Camci F.,Meliksah University |
Chinnam R.B.,Wayne State University
IEEE Transactions on Automation Science and Engineering | Year: 2010
Failure mechanisms of electromechanical systems usually involve several degraded health-states. Tracking and forecasting the evolution of health-states and impending failures, in the form of remaining-useful-life (RUL), is a critical challenge and regarded as the Achilles' heel of condition-based- maintenance (CBM). This paper demonstrates how this difficult problem can be addressed through Hidden Markov models (HMMs) that are able to estimate unobservable health-states using observable sensor signals. In particular, implementation of HMM based models as dynamic Bayesian networks (DBNs) facilitates compact representation as well as additional flexibility with regard to model structure. Both regular HMM pools and hierarchical HMMs are employed here to estimate online the health-state of drill-bits as they deteriorate with use on a CNC drilling machine. Hierarchical HMM is composed of sub-HMMs in a pyramid structure, providing functionality beyond an HMM for modeling complex systems. In the case of regular HMMs, each HMM within the pool competes to represent a distinct health-state and adapts through competitive learning. In the case of hierarchical HMMs, health-states are represented as distinct nodes at the top of the hierarchy. Monte Carlo simulation, with state transition probabilities derived from a hierarchical HMM, is employed for RUL estimation. Detailed results on health-state and RUL estimation are very promising and are reported in this paper. Hierarchical HMMs seem to be particularly effective and efficient and outperform other HMM methods from literature. © 2010 IEEE.
Sevkat E.,Meliksah University |
Liaw B.,City College of New York |
Delale F.,City College of New York
Materials and Design | Year: 2013
Drop-weight impact response of hybrid woven composite plates was studied. Hybrid S2 glass-IM7 graphite fibers/toughened epoxy composites with two lay-up arrangements were impacted using spherical, flat-ended cylindrical and straight-line Charpy impactors. The time-histories of impact-induced dynamic strains and impact forces were recorded. The damaged specimens were inspected using ultrasonic C-Scan methods. Experimental results exhibited that hybrid composites with glass outer skins had higher resistance to impact compared to second type. It also delaminated more than hybrid composites with graphite outer skins. The 3-D dynamic finite element software, LS-DYNA, was then used to simulate the experimental result of drop-weight tests. Good agreement between experimental and FE results was achieved when comparing dynamic force, strain histories and damage patterns between experimental measurements and finite element simulations. © 2013 Elsevier Ltd.
Celik T.,Agency for Science, Technology and Research Singapore |
Celik T.,Meliksah University
Pattern Recognition | Year: 2012
In this paper, we propose a two-dimensional histogram equalization (2DHE) algorithm which utilizes contextual information around each pixel to enhance the contrast of an input image. The algorithm is based on the observation that the contrast in an image can be improved by increasing the grey-level differences between each pixel and its neighbouring pixels. The image equalization is achieved by assuming that for a given image, the modulus of the grey-level differences between pixels and their neighbouring pixels are equally distributed. The well-known global histogram equalization algorithm is a special case of 2DHE when contextual information is not utilized. 2DHE is easy to implement requiring only a small number of simple arithmetic operations and is thus suitable for real-time contrast enhancement applications. Experimental results show that 2DHE produces better or comparable enhanced images than several state-of-the-art algorithms. The only parameter in 2DHE which requires tuning is the size of the spatial neighbourhood support which provides the contextual information for a given dynamic range of the enhanced image. An automated parameter selection algorithm is also presented. The algorithm can be applied to a wide range of image types. © 2012 Elsevier Ltd.
Civril A.,Meliksah University |
Magdon-Ismail M.,Rensselaer Polytechnic Institute
Theoretical Computer Science | Year: 2012
Sevkat E.,Meliksah University |
Tumer H.,Meliksah University
Materials and Design | Year: 2013
This paper presents an experimental and numerical study to investigate residual torsional properties of composite shafts subjected to impact loadings. E-glass/epoxy, carbon/epoxy and E-glass-carbon/epoxy hybrid composite shafts were manufactured by filament winding method. Composite shafts were impacted at 5, 10, 20 and 40. J energy levels. Force-time and energy-time histories of impact tests were recorded. One composite shaft with no impact, and four composite shafts with impact damage, five in total, were tested under torsion. Torque-twisting angle relations for each test were obtained. Reduction at maximum torque and maximum twisting angle induced by impact loadings were calculated. While 5. J impact did not cause significant reduction at maximum torque and maximum twisting angle, remaining impact loadings caused 34-67% reduction at maximum torque, and 30-61% reduction at maximum twisting angle. Reductions increased with increasing energy levels and varied depending on the material of composite shafts. The 3-D finite element (FE) software, Abaqus, incorporated with an elastic orthotropic model, was then used to simulate the torsion tests. Good agreement between experimental and numerical results was achieved. © 2013 Elsevier Ltd.
Ozturk I.,Cag University |
Aslan A.,University of Nevsehir |
Kalyoncu H.,Meliksah University
Energy Policy | Year: 2010
This paper uses the panel data of energy consumption (EC) and economic growth (GDP) for 51 countries from 1971 to 2005. These countries are divided into three groups: low income group, lower middle income group and upper middle income group countries. Firstly, a relationship between energy consumption and economic growth is investigated by employing Pedroni (1999) panel cointegration method. Secondly, panel causality test is applied to investigate the way of causality between the energy consumption and economic growth. Finally, we test whether there is a strong or weak relationship between these variables by using Pedroni (2001) method. The empirical results of this study are as follows: i) Energy consumption and GDP are cointegrated for all three income group countries. ii) The panel causality test results reveal that there is long-run Granger causality running from GDP to EC for low income countries and there is bidirectional causality between EC and GDP for middle income countries. iii) The estimated cointegration factor, Β, is not close to 1. In other words, no strong relation is found between energy consumption and economic growth for all income groups considered in this study. The findings of this study have important policy implications and it shows that this issue still deserves further attention in future research. © 2010 Elsevier Ltd.
Civril A.,Meliksah University
Journal of Computer and System Sciences | Year: 2014
We address two problems related to selecting an optimal subset of columns from a matrix. In one of these problems, we are given a matrix AεRm×n and a positive integer k, and we want to select a sub-matrix C of k columns to minimize â€-A- ΠCAâ€-F, where ΠC=CC+ denotes the matrix of projection onto the space spanned by C. In the other problem, we are given AεRm×n, positive integers c and r, and we want to select sub-matrices C and R of c columns and r rows of A, respectively, to minimize â€-A-CURâ€-F, where UεRc×r is the pseudo-inverse of the intersection between C and R. Although there is a plethora of algorithmic results, the complexity of these problems has not been investigated thus far. We show that these two problems are NP-hard assuming UGC. © 2014 Elsevier Inc.
Peker K.A.,Meliksah University
Proceedings - International Workshop on Content-Based Multimedia Indexing | Year: 2011
SIFT features are widely used in content based image retrieval. Typically, a few thousand keypoints are extracted from each image. Image matching involves distance computations across all pairs of SIFT feature vectors from both images, which is quite costly. We show that SIFT features perform surprisingly well even after quantizing each component to binary, when the medians are used as the quantization thresholds. Quantized features preserve both distinctiveness and matching properties. Almost all of the features in our 5.4 million feature test set map to distinct binary patterns after quantization. Furthermore, number of matches between images using both the original and the binary quantized SIFT features are quite similar. We investigate the distribution of SIFT features and observe that the space of 128-D binary vectors has sufficient capacity for the current performance of SIFT features. We use component median values as quantization thresholds and show through vector-to-vector distance comparisons and image-to-image matches that the resulting binary vectors perform comparable to original SIFT vectors. We also discuss computational and storage gains. Binary vector distance computation reduces to bit-wise operations. Square operation is eliminated. Fast and efficient indexing techniques such as the signatures used for chemical databases can also be considered. © 2011 IEEE.
Uzam M.,Meliksah University
International Journal of Advanced Manufacturing Technology | Year: 2012
The Supervisory Control Theory (SCT) introduced by Ramadge-Wonham (RW) is a general theory to design supervisors (controllers) for discrete event systems. Although over the two decades SCT has received wide attention in academia, industrial applications are very few, due to the fact that there is a discrepancy between the abstract RW supervisor and its physical implementation. In this paper, an easy to use, general and practical technique is proposed for the PLC-based implementation of RW supervisors with time delay functions. The applicability of the proposed method is demonstrated by means of a PLC-based real-time control of an experimental manufacturing system. © Springer-Verlag London Limited 2011.
Sevkat E.,Meliksah University
International Journal of Impact Engineering | Year: 2012
This paper presents an experimental and numerical study to estimate ballistic limit velocity, V 50, of plain-weave hybrid S2-glass-IM7 graphite fibers/toughened SC-79 resin composite beams. The tests were conducted on hybrid S2-glass-IM7 graphite fibers/toughened SC-79 resin and non-hybrid S2 glass-fiber/toughened SC-79 resin composites beams using a high-speed gas gun. The ballistic impact tests were then modeled using a 3-D dynamic nonlinear finite element (FE) code, LS-DYNA, modified with a proposed user-defined nonlinear-orthotropic damage model. The ballistic limit velocities, V 50, for both composite beams were then estimated using three approaches: (1) Only experimental tests, (2) combined experimental and numerical tests, (3) FE calculated residual velocities, and FE calculated residual energies. For each type of composite beams, the parameters for the well-known Lambert-Jones equation were also computed. Good agreement between experimental and numerical results was observed. © 2012 Elsevier Ltd. All rights reserved.