간행물
Journal of the Earthquake Engineering Society of Korea KCI 등재 한국지진공학회논문집
- 발행기관 한국지진공학회
- 자료유형 학술지
- 간기 격월간
- ISSN 1226-525x (Print)2234-1099 (Online)
- 수록기간 1997 ~ 2026
- 주제분류 공학 > 토목공학 공학 분류의 다른 간행물
- 십진분류KDC 531DDC 624
권호리스트/논문검색
제30권 제3호(통권 제171호) (2026년 5월) 4건
1.
2026.05
구독 인증기관 무료, 개인회원 유료
Rapid, real-time detection of anomalies and locate structural defects during earthquakes is critical for ensuring safety and enabling timely decision-making. Although deep learning-based structural health monitoring (SHM) has shown considerable promise, conventional supervised models are often impractical because labeled damage data from real-world structures are extremely scarce. To address this challenge, this paper proposes a Multi-Class Deep Support Vector Data Description (SVDD) framework for structural defect detection. The proposed Multi-Class Deep SVDD approach learns the boundary of normal data using only normal seismic acceleration responses. When new data are recorded, the system infers both the occurrence and location of defects by evaluating whether the responses fall within or deviate from the learned normal boundary. The framework is validated using the Los Alamos National Laboratory 3-story bookshelf structure benchmark dataset. Experimental results show that the proposed model achieves a peak average accuracy of 87.12% in a 4-dimensional latent space, substantially outperforming traditional baseline methods, including Kernel Density Estimation (KDE), SVDD, and One-Class Deep SVDD. These findings indicate that the Multi-Class Deep SVDD framework provides a robust and objective metric for rapid post-earthquake safety assessment without requiring prior exposure to faulty datasets.
4,000원
2.
2026.05
구독 인증기관 무료, 개인회원 유료
This study optimizes three machine learning models—Decision Tree, Random Forest (RF), and Gradient Boosting—to classify concrete structure types (C2, C3, C4, and C5) using information from a building register. Although the initial models achieved high overall accuracy, the minority class C5 exhibited relatively low performance due to class imbalance and inherent complexity. To address this, an exhaustive grid search over discrete parameter candidates was performed, and a class-weighting strategy was integrated into the RF model to prioritize accurate classification of the minority class. The optimized RF model preserved a high overall accuracy of 94% while markedly improving C5 recall from 0.81 to 0.86 and its F1-score from 0.85 to 0.87. These results demonstrate that strategic hyperparameter tuning with class weights can effectively enhance classification reliability for rare structural types. Future research should include feature importance analysis to refine data configurations and the expansion of minority class samples to further improve model robustness in practical applications.
4,200원
3.
2026.05
구독 인증기관 무료, 개인회원 유료
This study evaluates ground-motion (GM) acceleration conversion methods by applying them to strain-rate data from a horizontal Distributed Acoustic Sensing (DAS) array under both idealized and real-world conditions. We test four conversion methods—1) slant-stacking, 2) Lior’s method, 3) Lindsey’s method, and 4) Curvelet transform—through numerical modeling and by applying them to a publicly available 9 km horizontal DAS array dataset. Numerical simulations reveal critical calculation factors specific to each method and show that numerically derived apparent ground velocity can deviate from theoretical values when multiple elastic waves arrive simultaneously. In real-world applications, the slant-stacking and Lior’s methods are relatively insensitive to the measurement length of the straight DAS array. By contrast, the Curvelet method exhibits strong sensitivity to this factor, whereas Lindsey’s method shows weaker dependence. Implementing Lior’s method in the frequency-wavenumber domain also requires pre-determining water-levels by comparing adjacent seismograms. Additionally, we find that Lior’s method generates excessively high spectral levels above 13 Hz, which may lead to underestimation of the high-frequency spectral attenuation parameter (κ0), a key parameter in GM simulation. Collectively, these findings provide a technical guideline for the use of horizontal DAS arrays in future observational earthquake seismology.
4,300원
4.
2026.05
구독 인증기관 무료, 개인회원 유료
This study quantitatively evaluates the effects of embankment height and input excitation frequency on crest settlement—a key damage indicator—for railway embankments founded on liquefiable ground. Dynamic numerical analyses were conducted using FLAC2D, based on the cross-section adopted in a previous 1-g shaking table test. The parametric study considered four embankment heights (0, 2, 4, and 6 m) and three input frequencies (0.8, 2.5, and 5.0 Hz). To simulate liquefaction in the foundation soil, the PM4Sand constitutive model was employed within an effective-stress framework. Model validity was first examined by comparing computed time histories of excess pore-water pressure, acceleration, and settlement with experimental results, and by confirming qualitative agreement with observed settlement trends across different embankment heights. The results show that crest settlement does not increase monotonically with embankment height; instead, it reaches a maximum and then decreases beyond a critical range. The largest settlement occurs when the embankment height is approximately 15~25% of the liquefiable layer thickness. This behavior reflects the competition between increased overburden pressure, which enhances liquefaction resistance beneath the embankment, and amplified lateral spreading, which increases permanent deformation. Although excitation frequency influences settlement, its effect is smaller than that of embankment height.
4,000원
