• Geophysics and Geophysical Exploration
• /
• v.26 no.4
• /
• pp.171-184
• /
• 2023

Machine learning (ML)-based cavity detection using a large amount of survey data obtained from vehicle-mounted ground penetrating radar (GPR) has been actively studied to identify underground cavities. However, only simple image processing techniques have been used for preprocessing the ML input, and many conventional seismic and GPR data processing techniques, which have been used for decades, have not been fully exploited. In this study, based on the idea that a cavity can be identified using diffraction, we applied ML-based diffraction separation to GPR data to increase the accuracy of cavity detection using the YOLO v5 model. The original ML-based seismic diffraction separation technique was modified, and the separated diffraction image was used as the input to train the cavity detection model. The performance of the proposed method was verified using public GPR data released by the Seoul Metropolitan Government. Underground cavities and objects were more accurately detected using separated diffraction images. In the future, the proposed method can be useful in various fields in which GPR surveys are used.

• Tunnel and Underground Space
• /
• v.28 no.4
• /
• pp.358-371
• /
• 2018

The stability of openings in the underground mine is major concern in the operation of mines that must ensure productivity and safety. Among many rock conditions affecting cavities stability, the width and height of the opening is an important design factor. In this paper, we consider to determine the maximum unsupported span of a opening in a limestone mine by using the Q system among several rock classification schemes. In order to determine the span of the unsupported opening in the limestone mine, rock mass classifications were carried out at over 200 sites in the underground limestone mines. The relationships by using the Q system and the stability graph proposed by Mathews to determine the maximum span of the unsupported opening were derived and compared. We propose a new classification method that combines GSI and RMR rock classification systems to make it easy to use in a field.

• Tunnel and Underground Space
• /
• v.30 no.2
• /
• pp.149-163
• /
• 2020

This paper presents a logistic regression and GIS based urban ground sink susceptibility assessment using underground facility information considering soil particle loss. In the underground environment, the particle loss due to water flow or groundwater level change leads to the occurrence and expansion of cavities, which directly affect the ground sink. Four different contributory factors were selected according to the two underground facility domains (water pipeline area, sewer pipeline area) and subway line area. The logistic regression method was used to analyze the correlation and to derive the regression equation between the ground sink inventory and the contributory factors. Based on these results, three ground sink susceptibility maps were generated. The results obtained from this study are expected to provide basic data on the area susceptible to ground sink and needed to safety monitoring.

• Tunnel and Underground Space
• /
• v.10 no.3
• /
• pp.395-402
• /
• 2000

The influence of the mined-out caves on the stability of the high speed railway tunnel was investigated with a series of geological logging and in-situ tests on the one hand, and with the rock mass classification using the multiple regression analysis on the other hand. The rock mass in this area can be classified as 'fair', and the condition of the discontinuities plays the most important role in the classification of the rock mass. The results of the analysis obtained by the FLAC showed that the western part of the tunnel locating at 50m above the mine cavities could be affected by subsidence associated with a considerable deformation, the magnitude of which might depend on the properties of the rock mass.

• Journal of the Korean Geotechnical Society
• /
• v.35 no.11
• /
• pp.37-49
• /
• 2019

In order to analyze the effect of burial depth on the generation of ground cavities and cave-ins, a series of model experiments with different height of model ground were performed. Digital images of the model ground were captured to evaluate the internal deformation of the model grounds by adopting the PIV (Particle Image Velocimetry) technique. Additionally, the vertical displacement at the surface, the size of the cavity, and the weight of the discharged soil were measured in each test. The results indicate that the model ground with low burial depth, which does not satisfy the criterion, was more vulnerable to ground cavities and cave-ins than the model ground with high burial depth.

• Journal of the Korean GEO-environmental Society
• /
• v.21 no.3
• /
• pp.5-11
• /
• 2020

Recently, ground pits that have been occurring frequently in urban areas are hindering traffic flow and causing property damages and loss of human life, acting as factors that are threatening the safety of citizens. Therefore, sunken ground must be quickly restored and provisions must be made for additional damage but current domestic detailed standards regarding ground pits and accurate definitions regarding causes and measures to be taken for reoccurrences are lacking. Restoration methods of sunken ground include backfilling by reusing sunken soil or other fill material and paving the road and while this is the most often used method, this only prevents ground from sinking temporarily and can not serve as a fundamental solution. Also, additional ground pits can occur on ground that is reinforced using this method due to faulty backfill material or faulty hardening. This study used Eco-friendly High-Strength Material (EHSM) as restoration material that can be used in the restoration of underground cavities that have occurred due to ground subsidence to analyze the engineered characteristics of modified dredging clay and test pieces made from changed ratios of EHSM and weathered granite soil were uniaxial compression tests were conducted and freezing-thawing tests were conducted to study strength properties according to environmental changes of restoration material, and after tests were concluded by each level, uniaxial compression tests and dynamic elasticity tests were conducted for intensity analysis. Also, to evaluate strength characteristics of the restored ground, dynamic plate load tests were conducted to verify the improvement effectiveness of the restored ground.

• Journal of the Korean GEO-environmental Society
• /
• v.25 no.1
• /
• pp.5-11
• /
• 2024

Ground subsidence mainly occurs in urban areas with high population density, so it is necessary to clearly identify the cause of occurrence and prepare in advance. The main cause of ground subsidence is reported to be the creation of cavities in the ground due to damage to underground pipes, but the property information and influencing factors of underground pipes to predict and prepare for ground subsidence are not properly established. Therefore, in this study, factors showing a significant correlation with the occurrence of ground subsidence were selected among the underground facility property information and a regression equation was proposed through logistic regression analysis. For this purpose, data on underground structures and ground subsidence history information in the target area were collected, and the target area was divided into girds of 100m x 100m in size using QGIS. The underground facility attribute information and ground subsidence history information contained within the gird were extracted. Then, preprocessing was performed to construct a dataset and correlation analysis was performed. As a result, factors excluding the year of sewer pipes and communication pipes and the average depth of communication pipes, heat pipes, and gas pipes were found to have a significant correlation with ground subsidence. In addition, a regression equation for whether ground subsidence occurred in the target area is proposed through logistic regression analysis.

• Journal of the Korean GEO-environmental Society
• /
• v.23 no.8
• /
• pp.5-10
• /
• 2022

Ground Subsidence has been continuously occurring in densely populated downtown. The main cause of ground subsidence is the damaged underground facility like sewer. Currently, ground subsidence is being dealt with by discovering cavities in ground using GPR. However, this consumes large amount of manpower and cost, so it is necessary to predict hazardous area for efficient operation of GPR. In this study, ◯◯city is divided into 500 m×500 m grids. Then, data set was constructed using the characteristics of the underground facility and ground subsidence in grids. Data set used to machine learning model for ground subsidence risk grade prediction. The purposed model would be used to present a ground subsidence risk map of target area.

• Geomechanics and Engineering
• /
• v.16 no.2
• /
• pp.177-185
• /
• 2018

The subsidence mechanism of ground surface is a complex phenomenon when multiple seam coal mining operations are carried out. Particularly, the coal mining beneath karstic formations causes a very special form of subsidence. The subsidence causes elasto-plastic deformation of the karstic layers and the collapse of cavities leads to dolinization and/or sinkhole formation. In this study, a sinkhole with a depth of 90 m and a width of 25 m formed in Gelik district within the coal-basin of Zonguldak (NW, Turkey) induced by multiple seam coal mining operations in the past has been presented as a case-history together with two-dimensional numerical simulations and InSAR monitoring. The computational results proved that the sinkhole was formed as a result of severe yielding in the close vicinity of the faults in contact with karstic formation due to multiple seam longwall mining at different levels.

• Journal of Korean Society of Disaster and Security
• /
• v.12 no.2
• /
• pp.57-63
• /
• 2019

The road sinks in the sewer line or subway section are affected by the ground characteristics. Therefore, it is necessary to accurately identify the relationship between the soil properties and the ground motion in the area where cavities occurred in order to establish a countermeasure against the road sink. In this paper, simulation was performed by using EDEM program, which is one of the Discrete Element Method programs, for sandy soil and clayey soil, which are most common in alluvial deposits, with different locations and sizes of cavities in the underground. As a result, it was found that the sink size occurred more in the sandy soil than in the cohesive soil. Deeper and larger cavity is more likely to occur the road sink In the sand soil model while road sink in the clay model is easy to occur when the cavity is more shallower.