• Journal of Korean Tunnelling and Underground Space Association
• /
• v.21 no.3
• /
• pp.419-432
• /
• 2019

Most of deep learning model training was proceeded by supervised learning, which is to train labeling data composed by inputs and corresponding outputs. Labeling data was directly generated manually, so labeling accuracy of data is relatively high. However, it requires heavy efforts in securing data because of cost and time. Additionally, the main goal of supervised learning is to improve detection performance for 'True Positive' data but not to reduce occurrence of 'False Positive' data. In this paper, the occurrence of unpredictable 'False Positive' appears by trained modes with labeling data and 'True Positive' data in monitoring of deep learning-based CCTV accident detection system, which is under operation at a tunnel monitoring center. Those types of 'False Positive' to 'fire' or 'person' objects were frequently taking place for lights of working vehicle, reflecting sunlight at tunnel entrance, long black feature which occurs to the part of lane or car, etc. To solve this problem, a deep learning model was developed by simultaneously training the 'False Positive' data generated in the field and the labeling data. As a result, in comparison with the model that was trained only by the existing labeling data, the re-inference performance with respect to the labeling data was improved. In addition, re-inference of the 'False Positive' data shows that the number of 'False Positive' for the persons were more reduced in case of training model including many 'False Positive' data. By training of the 'False Positive' data, the capability of field application of the deep learning model was improved automatically.

• Journal of the Korean Crystal Growth and Crystal Technology
• /
• v.29 no.2
• /
• pp.61-70
• /
• 2019

Geopolymer is an eco-friendly construction material that has various advantages such as reduced $CO_2$ emission, fire resistance and low thermal conductivity compared to cement. However, it has not been many studies on the thermal behavior of the surface of the geopolymer panel when flame is applied to the surface. In this study, surface characteristics of hardened geopolymer on flame exposure was investigated to observe its characteristics as heat-resistant architectural materials. External structure changes and crack due to the heat shock were not observed during the exposure on flame. According to the residue of calcite and halo pattern of aluminosilicate gel, decarboxylation and dehydration were extremely limited to the surface and, therefore, it is thought that durability of hardened geopolymer was sustained. Gehlenite and calcium silicate portion was inversely proportional to quartz and calcite and significantly directly proportional to BFS replacement ratio. Microstructure changes due to the thermal shock caused decarboxylation and dehydration of crystallization and it was developed the pore and new crystalline phase like calcium silicate and gehlenite. It is thought that those crystalline phase worked as a densification and strengthening mechanism on geopolymer panel surface.

• Korean Chemical Engineering Research
• /
• v.59 no.4
• /
• pp.521-531
• /
• 2021

Vacuum Pressure Impregnation (VPI) is a process that enhances physical properties by coating some types of epoxy resins on windings of stator used in large rotators such as generators and motors. During vacuum and pressurization of the VPI process, resin gas is generated by vaporization of epoxy resin. When the tank is opened for curing after finishing impregnation, resin gas is leaked out of the tank. If the leaked resin gas spreads throughout the workplace, there are safety and environmental problems such as fire, explosion and respiratory problems. So, exhaust system for resin gas is required during the process. In this study, a case study of exhaust efficiency by location of vent was conducted using Computational Fluid Dynamics (CFD) in order to design a system for exhausting resin gas generated by the VPI process. The optimal exhaust system of this study allowed more than 90% of resin gas to be exhausted within 1,800 seconds and reduced the fraction of resin gas below the Low Explosive Limit (LEL).

• Journal of Korea Water Resources Association
• /
• v.54 no.5
• /
• pp.301-309
• /
• 2021

In this study, we tried to improve the performance of the existing U-net-based deep learning rainfall prediction model, which can weaken the meaning of time series order. For this, ConvLSTM2D U-Net structure model considering temporal consistency of data was applied, and we evaluated accuracy of the ConvLSTM2D U-Net model using a RainNet model and an extrapolation-based advection model. In addition, we tried to improve the uncertainty in the model training process by performing learning not only with a single model but also with 10 ensemble models. The trained neural network rainfall prediction model was optimized to generate 10-minute advance prediction data using four consecutive data of the past 30 minutes from the present. The results of deep learning rainfall prediction models are difficult to identify schematically distinct differences, but with ConvLSTM2D U-Net, the magnitude of the prediction error is the smallest and the location of rainfall is relatively accurate. In particular, the ensemble ConvLSTM2D U-Net showed high CSI, low MAE, and a narrow error range, and predicted rainfall more accurately and stable prediction performance than other models. However, the prediction performance for a specific point was very low compared to the prediction performance for the entire area, and the deep learning rainfall prediction model also had limitations. Through this study, it was confirmed that the ConvLSTM2D U-Net neural network structure to account for the change of time could increase the prediction accuracy, but there is still a limitation of the convolution deep neural network model due to spatial smoothing in the strong rainfall region or detailed rainfall prediction.

• Journal of the Korean Recycled Construction Resources Institute
• /
• v.9 no.1
• /
• pp.33-40
• /
• 2021

The objective of this work is to prove a possibility of void f illing through a carbonation f or the purpose of improving the quality of recycled aggregate. Carbonation can permanently immobilize CO2, which is a greenhouse gas, and thus provides additional benefit on environment. In this work, recycled fine aggregate was reacted using gaseous CO2 and supercritical CO2(scCO2) in a closed chamber, and the changes in physical properties of the recycled f ine aggregate bef ore and af ter carbonation were analyzed using the apparent density, skeletal density, pH, and FE-SEM measurements. Thereafter, a mortar specimen was prepared and a compressive strength was measured. According to the experimental results, it was found that the increase in the apparent density and the true density was higher by the reaction with scCO2, which was conducted at high temperature and high pressure compared to the reaction with gaseous CO2. In addition, the pH of the eluted water was found to have a larger initial decrease than that observed with samples from reaction by gaseous CO2. The shape and amount of calcium carbonate crystals were also found to be larger than that from gaseous CO2. The increase in compressive strength was the largest when using recycled fine aggregate reacted with scCO2. It was clear that quality improvement of recycled fine aggregate was higher with scCO2 than with gaseous CO2.

• Journal of the Korean earth science society
• /
• v.42 no.6
• /
• pp.623-631
• /
• 2021

The increased transport of marine hazardous and noxious substances (HNS) has resulted in frequent HNS spill accidents domestically and internationally. There are about 6,000 species of HNS internationally, and most of them have toxic properties. When an accidental HNS spill occurs, it can destroys the marine ecosystem and can damage life and property due to explosion and fire. Constructing a spectral library of HNS according to wavelength and developing a detection algorithm would help prepare for accidents. In this study, a ground HNS spill experiment was conducted in France. The toluene spectrum was determined through hyperspectral sensor measurements. HNS present in the hyperspectral images were detected by applying the spectral mixture algorithm. Preprocessing principal component analysis (PCA) removed noise and performed dimensional compression. The endmember spectra of toluene and seawater were extracted through the N-FINDR technique. By calculating the abundance fraction of toluene and seawater based on the spectrum, the detection accuracy of HNS in all pixels was presented as a probability. The probability was compared with radiance images at a wavelength of 418.15 nm to select abundance fractions with maximum detection accuracy. The accuracy exceeded 99% at a ratio of approximately 42%. Response to marine spills of HNS are presently impeded by the restricted access to the site because of high risk of exposure to toxic compounds. The present experimental and detection results could help estimate the area of contamination with HNS based on hyperspectral remote sensing.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.24 no.4
• /
• pp.305-316
• /
• 2022

Hydrogen energy is emerging as an alternative to the depletion of fossil fuels and environmental problems, and the use of hydrogen vehicles is increasing in the automobile industry as well. However, since hydrogen has a wide flammability limit of 4 to 75%, there is a high concern about safety in case of a hydrogen car accident. In particular, in semi-enclosed spaces such as tunnels and underground parking lots, a fire or explosion accompanied by hydrogen leakage is highly likely to cause a major accident. Therefore, it is necessary to review hydrogen safety through analysis of flammability areas caused by hydrogen leakage. Therefore, in this study, the effect of the air velocity in the tunnel on the flammability area was investigated by analyzing the hydrogen concentration according to the hydrogen leakage conditions of hydrogen vehicles and the air velocity in the tunnel in a road tunnel with standard section. Hydrogen leakage conditions were set as one tank leaking and three tanks leaking through the TPRD at the same time and a condition in which a large crack occurred and leaked. And the air velocity in the tunnel were considered 0, 1, 2.5, and 4.0 m/s. As a result of the analysis of the flammability area, it is shown that when the air velocity of 1 m/s or more exists, it is reduced by up to 25% compared to the case of air velocity of 0 m/s. But there is little effect of reducing the flammability area according to the increase of the wind speed. In particular, when a large crack occurs and completely leaks in about 2.5 seconds, the flammability area slightly increases as the air velocity increases. It was found that in the case of downward ejection, hydrogen gas remains under the vehicle for a considerably long time.

• Applied Chemistry for Engineering
• /
• v.33 no.3
• /
• pp.302-308
• /
• 2022

The PdO nanoparticle with large surface area was selected to solve the environmental pollution problem at fire range caused by high energy explosives research department explosive (RDX) and high melting explosive (HMX). By simulating water pollution, RDX and HMX nitramine explosives were dissolved in water, followed by the degradation reaction at 313 K by adding PdO. In order to measure the degradation reaction rate of explosives, ¹H NMR was used, which can monitor the reaction rate without losing sample during reaction, and observe the progress of the reaction through the spectrum. The results showed that the degradation of RDX and HMX by PdO nanoparticles are pseudo-first order reaction. The degradation of explosives compounds were observed via the chemical shift and peak intensity analysis of NMR peaks. The measured rate constants for these reactions of RDX and HMX were 2.10 × 10^-2 and 6.35 × 10^-4 h^-1, respectively. This study showed that the application of PdO nanoparticles for explosives degradation is a feasible option.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.28 no.spc
• /
• pp.1-10
• /
• 2022

As the volume of marine hazardous and noxious substances (HNSs) transported in domestic and overseas seas increases, the risk of HNS spill accidents is gradually increasing. HNS leaked into the sea causes destruction of marine ecosystems, pollution of the marine environment, and human casualties. Secondary accidents accompanied by fire and explosion are possible. Therefore, various types of HNSs must be rapidly detected, and a control strategy suitable for the characteristics of each substance must be established. In this study, the ground HNS spill experiment process and application result of detection algorithms were presented based on hyperspectral remote sensing. For this, styrene was spilled in an outdoor pool in Brest, France, and simultaneous observation was performed through a hyperspectral sensor. Pure styrene and seawater spectra were extracted by applying principal component analysis (PCA) and the N-Findr method. In addition, pixels in hyperspectral image were classified with styrene and seawater by applying spectral matching techniques such as spectral distance similarity (SDS), spectral correlation similarity (SCS), spectral similarity value (SSV), and spectral angle mapper (SAM). As a result, the SDS and SSV techniques showed good styrene detection results, and the total extent of styrene was estimated to be approximately 1.03 m². The study is expected to play a major role in marine HNS monitoring.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.29 no.1
• /
• pp.76-85
• /
• 2023

Demand for hydrogen as a renewable energy resource is increasing. However, unlike conventional fossil fuels, hydrogen requires a dedicated refueling station for fuel supply. A risk assessment of hydrogen refueling stations must be undertaken to secure the infrastructure. Therefore, in this study, a risk assessment for hydrogen refueling stations was conducted through both qualitative and quantitative risk assessments. For the qualitative evaluation, the hydrogen dispenser module was evaluated as two nodes using the hazard and operability (HAZOP) analysis. The risk due to filter clogging and high-pressure accidents was evaluated to be high according to the criticality estimation matrix. For the quantitative risk assessment, the Hydrogen Korea Risk Assessment Module (Hy-KoRAM) was used to indicate the shape of the fire and the range of damage impact, and to evaluate the individual and social risks. The individual risk level was determined of to be as low as reasonably practicable (ALARP). Additional safety measures proposed include placing the hydrogen refueling station about 100m away from public facilities. The social risk level was derived as 1E-04/year, with a frequency of approximately 10 deaths, falling within the ALARP range. As a result of the qualitative and quantitative risk assessments, additional safety measures for the process and a safety improvement plan are proposed through the establishment of a restricted area near the hydrogen refueling station.