• Proceedings of the Korean Information Science Society Conference
• /
• 2005.07a
• /
• pp.892-894
• /
• 2005

Graph partitioning provides an important tool for data clustering, but is an NP-hard combinatorial optimization problem. Spectral clustering where the clustering is performed by the eigen-decomposition of an affinity matrix [1,2]. This is a popular way of solving the graph partitioning problem. On the other hand, semidefinite relaxation, is an alternative way of relaxing combinatorial optimization. issuing to a convex optimization[4]. In this paper we present a semidefinite programming (SDP) approach to graph equi-partitioning for clustering and then we use eigen-decomposition to obtain an optimal partition set. Therefore, the method is referred to as semidefinite spectral clustering (SSC). Numerical experiments with several artificial and real data sets, demonstrate the useful behavior of our SSC. compared to existing spectral clustering methods.

• Journal of computational fluids engineering
• /
• v.22 no.1
• /
• pp.59-66
• /
• 2017

The guideline of selecting the number of snapshot dataset, $N_s$ in proper orthogonal decomposition(POD) was presented via the analysis of Eigen values based on the singular value decomposition(SVD). In POD, snapshot datasets from the solutions of Euler or Navier-Stokes equations are utilized to SVD and a reduced order model(ROM) is constructed as the combination of Eigen vectors. The ROM is subsequently applied to reconstruct the flowfield data with new set of flow conditions, thereby enhancing the computational efficiency. The overall computational efficiency and accuracy of POD is dependent on the number of snapshot dataset; however, there is no reliable guideline of determining $N_s$. In order to resolve this problem, the order of maximum to minimum Eigen value ratio, O(R) from SVD was analyzed and presented for the decision of $N_s$; in case of steady flow, $N_s$ should be determined to make O(R) be $10^9$. For unsteady flow, $N_s$ should be increased to make O(R) be $10^{11\sim12}$. This strategy of selecting the snapshot dataset was applied to two dimensional NACA0012 airfoil and vortex flow problems including steady and unsteady cases and the numerical accuracies according to $N_s$ and O(R) were discussed.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.53 no.3
• /
• pp.59-66
• /
• 2016

This paper presents an empirical evaluation on dimensionality reduction strategies by which dissimilarity-based classifications (DBC) can be implemented efficiently. In DBC, classification is not based on feature measurements of individual objects (a set of attributes), but rather on a suitable dissimilarity measure among the individual objects (pair-wise object comparisons). One problem of DBC is the high dimensionality of the dissimilarity space when a lots of objects are treated. To address this issue, two kinds of solutions have been proposed in the literature: prototype selection (PS)-based methods and dimension reduction (DR)-based methods. In this paper, instead of utilizing the PS-based or DR-based methods, a way of performing DBC in Eigen spaces (ES) is considered and empirically compared. In ES-based DBC, classifications are performed as follows: first, a set of principal eigenvectors is extracted from the training data set using a principal component analysis; second, an Eigen space is expanded using a subset of the extracted and selected Eigen vectors; third, after measuring distances among the projected objects in the Eigen space using $l_p$-norms as the dissimilarity, classification is performed. The experimental results, which are obtained using the nearest neighbor rule with artificial and real-life benchmark data sets, demonstrate that when the dimensionality of the Eigen spaces has been selected appropriately, compared to the PS-based and DR-based methods, the performance of the ES-based DBC can be improved in terms of the classification accuracy.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.15 no.2
• /
• pp.445-454
• /
• 1991

In this thesis, Mindlin plate element with nine nodes and three degrees-of-freedom at each node is formulated and is employed in eigen-analysis of a rectangular plates in order to alleviate locking phenomenon of eigenvalues. Eigenvalues and their modes may be locked if conventional $C_{0}$-isoparametric element is used. In order to reduce stiffness locking phenomenon, two methods (1, the general reduced and selective integration, 2, the new element that use of modified shape function) are studied. Additionally in order to reduce the error due to mass matrix, two mass matrixes (1, Gauss-Legendre mass matrix, 2, Gauss-Lobatto mass matrix) are considered. The results of eigen-analysis for two models (the square plate with all edges simply-supported and all edges built-in), computed by two methods for stiffness matrix and by two mass matrixes are compared with theoretical solutions and conventional numerical solutions. These comparisons show that the performance of the two methods with Gauss-Lobatto mass matrix is better than that of the conventional plate element. But, by considering the spurious rigid body motions, the element which employs modified shape function with full integration and Gauss-Lobatto mass matrix can elevate the accuracy and convergence of numerical solutions.

• Smart Structures and Systems
• /
• v.15 no.5
• /
• pp.1215-1232
• /
• 2015

In the present paper, a method for identifying damage in a multi storeyed shear building structure is presented using minimum number of modal parameters of the structure. A damage at any level of the structure may lead to a major failure if the damage is not attended at appropriate time. Hence an early detection of damage is essential. The proposed identification methodology requires experimentally determined sparse modal data of any particular mode as input to detect the location and extent of damage in the structure. Here, the first natural frequency and corresponding partial mode shape values are used as input to the model and results are compared by changing the sensor placement locations at different floors to conclude the best location of sensors for accurate damage identification. Initially experimental data are simulated numerically by solving eigen value problem of the damaged structure with inclusion of random noise on the vibration characteristics. Reliability of the procedure has been demonstrated through a few examples of multi storeyed shear structure with different damage scenarios and various noise levels. Validation of the methodology has also been done using dynamic data obtained through experiment conducted on a laboratory scale steel structure.

• Convergence Security Journal
• /
• v.12 no.3
• /
• pp.3-10
• /
• 2012

This study aimed at making a way of securing the access by using PCA. We got our result through using Box-Plot and PCA with the access data of the area of security level A~E at K(IPS)center. In order to perform PCA, We confirmed the extracted value of commonality has no problem in performing PCA because VIF is below 2.902. Based on this result, We classified people into Green-list, Blue-list, Red-list, and Black-list in a standard of security level with 1.453, as the eigen value of 1 main element, 1.283, as eigen value of 2 main elementm, 1.142, as the eigen value of 3 main element.

• Steel and Composite Structures
• /
• v.26 no.4
• /
• pp.473-484
• /
• 2018

Free vibration of cross-ply laminated plates using a higher-order shear deformation theory is studied. The arbitrary number of layers is oriented in symmetric and anti-symmetric manners. The plate kinematics are based on higher-order shear deformation theory (HSDT) and the vibrational behaviour of multi-layered plates are analysed under simply supported boundary conditions. The differential equations are obtained in terms of displacement and rotational functions by substituting the stress-strain relations and strain-displacement relations in the governing equations and separable method is adopted for these functions to get a set of ordinary differential equations in term of single variable, which are coupled. These displacement and rotational functions are approximated using cubic and quantic splines which results in to the system of algebraic equations with unknown spline coefficients. Incurring the boundary conditions with the algebraic equations, a generalized eigen value problem is obtained. This eigen value problem is solved numerically to find the eigen frequency parameter and associated eigenvectors which are the spline coefficients.The material properties of Kevlar-49/epoxy, Graphite/Epoxy and E-glass epoxy are used to show the parametric effects of the plates aspect ratio, side-to-thickness ratio, stacking sequence, number of lamina and ply orientations on the frequency parameter of the plate. The current results are verified with those results obtained in the previous work and the new results are presented in tables and graphs.

• The Transactions of the Korean Institute of Electrical Engineers A
• /
• v.49 no.5
• /
• pp.226-232
• /
• 2000

This paper presents application results of the augmented matrix eigen-sensitivity theories to TCSC siting problem for damping the inter-area low frequency oscillation in the large KEPCO system. First and second-order eigen-sensitivities of the inter-area low frequency oscillation in the large KEPCO system. First and second-order eigen-sensitivities of the inter-area mode are computed fro changes in susceptance of the transmission lines. The lines having high sensitivity are chosen as the initial candidates for installing TCSC. Then for each of the chosen candidates, Bodeplot of the transfer function with line susceptance as the input and the bus voltage at one side of the line as the output is computed. Using the Bode plots, the lines having any zeros near the inter-area mode are screened out since design of TCSC controller is very difficult in such a case. The $H_{\infty}$ TCSC controller installed at any finally chosen candidate is found to be effective in damping the inter-area oscillation, and the proposed TCSC siting algorithm is proved to be valid. Design of $H_{\infty}$ controller is described in Part IIof this paper.

• Journal of the Korean Society for Precision Engineering
• /
• v.21 no.8
• /
• pp.129-135
• /
• 2004

This paper discussed on the effect of an intermediate support on the stability of elastic material subjected to dry friction force. It is assumed in this paper that the dry frictional force between a tool stand and an elastic material can be modeled as a distributed follower force. The elastic material on the friction material is modeled for simplicity into an elastic beam on Winkler-type elastic foundation. The stability of beams on the elastic foundation subjected to distributed follower force is formulated by using finite element method to have a standard eigenvalue problem. The first two eigen-frequencies are obtained to investigate the dynamics of the beam. The eigen-frequencies yield the stability bound and the corresponding unstable mode. The considered beams lose its stability by flutter or divergence, depending on the location of intermediate support.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2006.05a
• /
• pp.320-325
• /
• 2006

The phenomenon of squeal in disc brakes has been, and stin is, a problem for the automotive industry. Extensive research has been done in an attempt to understand the mechanisms that cause it and in developing design procedures to reduce it to make vehicles more comfortable. In this paper, the study on squeal noise of disc brake is performed using complex eigen-value analysis, The first part describes the chassis-dynamometer and the testing procedure, and second part explains how the analysis is performed and shows some of the results from typical squeal tests. Finally, to reduce squeal nose of disc brake is investigated by the effects of brake design parameter.