을 수 있었다.

PURPOSES: The objective of this study is to evaluate the structural capacity of asphalt pavement in subsurface cavity sections using falling weight deflectometer (FWD) backcalculation method. METHODS: It is necessary to analyze the reduction of structural capacity in asphalt pavements due to the occurrence of subsurface cavities. The FWD testing was conducted on the cavity and intact asphalt pavement in the city of Seoul. The GAPAVE, backcalculation program for FWD deflections, was utilized to determine the layer moduli in asphalt pavements. The remaining life of asphalt pavements in cavity sections were predicted using the pavement performance model for fatigue cracking. The backcalculated layer moduli between cavity and intact sections were compared to determine the reduction of structural capacity due to subsurface cavity. The relationship between the reduction of layer modulus and cavity depth/length was analyzed to estimate the effect of cavity characteristics on the structural capacity degradation. RESULTS: According to the FWD backcalculation results, the modulus of asphalt layer, subbase, and subgrade in cavity sections are generally lower than those in intact sections. In the case of asphalt layers, the backcalculated modulus in cavity section was reduced by 50% compared to intact section. A study for the prediction of remaining life of cavity section shows that the occurrence of subsurface cavity induces the decrease of the pavement life significantly. It is found that there is no close relationship between the backcalculated modulus and cavity length. However, the reduction of asphalt layer modulus is highly correlated with the cavity depth and was found to increase with the decrease of cavity depth. CONCLUSIONS : This reduction of structural capacity due to the occurrence of cavities underneath asphalt pavements was determined using FWD backcalculation analysis. In the future, this approach will be utilized to establish the criteria of road collapse risk and predict the remaining life of cavity sections under numerous varied conditions.

The purpose of this study is to compare securing of walking roads’ cross section zones in the walking paths by street by street and city by city. The cities involved with this study are Seoul Metropolitan Area and Tokyo Metropolitan Area. The securing of widths are compared based on zone types such as frontage, pedestrian, and furniture zones. The widths are measured directly through on-site surveys. The widths measured are compared the standards given by the present paper. Among 22 sites, the satisfied sites are counted. In terms of frontage zones in Seoul Metropolitan Area, 17 out of 22 sites are counted as satisfied while in Tokyo Metropolitan Area, 18 out of 22 sites are counted as satisfied. In terms of pedestrian zone, in Seoul Metropolitan Area, 13 out of 22 sites are counted as satisfied while in Tokyo Metropolitan Area, 14 out of 22 sites are counted as satisfied. In terms of furniture zone, in Seoul Metropolitan Area, 14 out of 22 sites are counted as satisfied while in Tokyo Metropolitan Area, 11 out of 22 sites are counted as satisfied. In conclusion, although walking roads’ cross section zones such as frontage, pedestrian, and furniture zones are relatively a new standards in Korea and Japan, many sites are founded being satisfied with the standards. However, the rates of satisfaction differs city by city. Especially the rates of satisfaction in furniture zones in Seoul Metropolitan are somewhat higher while the rates of satisfaction in frontage and pedestrian zones are similar between two cities.

In recent years, pavement distresses have been caused by diverse factors such as spalling, deterioration of repaired sections, blow-up, and alkali aggregate reaction due to changing climate environment of a concrete pavement and its construction and maintenance conditions (supply of materials, increase in use of de-icers, etc,). As a leading repair method for deteriorated concrete pavements, partial-depth repair is implemented in accordance with guidelines of material properties for joints of a concrete pavement and field application evaluation systems, but still some of the repaired sections become deteriorated again at an early stage due to poor construction quality and failure of response to environmental impacts. Distresses that can be corrected with partial-depth repairs are largely divided into those of repair materials and of the existing pavement bonded to repair materials, and combined distress of repair materials and the existing pavement. Although re-repair methods should be different by distress type and scale than conventional pavement repair methods, appropriate repair methods and guidance for re-repairs have not been in place so far, and therefore currently, re-repair practices follow the existing manual of partial depth repairs. Therefore, this study evaluated concrete bond characteristics by removing method and repair scope for an experimental section of frequently distressed pavements to determine a re-repair scope and method for deteriorated partial depth repair sections of concrete pavement, the number of which has increased over time.

Highly developed, European industrial countries normally have a dense network of rural highways. The future focus of roadwork in Germany, where there is a high density of 1.6 km of rural highway per km² of land, will involve upgrading and rebuilding rural highways prone to accidents. As new guidelines are not completely relevant for upgrading existing highways and exceptional permits are often required, there is an increasing need to use simulation in the design process to assess discrepancies from the standards and their effect on traffic safety. Using virtual driving (automatically using a defined driver profile or a driving simulator), it is possible to objectively assess expected driving behaviour using characteristic feature graphs and quantitative parameters derived from them. Zwickau University of Applied Sciences has developed a new multi-stage methodology for upgrading rural highways by using a 3D design working place simulator. The highway is automatically retraced using survey data and genetic algorithms in stage 1 (preliminary review) and is broken down into the horizontal and vertical projection design elements. Stage 2 involves determining the shortcomings on the road by superimposing the design elements, the actual accident figures and virtual driving runs. The actual replanning work occurs in stage 3, taking into account the minimum and maximum standard figures for the design elements. The replanning work is feasible, even if the standard figures are not met along parts of the route, but the graph figures, parameters and virtual driving runs prove that there is no reduction in traffic safety. Otherwise, the route needs to be realigned. After finishing the testing period of the hard- and software components, the methodology and the working place can be used from designing engineers for the preparing process of practical projects.

In this paper, a local deformation effect in thin-walled box beams is investigated via a finite element modal analysis. The analysis is carried out for single-cell and multi-cell box beam configurations. The single-cell box beam with and without a neck, which mimics a simple wind-turbine blade, is analyzed first. The results obtained by shell elements are compared to those of one-dimensional(1D) beam elements. It is observed that the wall thickness plays a crucial role in the natural frequencies of the beam. The 1D beam analysis deviates from the shell analysis when the wall thickness is either thin or thick. The shell modes(local deformations) are dominant as it becomes thin, whereas the shear deformation effects are significant as it does thick. The analysis is extended to the single-cell box beam with a neck, in which the shell modes are confined to near the neck. Finally the multi-cell box beam with a taper, which is quite similar to real wind-turbine blade configuration, is considered to investigate the local deformation effect. The results reveal that the 1D beam analysis cannot match with the shell analysis due to the local deformation, especially for the lagwise frequencies. There are approximately 5∼7% errors even if the number of segments is increased.