• KIEAE Journal
• /
• v.16 no.1
• /
• pp.81-87
• /
• 2016

Purpose: Heat transfers phenomena are described by the second order partial differential equation and its boundary conditions. In a three-dimensional structure of a building, the heat transfer phenomena generally include more than one material, and thus, become complicate. The analytic solutions are useful to understand heat transfer phenomena, but they can hardly be applied in engineering or design problems. Engineers and designers have generally been forced to use numerical methods providing reliable results. Finite volume methods with the unstructured grid system is only the suitable means of the analysis for the complex and arbitrary domains. Method: To obtain an numerical solution, a discretization method, which approximates the differential equations, and the interpolation methods for temperature and heat flux between two or more materials are required. The discretization methods are applied to small domains in space and time, and these numerical solutions form the descretized equations provide approximated solutions in both space and time. The accuracy of numerical solutions is dependent on the quality of discretizations and size of cells used. The higher accuracy, the higher numerical resources are required. The balance between the accuracy and difficulty of the numerical methods is critical for the success of the numerical analysis. A simple and easy interpolation methods among multiple materials are developed. The linear equations are solved with the BiCGSTAB being a effective matrix solver. Result: This study provides an overview of discretization methods, boundary interface, and matrix solver for the 3-dimensional numerical heat transfer including two materials.

• The Journal of Advanced Prosthodontics
• /
• v.15 no.2
• /
• pp.72-79
• /
• 2023

PURPOSE. The purpose of this in vitro study was to compare the accuracy of various 3D printers and a milling machine. MATERIALS AND METHODS. The die model was designed using CAD (Autodesk Inventor 2018 sp3). The 30 ㎛ cement space was given to the die and the ideal crown of the mandibular left first molar was designed using CAD (ExoCAD). The crowns were produced using the milling machine (Imes-icore 250i) and the 3D printers (Zenith U, Zenith D, W11) and they were divided into four groups. In all groups, the interior of each crown was scanned (Identica blue) and superimposed (Geomagic Control X) with the previously designed die. The difference between the die and the actual crown was measured at specific points. The Kruskal-Wallis test, the Mann-Whitney test, and Bonferroni's method were performed with a statistical analysis software (P < .008 in inter-group comparison P < .001 in intra-group comparison). RESULTS. In all groups, the center of the occlusal area and the anti-rotational dimple area showed significantly greater difference and the marginal area showed the smallest difference comparatively. The mean value of the difference in each area and the sum of the differences were higher in order of W11, Imes-icore 250i, Zenith D, and Zenith U. CONCLUSION. The digital light processing (DLP) method shows higher accuracy compared to the sereolithography (SLA) method using the same resin material.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.5
• /
• pp.1433-1449
• /
• 2023

In recent years, the rapid development of social networks has led to a rapid increase in the amount of information available on the Internet, which contains a large amount of sensitive information related to pornography, politics, and terrorism. In the aspect of sensitive image detection, the existing machine learning algorithms are confronted with problems such as large model size, long training time, and slow detection speed when auditing and supervising. In order to detect sensitive images more accurately and quickly, this paper proposes a multiclassification sensitive image detection method based on lightweight Convolutional Neural Network. On the basis of the EfficientNet model, this method combines the Ghost Module idea of the GhostNet model and adds the SE channel attention mechanism in the Ghost Module for feature extraction training. The experimental results on the sensitive image data set constructed in this paper show that the accuracy of the proposed method in sensitive information detection is 94.46% higher than that of the similar methods. Then, the model is pruned through an ablation experiment, and the activation function is replaced by Hard-Swish, which reduces the parameters of the original model by 54.67%. Under the condition of ensuring accuracy, the detection time of a single image is reduced from 8.88ms to 6.37ms. The results of the experiment demonstrate that the method put forward has successfully enhanced the precision of identifying multi-class sensitive images, significantly decreased the number of parameters in the model, and achieved higher accuracy than comparable algorithms while using a more lightweight model design.

• Restorative Dentistry and Endodontics
• /
• v.21 no.1
• /
• pp.289-299
• /
• 1996

It is necessary to measure the length of a root canal in order to attain a satisfactory prognosis after root canal therapy. There are several methods for determining root canal length, such as tactile sensation by the dental practitioner, the utilization of x-ray film, and electronic root canal measurement. Among these, the electrical measurement methods, in which the impedence between the oral mucous membrane and periodontal membrane is determined, have advantages of simplicity and accuracy. During root canal treatment, the root canal contains a solution of high electrical conductivity such as pus, blood, sodium hypochlorite and so on. Recently a new electronic root canal measurement device of frequency-dependent type has been developed, which is capable of measuring the length of root canal under moist conditions. Endex and Root ZX, which are frequency-dependent type, were evaluated for accuracy of measuring root canal length in vivo by stereomicroscope. The result were as follows ; 1. 82.5% of Endex and 87.5% of Root ZX measured in the range of ${\pm}0.5$ mm from the apical foramen and both showed 57.5 % in the range of 0.1 mm to 0.5 mm. 2. Endex showed significantly higher accuracy in vital teeth than nonvital teeth(p<0.05). But in case of Root ZX, there was no significant difference between vital and nonvital teeth. 3. As a result of this study, there was no significant difference in accuracy between Endex and Root ZX, and both devices showed file passes the apical foramen in more than half of the cases, and it is thought that this must be considered clinically.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.15 no.3
• /
• pp.186-189
• /
• 2003

Various empirical formulas have been developed to describe the stability of breakwater armor stones. The Hudson formula is known to have needs to be refined in many ways. but it is most widely used because of its simplicity. The van der Meer formula is sometimes used for the cross-check, but it is not very popularly used due to its complexity and the uncertainty of the relevant parameters. Recently, on the other hand, Yoo(2003) proposed a formula of a simple form using a new non-dimensional number which is called 'action slope' (abbreviated as Yoo formula hereinafter). In this study, in order to provide coastal engineers with some idea of the use of these formulas, their accuracies were estimated by comparing with the experimental data reported by van der Meer in 1987 and 1988. It was found that the van der Meer formula showed the highest accuracy, while the Hudson formula of very low accuracy needed considered judgment on its use. On the other hand, the Yoo formula of slightly lower accuracy but simpler expression than the van der Meer formula seems to be able to be used in engineering practice if a slightly higher safety factor is taken.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.40 no.1
• /
• pp.15-22
• /
• 2022

In order to improve the accuracy of the deep learning object detection technique, the effect of magnification rate conditions and seasonal factors on detection accuracy in aerial photographs and drone images was analyzed through experiments. Among the deep learning object detection techniques, Mask R-CNN, which shows fast learning speed and high accuracy, was used to detect the vehicle to be detected in pixel units. Through Seoul's aerial photo service, learning images were captured at different screen magnifications, and the accuracy was analyzed by learning each. According to the experimental results, the higher the magnification level, the higher the mAP average to 60%, 67%, and 75%. When the magnification rates of train and test data of the data set were alternately arranged, low magnification data was arranged as train data, and high magnification data was arranged as test data, showing a difference of more than 20% compared to the opposite case. And in the case of drone images with a seasonal difference with a time difference of 4 months, the results of learning the image data at the same period showed high accuracy with an average of 93%, confirming that seasonal differences also affect learning.

• Journal of the Korea Society of Computer and Information
• /
• v.22 no.4
• /
• pp.41-49
• /
• 2017

Word sense disambiguation(WSD) that determines the exact meaning of homonym which can be used in different meanings even in one form is very important to understand the semantical meaning of text document. Many recent researches on WSD have widely used NNLM(Neural Network Language Model) in which neural network is used to represent a document into vectors and to analyze its semantics. Among the previous WSD researches using NNLM, RNN(Recurrent Neural Network) model has better performance than other models because RNN model can reflect the occurrence order of words in addition to the word appearance information in a document. However, since RNN model uses only the forward order of word occurrences in a document, it is not able to reflect natural language's characteristics that later words can affect the meanings of the preceding words. In this paper, we propose a WSD scheme using Bidirectional RNN that can reflect not only the forward order but also the backward order of word occurrences in a document. From the experiments, the accuracy of the proposed model is higher than that of previous method using RNN. Hence, it is confirmed that bidirectional order information of word occurrences is useful for WSD in Korean language.

• Journal of the Society of Naval Architects of Korea
• /
• v.42 no.2 s.140
• /
• pp.113-123
• /
• 2005

A higher order panel method based on B-spline representation for both the geometry and the solution is developed for the analysis of steady flow around marine propellers. The self-influence functions due to the normal dipole and the source are desingularized through the quadratic transformation, and then shown to be evaluated using conventional numerical quadrature. By selecting a proper order for numerical quadrature, the accuracy of the present method can be increased to the machine limit. The far- and near-field influences are shown to be evaluated based on the same far-field approximation, but the near-field solution requires subdividing the panels into smaller subpanels continuously, which can be effectively implemented due to the B-spline representation of the geometry. A null pressure jump Kutta condition at the trailing edge is found to be effective in stabilizing the solution process and in predicting the correct solution. Numerical experiments indicate that the present method is robust and predicts the pressure distribution on the blade surface, including very close to the tip and trailing edge regions, with far fewer panels than existing low order panel methods.

• Journal of computational fluids engineering
• /
• v.9 no.1
• /
• pp.34-40
• /
• 2004

Recently with growth in the capability of super computers and Parallel computers, shape design optimization is becoming easible for real problems. Also, Computational Fluid Dynamics(CFD) techniques have been improved for higher reliability and higher accuracy. In the shape design optimization, analysis solvers and optimization schemes are essential. In this work, the Roe's 2nd-order Upwind TVD scheme and DADI time march with multigrid were used for the flow solution with the Euler equation and FDM(Finite Differenciation Method), GA(Genetic Algorithm) and Kriging were used for the design optimization. Kriging were applied to 2-D airfoil design optimization and compared with FDM and GA's results. When Kriging is applied to the nonlinear problems, satisfactory results were obtained. From the result design optimization by Kriging method appeared as good as other methods.

• Journal of Institute of Control, Robotics and Systems
• /
• v.10 no.9
• /
• pp.840-846
• /
• 2004

In this study, the performance of a heavy load pointing system has been investigated. The PI controller are being widely used in industrial application because of simple, cheap, and excellent performance. However, the requirement for control precision becomes higher and higher, as well as the plants becomes more and more complex. In order to achieve the satisfied control performance, we have to consider the affection of nonlinear factor contained in plant. In this paper, the neural-PI control law have been evaluated. The proposed controller is compared with the existing controllers through simulations, and the results show that the pointing accuracy of the proposed control system is improved against the disturbance induced by vehicle running on the bump course.