• Geomechanics and Engineering
• /
• v.37 no.6
• /
• pp.539-554
• /
• 2024

This study explores development of prediction model for seismic site classification through the integration of machine learning techniques with horizontal-to-vertical spectral ratio (HVSR) methodologies. To improve model accuracy, the research employs outlier detection methods and, synthetic minority over-sampling technique (SMOTE) for data balance, and evaluates using seven machine learning models using seismic data from KiK-net. Notably, light gradient boosting method (LGBM), gradient boosting, and decision tree models exhibit improved performance when coupled with SMOTE, while Multiple linear regression (MLR) and Support vector machine (SVM) models show reduced efficacy. Outlier detection techniques significantly enhance accuracy, particularly for LGBM, gradient boosting, and voting boosting. The ensemble of LGBM with the isolation forest and SMOTE achieves the highest accuracy of 0.91, with LGBM and local outlier factor yielding the highest F1-score of 0.79. Consistently outperforming other models, LGBM proves most efficient for seismic site classification when supported by appropriate preprocessing procedures. These findings show the significance of outlier detection and data balancing for precise seismic soil classification prediction, offering insights and highlighting the potential of machine learning in optimizing site classification accuracy.

• The Journal of the Acoustical Society of Korea
• /
• v.38 no.6
• /
• pp.687-696
• /
• 2019

Conventional feature of seismic classification focuses on strong seismic classification, while it is not suitable for classifying micro-seismic waves. We propose a feature extraction method based on histogram and Principal Component Analysis (PCA) in frequency-time space suitable for classifying seismic waves including strong, micro, and artificial seismic waves, as well as noise classification. The proposed method essentially employs histogram and PCA based features by concatenating the frequency and time information for binary classification which consist strong-micro-artificial/noise and micro/noise and micro/artificial seismic waves. Based on the recent earthquake data from 2017 to 2018, effectiveness of the proposed feature extraction method is demonstrated by comparing it with existing methods.

• 한국지구물리탐사학회:학술대회논문집
• /
• 2005.09a
• /
• pp.37-42
• /
• 2005

Remote seismic tomography is regarded as one of the most valuable geophysical technique for the estimation of the rock mass classification in the tunnel area where hard data information such as drill logs are absent. But the results of rock mass classification based on the remote seismic tomography tend to be overestimated in practice. In this study, we propose the effective method to implement the seismic reference velocity ratio based on semblance for the improvement of rock mass classification. Also, to verify its feasibility, proposed technique was tested by using the real field data.

• Nuclear Engineering and Technology
• /
• v.55 no.2
• /
• pp.792-799
• /
• 2023

Non-reactor nuclear facilities are increasing remarkably in Korea combined with advanced technologies such as life and space engineering, and the diversification of the nuclear industry. However, the absence of a basic classification guideline related to the design of non-reactor nuclear facilities has created confusion whenever related projects are carried out. In this paper, related domestic and international technical guidelines are reviewed to present the classification criteria of non-reactor nuclear facilities in Korea. Based on these criteria, the classification of structures, systems and components (SSCs) for safety controls is presented. Using the presented classification criteria, classification of a hot cell facility, a representative non-reactor nuclear facility, was performed. As a result of the classification, the hot cell facility is classified as the hazard category 3, accordingly, the safety class was classified as non-nuclear safety, the seismic category as non-seismic (RW-IIb), and the quality class as manufacturers' standards (S).

• Journal of the Korea Society of Computer and Information
• /
• v.24 no.1
• /
• pp.33-39
• /
• 2019

In this paper, we propose deep convolutional neural network(CNN) with bottleneck structure which improves the performance of earthquake classification. In order to address all possible forms of earthquakes including micro-earthquakes and artificial-earthquakes as well as large earthquakes, we need a representation and classifier that can effectively discriminate seismic waveforms in adverse conditions. In particular, to robustly classify seismic waveforms even in low snr, a deep CNN with 1x1 convolution bottleneck structure is proposed in raw seismic waveforms. The representative experimental results show that the proposed method is effective for noisy seismic waveforms and outperforms the previous state-of-the art methods on domestic earthquake database.

• Earthquakes and Structures
• /
• v.15 no.1
• /
• pp.1-8
• /
• 2018

A new site classification system and site coefficients based on local site conditions in Korea were developed and implemented as a part of minimum design load requirements for general seismic design. The new site classification system adopted bedrock depth and average shear wave velocity of soil above the bedrock as parameters for site classification. These code provisions were passed through a public hearing process before it was enacted. The public hearing process recommended to modify the naming of site classes and adjust the amplification factors so that the level of short-period amplification is suitable for economical seismic design. In this paper, the new code provisions were assessed using dynamic centrifuge tests and by comparing the design response spectra (DRS) with records from 2016 Gyeongju earthquake, the largest earthquake in history of instrumental seismic observation in Korea. The dynamic centrifuge tests were performed to simulate the representative Korean site conditions, such as shallow depth to bedrock and short-period amplification characteristics, and the results corroborated with the new DRS. The Gyeongju earthquake records also showed good agreement with the DRS. In summary, the new code provisions are reliable for representing the site amplification characteristic of shallow bedrock condition in Korea.

• Journal of the Earthquake Engineering Society of Korea
• /
• v.26 no.4
• /
• pp.173-180
• /
• 2022

Most of the drinking water dams managed by the local governments in Korea are earthfill dams, and these dams have almost no geotechnical property information necessary for seismic performance evaluation. Nevertheless, in the rough planning stage for improving seismic safety for these dams, it is necessary to classify their relative seismic hazard against earthquakes and conduct an additional ground investigation. The zero seismic failure probability curve is a curve suggested in this study in which the probability of failure due to an earthquake becomes '0' regardless of the geotechnical properties of the earthfill dam. By examining the method and procedure for calculating failure probability due to an earthquake suggested in previous researches, the zero seismic failure probability curves for an earthquake in 1,000-year and 2,400-year return periods in Korea were presented in the form of a hyperbola on the plane of the dam height versus freeboard ratio (ratio of freeboard to dam height), respectively. The distribution characteristics of the dam height and the freeboard ratio of 81 Korean earthfill dams were presented. The two proposed zero seismic failure probability curves are shown on the plane of the dam height versus freeboard ratio, and the relative seismic hazard of 81 dams can be classified into three groups using these curves as boundaries. This study presented the method of classifying the relative seismic hazard and the classification result.

• Proceedings of the Korean Geotechical Society Conference
• /
• 2006.03a
• /
• pp.562-571
• /
• 2006

Mean shear wave velocity of the upper 30m $(V_s30)$ used as the current site classification criterion for determining seismic design ground motions in Korea was established based on the typical depth of site investigations in western US, in which the depth to bedrock is much deeper than that in Korea. In this study, to establish appropriate site classification system for site conditions of Korea, site investigations including in-situ seismic tests to determine shear wave velocity $(V_s)$ were carried out at total 72 sites in Korean peninsula. The mean $V_s's$ to the depths of 5m, 10m, 15m, 20m and 25m together with the $V_s30$ at the testing sites were determined, and the correlation between the mean $V_s$ to a depth shallower than 30m and the $V_s30$ was drawn and suggested for the efficient seismic site classification in Korea. The proposed correlation could be utilized for the seismic design in case of the $V_s$ profiles shallower than 30 m in depth. The correlation in this study, nevertheless, requires further modification by means of the accumulation of various site data in Korea.

• Earthquakes and Structures
• /
• v.9 no.5
• /
• pp.983-997
• /
• 2015

This research evaluates the soil conditions for seismic stations situated in Romania using the horizontal-to-vertical spectral ratio (HVSR). The strong ground motion database assembled for this study consists of 179 analogue and digital strong ground motion recordings from four intermediate-depth Vrancea seismic events with $M_w{\geq}6.0$. In the first step of the analysis, the influence of the earthquake magnitude and source-to-site distance on the H/V curves is evaluated. Significant influences from both the earthquake magnitude and hypocentral distance are found especially for soil class A sites. Next, a site classification method proposed in the literature is applied for each seismic station and the soil classes are compared with those obtained from borehole data and from the topographic slope method. In addition, the success and error rates of this method are computed and compared with other studies from the literature. A more in-depth analysis of the H/V results is performed using data from seismic stations in Bucharest and a comparison of the free-field and borehole H/V curves is done for three seismic stations. The results show large differences between the free-field and the borehole curves. As a conclusion, the results from this study represent an intermediary step in the evaluation of the soil conditions for seismic stations in Romania and the need to perform more detailed soil classification analysis is highly emphasized.

• 한국지구물리탐사학회:학술대회논문집
• /
• 2009.05a
• /
• pp.35-45
• /
• 2009

Reservoir Characterization (RC) using 3-D seismic attributes analysis can provide properties of the oil sand reservoirs, beyond seismic resolution. For example, distributions and temporal bed thicknesses of reservoirs could be characterized by Spectral Decomposition (SD) and additional seismic attributes such as wavelet classification. To extract physical properties of the reservoirs, we applied 3-D seismic attributes analysis to the oil sand reservoirs in McMurray formation, in BlackGold Oilsands Lease, Alberta Canada. Because of high viscosity of the bitumen, Enhanced Oil Recovery (EOR) technology will be necessarily applied to produce the bitumen in a steam chamber generated by Steam Assisted Gravity Drainage (SAGD). To optimize the application of SAGD, it is critical to identify the distributions and thicknesses of the channel sand reservoirs and shale barriers in the promising areas. By 3-D seismic attributes analysis, we could understand the expected paleo-channel and characteristics of the reservoirs. However, further seismic analysis (e.g., elastic impedance inversion and AVO inversion) as well as geological interpretations are still required to improve the resolution and quality of RC.