https://doi.org/10.12652/Ksce.2024.44.6.0743

박상원(Park, Sangwon) ; 장민우(Chang, Minwoo) ; 야즈단파나 오미드(Yazdanpanah, Omid)

Precise diagnosis of structural condition is essential indicator for the sustained operation of infrastructure. The load tests are widely used to assess the serviceability and safety condition of bridges that are continually exposed to various forms of repeated loading. Recently, advancement in smart diagnostic technologies have introduced methods for assessing structural responses using image data. However, when utilizing image data, challenges such as distortions from lighting and angles, as well as difficulties in collecting information from blind spots, can hinder the comprehensive understanding of structural behavior. To address these limitations, this study presents an artificial intelligence (AI) based method for estimating vertical deflections using strain data measured during static tests on girders. The strain data were obtained from load tests conducted on 11 girders at the Hybrid Structural Testing Center of Myongji University. The deep learning model was developed using a bidirectional Long Short-Term Memory (LSTM) network, with multiple input parameters including girder type, static load, and strain data. The performance of AI model was further enhanced through hyperparameter optimization, and its accuracy was verified through consistency evaluation. Additionally, the effectiveness of the proposed method was evaluated by analyzing prediction accuracy and computational time in relation to the quantity of training data.

https://doi.org/10.12652/Ksce.2024.44.6.0753

최길호(Choi, Kil Ho) ; 이승건(Lee, Seung Geon) ; 박정호(Park, Jung Ho) ; 박성균(Park, Sung Kyun) ; 조성훈(Cho, Sung Hun) ; 박찬민(Park, Chan Min)

Interest in modular floating structures is growing as a solution for efficient utilization of ocean spaces. Modular floating structures are formed by arranging individual modules into a single structure, so careful consideration of the arrangement is necessary in the initial design phase. This study investigates the influence of modular arrangement (orientation, gap, and the number of modules) on the dynamic response of modular floating structures installed in coastal areas. The Jonswap spectrum, which appropriately represents the characteristics of coastal environments, was applied, and numerical analysis focused on the vertical acceleration at the corners of each module. The arrangement analyzed ranged from 1×2 to 1×6 modules, and the connections between the modules were considered as hinged joints. The results showed that the maximum response occurred at a heading angle of 0°, with the highest response at the module first exposed to the waves. Additionally, the response remained relatively unchanged beyond a certain gap, with a significant reduction in response after the first module. This highlights the importance of focusing on vibration control for the first module. These findings provide essential reference data for optimizing the layout strategies in the design of modular floating structures.

https://doi.org/10.12652/Ksce.2024.44.6.0765

김건수(Kim, Kun-Soo) ; 정규산(Jung, Kyu-San) ; 조한민(Cho, Han-Min) ; 박기태(Park, Ki-Tae)

Precast retaining walls using cast-in-place foundation slabs mainly use lap splices, hooks or loops to secure the slab reinforcement, and sometimes couplers. When the retaining wall is subjected to soil load and the foundation soil around the retaining wall is lost, the load is concentrated on the foundation slab, which has a negative effect on the structure. In this study, the effect of contactless rebar couplers on the structural behavior of cold joints in precast concrete retaining walls requiring cast-in-place foundation slabs was experimentally analyzed. The connection conditions of the foundation slabs applied in the full-scale test were lap splice and contactless rebar coupler conditions. The ultimate failure mode of both experimental conditions was haunch failure due to bending moment, and the slopes of the initial load-displacement curves were the same. However, the retaining wall using contactless rebar couplers was confirmed to show higher yield strength and stiffness after the yield stage. In addition, the specimens using non-contact couplers showed better results than the specimens using lap splices by attaching strain gauges to the reinforcement and concrete. It was confirmed that this was due to the difference in joint strength depending on the joint conditions of the field-cast rebar.

https://doi.org/10.12652/Ksce.2024.44.6.0773

김용탁(Kim, Yong-Tak) ; 김진영(Kim, Jin-Young) ; 한영천(Han, Young-Chun) ; 권현한(Kwon, Hyun-Han)

This study presents an areal rainfall analysis conducted using three automated rainfall field exploration methods: Box-tracking, Point-tracking, and Advanced Point-tracking, with a focus on major storm events in South Korea. The characteristics of each method were compared, highlighting their ability to track rainfall distribution. The Box-tracking method allowed for the consideration of multiple storm centers through area expansion but showed limitations in accurately capturing storm shapes. In contrast, the Point-tracking method retained the storm center, enabling detailed representation of the storm shape but faced challenges in analyzing multiple storm centers. The Advanced Point-tracking method addressed these limitations by automatically identifying multiple storm centers and expanding the area to more accurately reflect the actual rainfall distribution. The differences in results provided by each method offered critical insights for understanding the intensity and distribution patterns of rainfall, which are essential for refining extreme rainfall predictions. These findings contribute valuable methodologies for more reliable Depth-Area-Duration (DAD) analysis of extreme rainfall events, particularly in the context of climate change, and provide crucial information for developing effective disaster mitigation strategies.

https://doi.org/10.12652/Ksce.2024.44.6.0781

옥광민(Ok, Gwangmin) ; 이호수(Lee, Hosoo) ; 권영화(Gwon, Yeonghwa) ; 김동수(Kim, Dongsu) ; 김영도(Kim, Young Do)

This study aimed to determine the optimal light source and intensity for monitoring sewer water depth using spectral data. While water depth monitoring methods utilizing spectral data have been continuously researched and applied, most studies have focused on natural light conditions. There is a lack of research in environments such as sewers where specific light sources are absent. The estimation of water depth using spectral data can vary significantly depending on the installed artificial light source, its angle, and intensity. In this study, indoor flume experiments were conducted to select the optimal light source and intensity for accurate water depth estimation. Halogen and LED light sources were used, and spectral data were collected under darkroom conditions using a multispectral camera with various light intensities and sediment heights. The collected multispectral images were processed to calculate water depth using the Optimal Band Ratio Analysis (OBRA) method. The accuracy of water depth estimation was compared across experimental cases to identify the optimal light source and intensity. The experimental results using the LED light source showed that the R2 values ranged widely from a minimum of 0.181 to a maximum of 0.799. Notably, only two cases exhibited an R2 value above 0.7, indicating that the majority of cases demonstrated low accuracy, suggesting that LED light sources provide limited precision in depth estimation. In contrast, the experiments with halogen light sources yielded R2 values ranging from a minimum of 0.800 to a maximum of 0.984, with particularly high accuracy observed in cases with illuminance levels between 600 and 800 lux. This study proposes optimal light sources and illumination levels for depth monitoring using spectral information and is expected to make significant contributions to improving the accuracy of depth monitoring in sewer systems. This translation incorporates more formal and technical language commonly used in academic papers while maintaining clarity and precision.

https://doi.org/10.12652/Ksce.2024.44.6.0793

진종훈(Jin, Jonghun) ; 김동규(Kim, Dongkyu) ; 오상율(Oh, Sangyul) ; 이현지(Lee, Hyeonji)

This study assessed the role and impact of ground rainfall data in adjusting radar rainfall measurements. Radar rainfall is widely used due to the limited spatial resolution and coverage of rain gauges. Due to the inherent uncertainty of radar rainfall, many studies have been done to calibrate it with ground rainfall. While technical methods for improving radar rainfall quality are important, the characteristics of the data used are equally critical but less studied. To assess the impact of ground rainfall data on adjusted results, we selected Gangwon State which is an area with a low density of Korea Meteorological Administration (KMA) rain gauges. We analyzed the effects of utilizing ground rainfall from the Korea Forest Service's mountain meteorological observation network alongside KMA data. The results showed that incorporating both datasets improved accuracy in terms of intensity by 27-47 % (based on the sum of squared errors) and enhanced temporal variability by 7-45 % (based on correlation coefficients) compared to using only KMA data. The diversity and spatial distribution of ground data can enhance the accuracy of adjusted results, especially in areas with complex terrain.

https://doi.org/10.12652/Ksce.2024.44.6.0807

이강민(Lee, Kangmin) ; 정민규(Jung, Minkyu) ; 이주헌(Lee, Joo-Heon) ; 권현한(Kwon, Hyun-Han)

This study aims to evaluate the impact of recent droughts on the variability of current water supply adjustment baselines for dams in response to frequent droughts due to climate change. Using the PARMA model, inflow variability was analyzed for Juam Dam in the southern region, and the sensitivity of water supply adjustment baselines was assessed depending on the inclusion of major drought events. The results showed that increasing the number of stochastic simulations improved the reliability of the water supply adjustment baselines; specifically, while a 15 % difference in reservoir levels was observed in the 100-year simulation, the difference decreased to within 5 % for simulations of 500 years or more, suggesting the use of at least 500-years simulations. Moreover, comparing scenarios including and excluding the drought years of 2017 and 2022, it was found that incorporating drought years led to an average 20 % increase in the target reservoir levels, indicating difficulties in achieving 95 % water supply reliability under current standards. In consecutive multi-year drought scenarios, the target reservoir level increased by up to 30 %, highlighting the need for further discussions on flood management. Based on these findings, the study proposes revising water supply standards to enhance the reliability of dam operations in response to climate change. Future research should focus on developing more sophisticated water resource management strategies and strengthening the national integrated drought response system.

https://doi.org/10.12652/Ksce.2024.44.6.0817

양순보(Yang, Soonbo) ; 문경태(Moon, Kyoungtae) ; 박상렬(Park, Sangyeol) ; 부상필(Boo, Sangpil)

Jeju volcanic rock has a porous structure with irregularly developed vesicles of varying sizes and shapes, resulting from factors such as the physical properties of lava and its flow and cooling processes. Moreover, its mechanical properties exhibit regional variations. In this study, various laboratory tests were conducted on volcanic rock specimens from the northern part of Jeju Island. The relationships between various physical properties, including total porosity, and their correlations with uniaxial compressive strength were examined. The results confirmed that the relationship between absorption and bulk specific gravity of the Jeju volcanic rocks can be divided into two linear relationships (Linear Regression (1) and Linear Regression (2)). In contrast, there was a single linear relationship between total porosity and bulk specific gravity (or dry density) of the Jeju volcanic rocks. Regarding the relationships between uniaxial compressive strength and both bulk specific gravity and absorption of the Jeju volcanic rocks, each rock corresponding to either Linear Regression (1) or Linear Regression (2) exhibited different behaviors. However, unlike the above results, the relationship between uniaxial compressive strength and total porosity showed a single correlation, regardless of the two linear relationships between absorption and bulk specific gravity.

https://doi.org/10.12652/Ksce.2024.44.6.0827

김덕희(Kim, Deokhee) ; 이수일(Lee, Sooil)

Developed countries are trying to form good driving habits through education and strict management policies that allow novice drivers to adapt to various road environments during their early driving period, but our country is only focusing on strengthening the driver's license test. In addition, even if they obtain a driver's license, there are not many drivers who drive right away, so it is difficult to accurately understand the scale and accident characteristics of novice drivers. This study analyzed the accident characteristics of novice drivers based on their starting driving experience using insurance subscription history data. In addition, the characteristics of gaze changes during driving of novice drivers were derived using an eye tracker. The research results can be briefly summarized into three points as follows. First, the accident characteristics of novice drivers are very high at 29.7 % in the first year of insurance subscription, and side collisions and rear-end collisions occur frequently. Second, according to the survey results for beginner drivers, they had difficulty driving in places that require driver judgment, such as unprotected left turns and back roads. Third, the results of the Eye Tracker experiment showed that novice drivers only looked forward while driving, and their field of vision was 18 degrees, which is one-fifth of that of experienced drivers. Therefore, for beginner drivers, it is necessary to reinforce education on how to widen the left and right field of vision while driving and education on the right of way when there is no signal.

https://doi.org/10.12652/Ksce.2024.44.6.0835

조현철(Cho, Hyunchul) ; 이수일(Lee, Sooil)

This study aimed to identify the changes in driver behavior and the impact of reducing accidents due to the operation of ADAS vehicles following the popularization of advanced driver assistance systems (ADAS). Existing studies have focused only on changes in driving behavior after installing ADAS, and the reliability of the accident reduction effect was low due to the small number of experimental vehicles. The data used in this study were the results of the Seoul City corporate taxi pilot test and insurance company accident statistics, which were used to increase the reliability of changes in driving behavior and accident reduction effects.The research results can be summarized in three points as follows. First, taxi drivers' dangerous driving behaviors were high in the following order: right lane departure (28 %), failure to maintain distance between cars (27.4 %), and left lane departure (22.9 %). In driving tests with general drivers, failure to maintain distance between cars was also the most common. Second, after installing ADAS, about 20 % of drivers' risky driving behaviors decreased, and among them, maintaining a safe distance from the car in front improved the most. Third, after installing ADAS, the effect of reducing accidents was minimal in terms of the number of accidents, and the severity of accidents, as analyzed by insurance payments per case, decreased by 42.6 %. Analysis of insurance statistics showed a significant decrease in insurance payments per case for rear-end collisions and side collisions. As autonomous vehicles evolve based on ADAS technology in the future, changes in driver behavior and accidents should be continuously analyzed and utilized in policy making.

https://doi.org/10.12652/Ksce.2024.44.6.0843

김정수(Kim, Jeongsoo)

With the mandatory application of Building Information Modeling (BIM) in construction projects, the analysis of BIM requirements from the bidding stage has become increasingly important. To effectively handle the vast amount of bid text information, research on natural language processing has been actively conducted. However, domestic bid documents often contain a mixture of Korean and English, and the inconsistent representation of identical expressions limits text quality improvement even after pre-processing, which in turn affects the performance of morphological analyzers. These challenges contribute to the degradation of analytical model performance. To mitigate performance degradation caused by pre-processing, this study proposes a deep learning model that classifies the difficulty level of BIM requirements in bid sentences using character embeddings. The study analyzes the impact of key parameters on three types of character embedding models: a simple character embedding, a character embedding combined with the Bi-directional long short-term memory layer, and a character embedding combined with the convolutional neural networks (CNN) layer. Based on this analysis, the optimal model was derived, and the performance of BIM requirement difficulty classification was compared across the character embedding models. Additionally, the results were compared with those of other models based on word embeddings. Among the three models, the character embedding model utilizing CNN achieved the highest F1 score (0.98), with the number and size of convolutional filters having the most significant impact on F1 score variation. Furthermore, it was confirmed that the use of character embedding models resulted in an approximately 15% improvement in the F1 score compared to word embedding models.

https://doi.org/10.12652/Ksce.2024.44.6.0855

김병일(Kim, Byungil) ; 정도범(Chung, Do-Bum)

The financial stability of construction firms constitutes a critical criterion for evaluation by clients and stakeholders. Notably, the issuance of a disclaimer of opinion by auditors can profoundly affect a firm's project sustainability and overall credibility. This study investigates revenue fluctuation patterns preceding the issuance of a disclaimer of opinion to identify early indicators of financial risk in construction firms. A paired sample t-test conducted on seven years of revenue data from 50 construction firms demonstrates a statistically significant decline in revenue during the year immediately preceding the issuance of a disclaimer. Complementary analysis, utilizing five years of data from 83 firms, further reveals a pronounced revenue decline commencing two years prior to the issuance. The findings of this research offer valuable insights for clients and auditors by providing a basis for early detection and assessment of financial risk in construction firms. They underscore the significance of time-series analysis of revenue trends as an integral component of risk management within the construction industry. Given the strong correlation between revenue declines and adverse audit opinions, the study highlights the potential fo r systematic monitoring of such patterns to enhance project success rates. This research contributes to the literature on financial stability and sustainability evaluation in the construction sector and emphasizes the necessity of vigilant scrutiny of revenue trajectories during the auditing process. Moreover, the findings hold promise for advancing predictive models aimed at improving the accuracy of audit opinion forecasts in subsequent studies.

https://doi.org/10.12652/Ksce.2024.44.6.0863

임종원(Lim, Jong Won) ; 정상윤(Jung, Sang Yoon) ; 강인석(Kang, Leen Seok)

In this paper, in order to overcome the limitation caused by the lack of linkage between design BIM and construction BIM, a study was conducted to predict the process sequence using attribute information of the design 3D model. The LSTM algorithm was used as the prediction algorithm, and the research was conducted on a total of three cases. In the case of the same process single prediction LSTM model, the loss value was predicted through the precedence relationship in which the process name and number of the design BIM model were established, and in the case of the same process simultaneous prediction LSTM model, it was confirmed that the number of numbers specified by the user was grouped and predicted. Finally, in the case of the two-way process simultaneous prediction LSTM model, it was found that the prediction was made in the order of the number of processes far from each other in the order of No. 1 abandoned concrete No. 5 and No. 2 No. 4. As a result of conducting three case 4D simulations on the two design BIM models to verify the prediction results, it was confirmed that the design BIM models were simulated in the predicted order.

https://doi.org/10.12652/Ksce.2024.44.6.0871

서동옥(Seo, Dongok) ; 김시곤(Kim, Sigon)

With the development of modern science and technology, the railway sector has also become more automated and advanced in systems, resulting in higher train speeds and mass transportation of passengers. While the probability of accidents has decreased compared to the past, the severity of damage is greatly increasing. With the recent enforcement of the Serious Accident Punishment Act, only one worker can be killed. If an accident caused by human error occurs as an industrial or civil disaster, it will cause great personal and social repercussions, so countermeasures are needed. According to the results of analysis of accidents and failures over the past five years after the integration of the Seoul Metro Rapid Corporation, the highest number of accidents caused by human error occurred in the crew sector (65.3 %), and human errors such as careless handling and negligent verification occurred during repetitive work performance. It appears that it is doing so. Urban railway crew members board and drive moving trains, stopping at each station every 2 minutes and 30 seconds and performing repetitive tasks handling an unspecified number of passengers. Due to the nature of the work, human errors mostly occur during the work process. Work process risk management is necessary. This study is based on the human error analysis (HEA) technique and is a railway crew work process risk assessment measurement tool to identify the degree of accident risk due to human errors such as careless handling and neglect of confirmation for each task during the work process performance of railway crew members. It is a study on development.

https://doi.org/10.12652/Ksce.2024.44.6.0879

문제웅(Moon, Je Woong) ; 김시곤(Kim, Si Gon)

As urban railways are extended and new stations are built, railway operators must establish railway operation plans to protect lives and property from expected disasters and safety accidents and minimize damage from accidents, thereby preventing major disasters and promoting railway safety, and providing convenient and safe railway services to urban railway users. Based on the safety analysis results that were limited to fire scenarios in the Urban Railway Station and Transfer/Convenience Facility Design Guidelines for the basic and detailed design of the 4th section of the Byeollae Line (Amsa~Byeollae) double-track electric railway construction project, a study on safety review scenarios for urban railway stations necessary for railway operation was conducted through the case of Byeollae Line Guri Station. The safety review scenario was divided into normal and disaster situations. The normal situation was classified into Normal hour, Rush hour, and escalator failure situations, and the escalator failure was divided into three types (1 failure, 2 failures, 3 failures). The disaster situation was classified into waiting room fire, Blackout, and flooding. Through this, the safety review scenario was set up, the number of people needed for safety review was calculated, and the urban railway station safety review scenario was developed through the case of Byeollae Line Guri Station.

https://doi.org/10.12652/Ksce.2024.44.6.0887

오주경(Oh, Ju Kyoung) ; 김시곤(Kim, Si Gon)

Currently, high-speed rail KTX has been actively operating since its opening in 2004. However, as the number of operations is mostly concentrated in major bases (metropolitan cities) such as Seoul, Gwangju, and Busan, the demand for use varies greatly according to the number of operations on each line. In this study, by reviewing the current status of KTX operations for high-speed rail, the justification for expanding the number of operations was presented by analyzing the distance per rail extension, reviewing the future development plan, and the utilization rate saturation for areas with a relatively small number of operations. In addition, Pohang Station was selected to conduct a detailed analysis of related routes, and by analyzing the justification for expanding the number of operations, a more rational expansion plan was suggested. As a result of analysis in this study, in the case of the Donghae Line, the number of users per railroad extension (km) was the highest at 49 people/day compared to lines with similar number of operations, and the utilization rate was also calculated to be the highest at 104 % compared to other regions. In addition, since a number of future urban development plans are being promoted to secure the number of residents and workers that affect the number of users, it is considered that the justification for expanding the number of train operations was secured. Through this study, it is expected that the number of reasonable operations will be secured by comprehensively judging various matters other than the number of users for simple profit.