• Journal of Korean Tunnelling and Underground Space Association
• /
• v.23 no.6
• /
• pp.549-558
• /
• 2021

In order to keep infrastructures such as tunnels and underground facilities safe, cracks of concrete lining in tunnel should be detected by regular inspections. Since regular inspections are accomplished through manual efforts using maintenance lift vehicles, it brings about traffic jam, exposes works to dangerous circumstances, and deteriorates consistency of crack inspection data. This study aims to provide methodology to automatically extract cracks from tunnel concrete lining images generated by the existing tunnel image acquisition system. Specifically, we train a deep learning based semantic segmentation model with open dataset, and evaluate its performance with the dataset from the existing tunnel image acquisition system. In particular, we compare the model performance in case of using all of a public dataset, subset of the public dataset which are related to tunnel surfaces, and the tunnel-related subset with negative examples. As a result, the model trained using the tunnel-related subset with negative examples reached the best performance. In the future, we expect that this research can be used for planning efficient model training strategy for crack detection.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.21 no.1
• /
• pp.201-210
• /
• 2019

Recently, due to the efforts to mitigate traffic congestion and expansion of space efficiency, the construction of underground roads has been increased in big-scale cities. Since tunnels in the city have a higher chance for a fire leading to a great tragedy during a severe traffic jam than mountain tunnels, it is highly likely that it will be constructed as a tunnel, having a small cross section, for small vehicles. However, if they are constructed as such small-vehicle tunnels, it would be possible to reduce the design fire intensity while the concentration of harmful gases would increase due to a reduction in the small cross sectional area, led by a decrease in the tunnel height. In this study, behaviors of fire smoke by the installation interval and format of large-scale exhaust-gas ports were examined and compared in the analysis of temperatures and CO concentrations of a tunnel and its results were as the following. Although there were no significant differences in the smoke spreading distance between installation intervals, but in this study, 100 m was found to be the most effective installation interval. The smoke exhaustion performance was found to be excellent in the order of $4m{\times}3m$, $6m{\times}2m$, and $3m{\times}2m$ (2 lane) of the smoke spreading distance. Although there was no significant difference in the smoke spreading distance between formats of large-scale exhaust-gas ports, it was found that the smoke spreading distance was larger than other cases when it was $3m{\times}2m$ in the fire growing process. The analysis of smoke spreading distances by the aspect ratio showed that a smoke spreading distance was shorted when its the smoke spreading distance was found to be shorter when its traverse distance was relatively longer than its longitudinal distance.