Showing papers by "Ghulam Ishaq Khan Institute of Engineering Sciences and Technology published in 2006"

Finite element prediction of thermal stresses and deformations in layered manufacturing of metallic parts

Abstract: Residual stress induced deformations are a major cause of loss in tolerances in Solid Freeform Fabrication processes employing direct metal deposition. In this paper, a 2D finite element thermo-mechanical model is presented to predict the residual stress induced deformations with application to processes where material is added using a distributed, moving heat source. A sequentially coupled thermo-mechanical analysis is performed using a kinematic thermal model and plane strain structural model. Temperature dependent material properties are used with the material modeled as elastic perfectly plastic. An interpass cooling between successive depositions is employed in accordance to the requirement of experiment. The simulation results are compared with experimental data for successive sections along deposition and it is found that, with the exception of deposition center and plate edges, the two are in very good agreement. The error at plate edges can be as high as 45%, and the reason is that a 2D model cannot capture the effect of plate bolting accurately. A case of continuous deposition, without interpass cooling, has been compared with the base case of employing interpass cooling. It has been found that continuous deposition results in higher preheating of the substrate which consequently reduces the deformation.

Combination of support vector machines using genetic programming

Abstract: This paper describes the combination of support vector machine (SVM) classifiers using Genetic Programming (GP) for gender classification problem. In our scheme, individual SVM classifiers are constructed through the learning of different SVM kernel functions. The predictions of SVM classifiers are then combined using GP to develop Optimal Composite Classifier (OCC). In this way, the combined decision space is more informative and discriminant. OCC has shown improved performance than that of optimized individual SVM classifiers using grid search. Another advantage of our GP combination scheme is that it automatically incorporates the issues of optimal kernel function and model selection to achieve high performance classification model. The classification performance is reported by using Receiver Operating Characteristics (ROC) Curve. Experiments are conducted under various feature sets to show that OCC is more informative and robust as compared to their individual SVM classifiers. Specifically, it attains high margin of improvement for small feature sets.

Intelligent perceptual shaping of a digital watermark: Exploiting Characteristics of human visual system

Abstract: In this paper, we present a method for developing a Genetic Perceptual Model (GPM) applicable to a watermarking system. The proposed technique exploits the characteristics of human visual system using a Genetic Programming (GP) approach. We employ a tradeoff between watermark robustness and imperceptibility, as an optimization criterion in the GP search. The resultant GPM is a combination of frequency, luminance sensitivity and contrast masking, enabling us to shape the watermark according to the cover image. Our investigations have shown that the evolved GPM provides maximum allowable imperceptible alterations to the Discrete Cosine Transform coefficients of a cover image. Comparative studies in terms of watermark imperceptibility and bit correct ratio performance have been carried out. The performance of the GPM has been analyzed for various watermarking schemes.

Investigation of the characteristics of the fuel injection pump of a diesel engine fuelled with viscous vegetable oil-diesel oil blends

Abstract: The potential of vegetable oils as an alternative to diesel fuel has already been proved. Nevertheless, the high viscosity of vegetable oils poses some problems concerning their use in unmodified vehicle diesels. In particular, distortion of the speed characteristics of a distributor fuel injection pump with a maximum-minimum speed mechanical governor when using sunflower oil-diesel oil blends has been observed. A diesel engine fuelled by a viscous blend could not develop its rated power. After fuel delivery was increased to retain the rated power, the engine was overfuelled at lower speeds and the smoke limit was exceeded. In the present paper, measures required to retain the characteristics of an injection pump fuelled by viscous blends are described. This can be done by readjusting the pump governor and changing the hydraulic characteristics of the fuel delivery system.

Enhancing Contrast of Compressed Images: Reducing Block Artifacts Adaptively

Abstract: Block artifacts in DCT-based compression are a significant problem. A small reduction in block artifacts could easily improve the quality of the resultant image. For this purpose, in the present work, we propose a novel approach that adaptively enhances the DCT coefficients relying on their band importance. The proposed technique is designed to enhance the contrast in DCT domain and moreover preserves edges and details. Since the image enhancement is performed in DCT domain, therefore, the compressibility of the original image is not affected. Experimental results on standard images validate the concept of adaptively modifying different bands.

Oxides of Nitrogen (NOx) Emission Levels of Diesel Engines of Switch Locomotives

Abstract: This paper presents the results of a study that was carried out to determine the emission levels of engines of switch locomotives used in Ukraine and the compliance of the engines with existing and future standards. Test results showed that none of the engines tested fully complied with United States Environmental Protection Agency and European Union standards. An experimental investigation to evaluate some of the emission reduction technologies for oxides of nitrogen, such as optimization of operating characteristics and injection timing retardation, was also conducted.

A Novel Color Image Retrieval System Based On Dual Tree Complex Wavelet Transform and Support Vector Machines

Abstract: This paper presents a novel color image retrieval system (SVMBIR) based on dual tree complex wavelet transform (CWT) and support vector machines (SVM). We have shown that how one can improve the performance of image retrieval systems by assuming two attributes. Firstly, images that user needs through query image are similar to a group of images with same conception. Secondly, there exists non-linear relationship between feature vectors of different images and can be exploited very efficiently with the use of support vector machines. At first level, for low level feature extraction we have used dual tree complex wavelet transform because recently it is proven that CWT can be used for efficient color feature extraction just like texture feature extraction. At second level to extract semantic concepts, we grouped images of typical classes with the use of one against all support vector machines. We have also shown how one can use a correlation based distance metric for comparison of SVM distance vectors. The experimental results show that the proposed approach has superior retrieval performance over the existing linear feature combining techniques.

Affine Normalized Invariant functionals using Independent Component Analysis

Abstract: The paper presents a hybrid technique for affine invariant feature extraction with the view of object recognition based on parameterized contour The presented technique first normalizes an input image by removing affine distortions using independent component analysis which also reduces the effect of noise introduced during contour parameterization Then two invariant functionals are constructed, one using the normalized object contour and the other using the dyadic wavelet transform Experimental results conducted using three different standard datasets confirm the validity of the proposed approach Beside this the error rates obtained in terms of invariant stability are significantly lower when compared to other wavelet based techniques and the proposed invariants exhibit higher feature disparity than the method of Fourier descriptors

Visual Sign Language Interpretation Using Spatial Temporal Neural Processing

Abstract: An approach to recognizing human hand gestures from a monocular temporal sequence of color images is presented in this paper. The focus is on the representation and recognition of hand movements that are used in single handed (right hand is the dominant hand) American sign language (ASL). This work has been done with the help of markov chains while we have carried it out with STANN. The first level of decomposition is in terms of three sets of primitives, hand shape, location and movement. Further levels of decomposition involve the lexical and sentence levels and are part of future work. The approach has been trained and tested with an overall recognition of 100%

Tuned Spiral Resonators for Composite Metamaterials

Abstract: Metamaterials composed of thin wires and spiral resonators have been studied in this paper. Technique to tune spiral resonators have been proposed. Transmission and reflection properties of metamaterials composed of thin wires and tuned spiral resonators have been simulated and compared with conventional metamaterials. The center frequency and transmission bandwidth of metamaterials have been varied through tuning of spiral resonators.

Affine Normalized Invariant Feature Extraction using Multiscale Gabor Autoconvolution

Abstract: The paper presents a hybrid technique for affine invariant feature extraction with the view of object recognition. The proposed technique first normalizes an input image by removing affine distortions from it and then spatially re-samples the affine normalized image across multiple scales, next the Gabor transform is computed for the resampled images over different frequencies and orientations. Finally autoconvolution is performed in the transformed domain to generate a set of 384 invariants. Experimental results conducted using four different standard datasets confirm the validity of the proposed approach. Beside this the error rates obtained in terms of invariant stability are significantly lower when compared to Fourier based MSA, which has proven itself to be better than moment invariants

Mathematical and numerical modeling of melting in metal-ceramic composite beds

Abstract: Many of the ceramics used in the manufacture of the metal-ceramic composites are available as powders, which has encouraged investigators to develop tool less manufacturing techniques. Innovative processes like combustion synthesis and laser sintering processes offer great potential for specific material combinations. These processes involve heat and mass transfer in porous matrix with melting, solidification and may be chemical reaction. Since the resulting transport is relatively complex, there are few mathematical models available. In the present study, melting and transport of metal is demonstrated for a one-dimensional bed with uniform heat generation and convectively cooled boundaries. The effects of different Bond, Biot and Prandtl numbers are discussed using constant and temperature varying material properties. At low bond number the transport is controlled by capillary forces and temperature and saturation distribution is symmetric about the center of the bed. For Biot number larger than ten the process is internally controlled. Effect of Prandtl number is opposite of that of the Bond number. Use of constant properties overpredicts the process rate by 13%.

Lyapunov equations approximation using implicit restart scheme

Abstract: Large scale systems with few inputs and outputs as compared to high order state variables occur frequently in control systems. These systems have input-output union lead by a few states. The aim is to make use of those dominant states and to find a low rank approximation that is a better representation of original solution of Lyapunov equations. This approximation can be used in analysis, design and model reduction of control systems. In this paper a numerical method for computation of low rank approximation of Lyapunov equation is presented. The proposed method is based on Arnoldi Krylov subspace projection method with implicit restart scheme as a supplement to refine the results. Simulation of a system has been shown to validate the proposed technique

Algebraic Riccati Equation's Approximation Using Implicit Restart Scheme

Abstract: Analysis and design of complex systems which are composed of larger number of different devices arise in several engineering applications These systems have a few dominant states which influence output of the system to a greater extent and are more sensitive to the input of system The aim is to make use of those dominant states and to find a low rank approximation that is a better representation of original solution of Algebraic Riccati equations In this paper a numerical method for computation of low rank approximation of algebraic Riccati equation is presented The proposed method is based on Arnoldi Krylov subspace projection method with implicit restart scheme as an enhancement to refine the results Simulation of a system has been shown to authenticate the proposed technique