• 한국정밀공학회지
• /
• 제20권2호
• /
• pp.146-153
• /
• 2003

Ultrasonic inspection of defects has been focused on the existence of defect in structural material and need has much time and expenses in inspecting all the coordinates (x, y) on material surface. Neural networks can have an application to coordinates (x, y) of defects by multi-point inspection method. Ultrasonic inspection modeling is optimized by neural networks Neural networks has trained training example of absolute and relative coordinate of defects, and defect pattern. This method can predict coordinates (x, y) of defects within engineering estimated mean error $\psi$.

• 상하수도학회지
• /
• 제27권6호
• /
• pp.731-741
• /
• 2013

CCTV inspection method has been used in Korea for more than 20 years, but there is no proper assessment system for sewer failure severity that considers the domestic circumstances. This study classified the defects caused by the overburden load of reinforced concrete sewer pipes depending on severity and developed defect code by analyzing the domestic CCTV inspection videos. The defect score was assigned to each defect code, and it was classified into 5 grades for the decision-making of repair and rehabilitation. The result of this study is expected to be useful for domestic CCTV inspectors to assess the sewer condition and helpful for managers to make a decision of repair and rehabilitation.

• Advances in nano research
• /
• 제16권1호
• /
• pp.91-109
• /
• 2024

Carbon nanotubes are drawing wide attention of research communities and several industries due to their versatile capabilities covering mechanical and other multi-physical properties. However, owing to extreme operating conditions of the synthesis process of these nanostructures, they are often imposed with certain inevitable structural deformities such as single vacancy and nanopore defects. These random irregularities limit the intended functionalities of carbon nanotubes severely. In this article, we investigate the mechanical behaviour of double-wall carbon nanotubes (DWCNT) under the influence of arbitrarily distributed single vacancy and nanopore defects in the outer wall, inner wall, and both the walls. Large-scale molecular simulations reveal that the nanopore defects have more detrimental effects on the mechanical behaviour of DWCNTs, while the defects in the inner wall of DWCNTs make the nanostructures more vulnerable to withstand high longitudinal deformation. From a different perspective, to exploit the mechanics of damage for achieving defect-induced shape modulation and region-wise deformation control, we have further explored the localized longitudinal and transverse spatial effects of DWCNT by designing the defects for their regional distribution. The comprehensive numerical results of the present study would lead to the characterization of the critical mechanical properties of DWCNTs under the presence of inevitable intrinsic defects along with the aspect of defect-induced spatial modulation of shapes for prospective applications in a range of nanoelectromechanical systems and devices.

• 대한기계학회논문집A
• /
• 제30권8호
• /
• pp.931-940
• /
• 2006

Defect assessment of a weld zone is important in fitness-for-service evaluation of plant components. Typically a J and $C^*$ estimation method for a defective homogeneous component is extended to a mismatched component, by incorporating the effect due to the strength mismatch between the weld metal and the base material. The key element is a mismatch limit load. For instance, the R6/R5 procedure employs an equivalent material concept, defined by a mismatch limit load. A premise is that if a proper mismatch limit load solution is available, the same concept can be used for any defect location (either a weld centre defect or a heat affected zone (HAZ) defect) and for any material combination (either two-material or multi-material combinations; either similar or dissimilar joints). However, validation is still limited, and thus a more systematic investigation is needed to generalise the suggestion to any geometry, any defect location and any material combination. This paper describes the effect of structural geometry on the $C^*$ integral for defective similar welds, based on systematic elastic-creep 2-D and 3-D finite element (FE) analyses, to attempt to elucidate the questions given above. It is found that the existing 'equivalent material' concept is valid only for limited cases, although it provides conservative estimates of $C^*$ for most of cases. A modification to the existing equivalent material concept is suggested to improve accuracy.

• Smart Structures and Systems
• /
• 제31권4호
• /
• pp.383-392
• /
• 2023

Post-earthquake building condition assessment is crucial for subsequent rescue and remediation and can be automated by emerging computer vision and deep learning technologies. This study is based on an endeavour for the 2nd International Competition of Structural Health Monitoring (IC-SHM 2021). The task package includes five image segmentation objectives - defects (crack/spall/rebar exposure), structural component, and damage state. The structural component and damage state tasks are identified as the priority that can form actionable decisions. A multi-task Convolutional Neural Network (CNN) is proposed to conduct the two major tasks simultaneously. The rest 3 sub-tasks (spall/crack/rebar exposure) were incorporated as auxiliary tasks. By synchronously learning defect information (spall/crack/rebar exposure), the multi-task CNN model outperforms the counterpart single-task models in recognizing structural components and estimating damage states. Particularly, the pixel-level damage state estimation witnesses a mIoU (mean intersection over union) improvement from 0.5855 to 0.6374. For the defect detection tasks, rebar exposure is omitted due to the extremely biased sample distribution. The segmentations of crack and spall are automated by single-task U-Net but with extra efforts to resample the provided data. The segmentation of small objects (spall and crack) benefits from the resampling method, with a substantial IoU increment of nearly 10%.

• Structural Engineering and Mechanics
• /
• 제28권5호
• /
• pp.525-548
• /
• 2008

This article deals the theory for solving an inverse problem of plate structures using the frequency-domain information instead of classical time-domain delays or free vibration eigenmodes or eigenvalues. A reduced set of output parameters characterizing the defect is used as a regularization technique to drastically overcome noise problems that appear in imaging techniques. A deconvolution scheme from an undamaged specimen overrides uncertainties about the input signal and other coherent noises. This approach provides the advantage that it is not necessary to visually identify the portion of the signal that contains the information about the defect. The theoretical model for Quantitative nondestructive evaluation, the relationship between the real and ideal models, the finite element method (FEM) for the forward problem, and inverse procedure for detecting the defects are developed. The theoretical formulation is experimentally verified using dynamic responses of a steel plate under impact loading at several points. The signal synthesized by FEM, the residual, and its components are analyzed for different choices of time window. The noise effects are taken into account in the inversion strategy by designing a filter for the cost functional to be minimized. The technique is focused toward a exible and rapid inspection of large areas, by recovering the position of the defect by means of a single accelerometer, overriding experimental calibration, and using a reduced number of impact events.

• Structural Engineering and Mechanics
• /
• 제9권2호
• /
• pp.195-208
• /
• 2000

To confirm the theory and static defect energy (SDE) equations proposed in the first part, extensive numerical simulation studies are performed in this portion. Stiffness method is applied to calculate the components of the stresses and strains from which the energy components and finally, the SDE are obtained. Examples are designed to cover almost all kinds of possibilities. Variables include structural type, material, cross-section, support constraint, loading type, magnitude and position. The SDE diagram is unique in the way of presenting damage information: two different energy constants are separated by a sharp vertical drop right at the damage location. Simulation results are successfully implemented for both methods in all the cases.

• 한국구조물진단유지관리공학회 논문집
• /
• 제2권1호
• /
• pp.98-107
• /
• 1998

There are many weld defects such as surface crack, lack of fusion, and imcomplete penetration in the butt joint of the existing steel bridges. The crack-like defects may significantly reduce the fatigue life of the structure. This paper presents the procedure and the results of the fatigue assessment of the butt joints with weld defect in the existing steel girder bridge. The butt joints with imcomple penetration were instrumented with strain gages to determine the stress histogram under normal traffic. Based on the measured stress histogram the crack propagation analyses were performed for the fatigue assesment. By using the suggested procedure and methodology, one can decide the time of periodic inspection and the necessity of repair of the butt joints with serious weld defects in the existing steel bridge.

• 소성∙가공
• /
• 제31권4호
• /
• pp.229-239
• /
• 2022

Convolution Neural Network(CNN) is a class of deep learning algorithms and can be used for image analysis. In particular, it has excellent performance in finding the pattern of images. Therefore, CNN is commonly applied for recognizing, learning and classifying images. In this study, the surface defect classification performance of Al 6061 extruded material using CNN-based algorithms were compared and evaluated. First, the data collection criteria were suggested and a total of 2,024 datasets were prepared. And they were randomly classified into 1,417 learning data and 607 evaluation data. After that, the size and quality of the training data set were improved using data augmentation techniques to increase the performance of deep learning. The CNN-based algorithms used in this study were VGGNet-16, VGGNet-19, ResNet-50 and DenseNet-121. The evaluation of the defect classification performance was made by comparing the accuracy, loss, and learning speed using verification data. The DenseNet-121 algorithm showed better performance than other algorithms with an accuracy of 99.13% and a loss value of 0.037. This was due to the structural characteristics of the DenseNet model, and the information loss was reduced by acquiring information from all previous layers for image identification in this algorithm. Based on the above results, the possibility of machine vision application of CNN-based model for the surface defect classification of Al extruded materials was also discussed.

• Smart Structures and Systems
• /
• 제34권2호
• /
• pp.129-143
• /
• 2024

In recent years, with the vigorous development of visual algorithms, a large amount of research has been conducted on blade surface defect detection methods represented by deep learning. Detection methods based on deep learning models must rely on a large and rich dataset. However, the geographical location and working environment of wind turbines makes it difficult to effectively capture images of blade surface defects, which inevitably hinders visual detection. In response to the challenge of collecting a dataset for surface defects that are difficult to obtain, a multi-class blade surface defect generation method based on the StyleGAN3 (Style Generative Adversarial Networks 3) deep learning model and PBGMs (Physics-Based Graphics Models) method has been proposed. Firstly, a small number of real blade surface defect datasets are trained using the adversarial neural network of the StyleGAN3 deep learning model to generate a large number of high-resolution blade surface defect images. Secondly, the generated images are processed through Matting and Resize operations to create defect foreground images. The blade background images produced using PBGM technology are randomly fused, resulting in a diverse and high-resolution blade surface defect dataset with multiple types of backgrounds. Finally, experimental validation has proven that the adoption of this method can generate images with defect characteristics and high resolution, achieving a proportion of over 98.5%. Additionally, utilizing the EISeg annotation method significantly reduces the annotation time to just 1/7 of the time required for traditional methods. These generated images and annotated data of blade surface defects provide robust support for the detection of blade surface defects.