• Tunnel and Underground Space
• /
• v.29 no.3
• /
• pp.184-196
• /
• 2019

This study aims to extract a 3D image of micro-cracks generated by hydraulic fracturing tests, using the deep learning method and X-ray computed tomography images. The pixel-level cracks are difficult to be detected via conventional image processing methods, such as global thresholding, canny edge detection, and the region growing method. Thus, the convolutional neural network-based encoder-decoder network is adapted to extract and analyze the micro-crack quantitatively. The number of training data can be acquired by dividing, rotating, and flipping images and the optimum combination for the image augmentation method is verified. Application of the optimal image augmentation method shows enhanced performance for not only the validation dataset but also the test dataset. In addition, the influence of the original number of training data to the performance of the deep learning-based neural network is confirmed, and it leads to succeed the pixel-level crack detection.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.36 no.5
• /
• pp.339-346
• /
• 2023

The uncertainties of microstructural features affect the properties of materials. Numerous pores that are randomly distributed in materials make it difficult to predict the properties of the materials. The distribution of pores in cementitious materials has a great influence on their mechanical properties. Existing studies focus on analyzing the statistical relationship between pore distribution and material responses, and the correlation between them is not yet fully determined. In this study, the mechanical response of cementitious materials is predicted through an image-based data approach using a convolutional neural network (CNN), and the correlation between pore distribution and material response is analyzed. The dataset for machine learning consists of high-resolution micro-CT images and the properties (tensile strength) of cementitious materials. The microstructures are characterized, and the mechanical properties are evaluated through 2D direct tension simulations using the phase-field fracture model. The attributes of input images are analyzed to identify the spot with the greatest influence on the prediction of material response through CNN. The correlation between pore distribution characteristics and material response is analyzed by comparing the active regions during the CNN process and the pore distribution.

• Journal of the Korean Geotechnical Society
• /
• v.32 no.1
• /
• pp.55-62
• /
• 2016

This study presents the numerical results of soil-water characteristic curve for sandy soil by pore network model. The Jumunjin sand is subjected to the high resolution 3D X-ray computed tomographic imaging and its pore structure is constructed by the web of pore body and pore channel. The channel radius, essential to the computation of capillary pressure, is obtained based on the skeletonization and Euclidean Distance transform. The experimentally obtained soil-water characteristic curve corroborates the numerically estimated one. The pore channel radius defined by minimum radii of pore throat results in the slightly overestimation of air entry value, while the overall evolution of capillary pressure resides in the acceptable range. The relative permeability computed by a series of suggested models runs above that obtained by pore network model at high degree of saturation.

• Journal of the Korean Geotechnical Society
• /
• v.32 no.6
• /
• pp.41-48
• /
• 2016

This study presents the correlations between quantified particle shape parameters and internal friction angles for nine sand specimens including six natural sands and three crushed sands. Specimens are subjected to 3D X-ray computed tomographic imaging and their particles are segmented through the aid of image processing techniques. Shapes of segmented particles are then quantified through two shape parameters such as sphericity and elongation. The direct shear apparatus enables us to measure peak and critical state friction angles of sand specimens of distinct relative densities. The gathered data show that decreasing sphericity and increasing elongation cause increases in peak and critical state friction angle with similar gradients.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.23 no.7 s.166
• /
• pp.1112-1119
• /
• 1999

Fracture resistance(J-R) curves, which are used for the elastic-plastic fracture mechanics analyses, are known to be dependent on the cyclic loading history. The objective of this paper is to study the effect of reverse cyclic loading on J-R curves in CT specimens. The effect of two parameters was observed on the J-R curves during the reverse cyclic loading. One was the minimum-to-maximum load ratio(R) and the other was the incremental plastic displacement(${\delta}_{cycle}/{\delta}_i$), which is related to the amount of crack growth that occurs in a cycle. Fracture resistance test on CT specimens with varying load ratio and incremental plastic displacement were performed. For the SA 516 Gr. 70 steel, the results showed that the J-R curves were decreased with decreasing the load ratio and the incremental plastic displacement. When the load ratio was set to -1, the results of the J-R curves and the $J_i$ value were about $40{\sim}50$ percent of those for the monotonic loading condition. Also on condition that the incremental plastic displacement reached 1/40, the J-R curves and the $J_i$ value were about $50{\sim}60$ percent of those for the incremental plastic displacement of 1/10.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.10 no.1
• /
• pp.25-33
• /
• 1986

In this paper, the frontal method based elastic-plastic F.E.M. program for mini-computer was developed. Since, the executable source program size was restricted by the system core memory size on the mini-computer, the active variables were memorized by the element base and the nonactive varables were memorized to the external disc file. The active variables of the finally developed program were reduced enough to execute about 1,000 freedom finite element on the mini-computer on which available variables were restricted as 32,767 integers. A modified CT fracture test specimen was examined to test the developed program. The calculated results were compared with experimental results concerning on the crack tip plastic deformation zone. Recrystallization technique was adopted to visualize the intensive plastic deformation regions. The Von-Mises criterion based calculation results were well agreed with the experimental results in the intensive plastic region which was over than 2% offset strain. The F.E.M. results using the developed program were well agreed with the theoritical plastic boundary which was calculated by the stress intensity factor as r$_{p}$=(K$_{1}$$^{2}$/2.pi..sigma.$_{y}$$^{2}$).f(.theta.).).).

• Journal of the Korean Society of Propulsion Engineers
• /
• v.20 no.2
• /
• pp.11-17
• /
• 2016

Powder metallurgy nickel based superalloy has been used in a high temperature part of turbine engine for airplane. The fatigue crack growth behavior was investigated using CT specimens for the materials at room temperature(R.T.), $600^{\circ}C$ and $700^{\circ}C$. The direct current potential drop(DCPD) method suggested by ASTM E647 was used to measure the crack length during fatigue crack growth at various stress ratios. The fatigue crack growth rate at R=0.5 was faster than that at R=0.1 for all temperature conditions and increased with the increase of stress ratio and temperature. Fractography was conducted for analysis of fracture mechanism.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.36 no.12
• /
• pp.1599-1604
• /
• 2012

In general, internal defects, such as shrinkage in casting, cause stress concentration and can be a starting point for cracks. Therefore, it is important to understand the effects of internal defects on the mechanical properties including the impact behavior. This study aim is to evaluate the effects of internal casting defects on the impact performance of Al-alloy castings. Both an experimental method and computational analysis were used to achieve the research objective. The internal defects in the casting were scanned using an industrial CT scanner, and their shape was simplified using ellipsoidal primitives for impact analysis. The good agreement between the experimental and computer simulation results verified the reliability of the proposed computational method for the FEA of casting components with internal defects.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.35 no.10
• /
• pp.1243-1248
• /
• 2011

Internal defects are a major concern in the casting process because they have a significant influence on the strength and fatigue life of casting products. In general, they cause stress concentration and can be a starting point of cracks. Therefore, it is important to understand the effects of internal defects on mechanical properties such as fatigue life. In this study, fatigue experiments on tensile specimens with internal defects were performed. The internal defects in the casting product were scanned by an industrial CT scanner, and its shape was simplified by ellipsoidal primitives for the structural and fatigue analysis. The analysis results were compared with experimental results for casting products with internal defects. It was demonstrated that it is possible to consider internal defects of casting products in stress and fatigue analysis. The proposed method provides a tool for the prediction of the fatigue life of casting products and the investigation of the effects of internal defects on mechanical performance.

• Nuclear Engineering and Technology
• /
• v.21 no.3
• /
• pp.165-170
• /
• 1989

The strain rate effects on fracture toughness and fracture resistance characteristics of A533B-1 nuclear pressure vessel steels were examined in the quasi-dynamic test conditions through the microscopic investigation of the intense strain region around crack tip and the microroughness of fracture surface. J-value calculated from the recrystallization etch technique was the same as calculated from the modified-J when the crack extension is less than 1.5mm in a 1/2T-CT specimen. Local fracture strain was calculated from the fracture surface micro-roughness. The local strains were calculated to be the values of 1.8 and 2.0 and were much higher than the macroscopically measured values. It was nearly independent on strain rate and was regarded as a material constant in ductile dimpled rupture. The fracture toughness increased with increase in strain rate while the tearing modulus showed little variation.