Convenient construction of tetrahedral finite elements for the quasi-harmonic equation

It is shown that in the finite-element formulation of the general quasi-harmonic equation using tetrahedral elements, for every member of the element family there exists just one numerical universal matrix indpendent of the size, shape and material properties of the element. Thus the element matrix is conveniently constructed by manipulating this single matrix along with a set of reverse sequence codes at the same time accounting for the size, shape and material properties in a simple manner.

Constant conductance fluctuations in disordered metals

We discuss the effect of fluctuations of the random potential in directions transverse to the current flow in a modified Migdal-Kadanoff approach to probabilistic scaling of conductance with size L, in d-dimensional metallic systems. The conductance cumulants are finite and vary as Ld−1−n for n greater-or-equal, slanted 2 i.e. conductance fluctuations are constant for d = 3. The mean conductance has a non-classical correction with Image Full-size image (<1K) for d greater-or-equal, slanted 2. The form of the higher cumulants is strongly influenced by the transverse potential fluctuations and may be compared with the results of perturbative diagrammatic approaches.

Sol → gel transition with rings in a finite polycondensing system in solution

We report a theoretical formulation for the mean cluster size distribution in a finite polycondensing system. Expressions for the mean number of n-mers with j bonds ( nj) are developed. Numerical calculations show that while the non-cyclic molecules make the dominant contribution to the small clusters, the large clusters are dominated by cyclic structures. The number of particles in ringless chains, n n,n-1, decays monotonically with n at all extents of reaction, but n n becomes bimodal near the gel point. We also find that the solvent plays an important role in the cluster size distribution.

Size optimization of a cantilever beam under deformation-dependent loads with application to wheat stalks

In this paper, we consider the optimization of the cross-section profile of a cantilever beam under deformation-dependent loads. Such loads are encountered in plants and trees, cereal crop plants such as wheat and corn in particular. The wind loads acting on the grain-bearing spike of a wheat stalk vary with the orientation of the spike as the stalk bends; this bending and the ensuing change in orientation depend on the deformation of the plant under the same load.The uprooting of the wheat stalks under wind loads is an unresolved problem in genetically modified dwarf wheat stalks. Although it was thought that the dwarf varieties would acquire increased resistance to uprooting, it was found that the dwarf wheat plants selectively decreased the Young's modulus in order to be compliant. The motivation of this study is to investigate why wheat plants prefer compliant stems. We analyze this by seeking an optimal shape of the wheat plant's stem, which is modeled as a cantilever beam, by taking the large deflection of the stem into account with the help of co-rotational finite element beam modeling. The criteria considered here include minimum moment at the fixed ground support, adequate stiffness and strength, and the volume of material. The result reported here is an example of flexibility, rather than stiffness, leading to increased strength.

Performance of slotted ALOHA satellite channels with finite buffer

The behaviour of the slotted ALOHA satellite channel with a finite buffer at each of the user terminals is studied. Approximate relationships between the queuing delay, overflow probabilities and buffer size are derived as functions of the system input parameters (i.e. the number of users, the traffic intensity, the transmission and the retransmission probabilities) for two cases found in the literature: the symmetric case (same transmission and retransmission probabilities), and the asymmetric case (transmission probability far greater than the retransmission probability). For comparison, the channel performance with an infinite buffer is also derived. Additionally, the stability condition for the system is defined in the latter case. The analysis carried out in the paper reveals that the queuing delays are quite significant, especially under high traffic conditions.

Performance of slotted ALOHA satellite channels with finite buffer

The behaviour of the slotted ALOHA satellite channel with a finite buffer at each of the user terminals is studied. Approximate relationships between the queuing delay, overflow probabilities and buffer size are derived as functions of the system input parameters (i.e. the number of users, the traffic intensity, the transmission and the retransmission probabilities) for two cases found in the literature: the symmetric case (same transmission and retransmission probabilities), and the asymmetric case (transmission probability far greater than the retransmission probability). For comparison, the channel performance with an infinite buffer is also derived. Additionally, the stability condition for the system is defined in the latter case. The analysis carried out in the paper reveals that the queuing delays are quite significant, especially under high traffic conditions.

An FETI-preconditioned conjuerate gradient method for large-scale stochastic finite element problems

In the spectral stochastic finite element method for analyzing an uncertain system. the uncertainty is represented by a set of random variables, and a quantity of Interest such as the system response is considered as a function of these random variables Consequently, the underlying Galerkin projection yields a block system of deterministic equations where the blocks are sparse but coupled. The solution of this algebraic system of equations becomes rapidly challenging when the size of the physical system and/or the level of uncertainty is increased This paper addresses this challenge by presenting a preconditioned conjugate gradient method for such block systems where the preconditioning step is based on the dual-primal finite element tearing and interconnecting method equipped with a Krylov subspace reusage technique for accelerating the iterative solution of systems with multiple and repeated right-hand sides. Preliminary performance results on a Linux Cluster suggest that the proposed Solution method is numerically scalable and demonstrate its potential for making the uncertainty quantification Of realistic systems tractable.

Cluster based training for scaling non-linear support vector machines

Support Vector Machines(SVMs) are hyperplane classifiers defined in a kernel induced feature space. The data size dependent training time complexity of SVMs usually prohibits its use in applications involving more than a few thousands of data points. In this paper we propose a novel kernel based incremental data clustering approach and its use for scaling Non-linear Support Vector Machines to handle large data sets. The clustering method introduced can find cluster abstractions of the training data in a kernel induced feature space. These cluster abstractions are then used for selective sampling based training of Support Vector Machines to reduce the training time without compromising the generalization performance. Experiments done with real world datasets show that this approach gives good generalization performance at reasonable computational expense.

Spatial survival probability for one-dimensional fluctuating interfaces in the steady state

We report numerical and analytic results for the spatial survival probability for fluctuating one-dimensional interfaces with Edwards-Wilkinson or Kardar-Parisi-Zhang dynamics in the steady state. Our numerical results are obtained from analysis of steady-state profiles generated by integrating a spatially discretized form of the Edwards-Wilkinson equation to long times. We show that the survival probability exhibits scaling behavior in its dependence on the system size and the "sampling interval" used in the measurement for both "steady-state" and "finite" initial conditions. Analytic results for the scaling functions are obtained from a path-integral treatment of a formulation of the problem in terms of one-dimensional Brownian motion. A "deterministic approximation" is used to obtain closed-form expressions for survival probabilities from the formally exact analytic treatment. The resulting approximate analytic results provide a fairly good description of the numerical data.

An improved iterative finite element solution for pin joints

Accurate, reliable and economical methods of determining stress distributions are important for fastener joints. In the past the contact stress problems in these mechanically fastened joints using interference or push or clearance fit pins were solved using both inverse and iterative techniques. Inverse techniques were found to be most efficient, but at times inadequate in the presence of asymmetries. Iterative techniques based on the finite element method of analysis have wider applications, but they have the major drawbacks of being expensive and time-consuming. In this paper an improved finite element technique for iteration is presented to overcome these drawbacks. The improved iterative technique employs a frontal solver for elimination of variables not requiring iteration, by creation of a dummy element. This automatically results in a large reduction in computer time and in the size of the problem to be handled during iteration. Numerical results are compared with those available in the literature. The method is used to study an eccentrically located pin in a quasi-isotropic laminated plate under uniform tension.

Scaling analysis of momentum and heat transport in gas tungsten arc weld pools

A systematic procedure is outlined for scaling analysis of momentum and heat transfer in gas tungsten arc weld pools. With suitable selections of non-dimentionalised parameters, the governing equations coupled with appropriate boundary conditions are first scaled, and the relative significance of various terms appearing in them is analysed accordingly. The analysis is then used to predict the orders of magnitude of some important quantities, such as the velocity scene lit the top surface, velocity boundary layer thickness, maximum temperature increase in the pool, and time required for initiation of melting. Some of the quantities predicted from the scaling analysis can also be used for optimised selection of appropriate grid size and time steps for full numerical simulation of the process. The scaling predictions are finally assessed by comparison with numerical results quoted in the literature, and a good qualitative agreement is observed.

A scaling analysis of momentum and heat transport in laser surface melting

In this paper, we outline a systematic procedure for scaling analysis of momentum and heat transfer in laser melted pools. With suitable choices of non-dimensionalising parameters, the governing equations coupled with appropriate boundary conditions are first scaled, and the relative significance of various terms appearing in them are accordingly analysed. The analysis is then utilised to predict the orders of magnitude of some important quantities, such as the velocity scale at the top surface, velocity boundary layer thickness, maximum temperature rise in the pool, fully developed pool-depth, and time required for initiation of melting. Using the scaling predictions, the influence of various processing parameters on the system variables can be well recognised, which enables us to develop a deeper insight into the physical problem of interest. Moreover, some of the quantities predicted from the scaling analysis can be utilised for optimised selection of appropriate grid-size and time-steps for full numerical simulation of the process. The scaling predictions are finally assessed by comparison with experimental and numerical results quoted in the literature, and an excellent qualitative agreement is observed.

An extended Cahn-Hilliard model for interfaces with cubic anisotropy

For studying systems with a cubic anisotropy in interfacial energy sigma, we extend the Cahn-Hilliard model by including in it a fourth-rank term, namely, gamma (ijlm) [partial derivative (2) c/(partial derivativex(i) partial derivativex(j))] [partial derivative (2) c/(partial derivativex(l) partial derivativex(m))]. This term leads to an additional linear term in the evolution equation for the composition parameter field. It also leads to an orientation-dependent effective fourth-rank coefficient gamma ([hkl]) in the governing equation for the one-dimensional composition profile across a planar interface. The main effect of a non-negative gamma ([hkl]) is to increase both sigma and interfacial width w, each of which, upon suitable scaling, is related to gamma ([hkl]) through a universal scaling function. In this model, sigma is a differentiable function of interface orientation (n) over cap, and does not exhibit cusps; therefore, the equilibrium particle shapes (Wulff shapes) do not contain planar facets. However, the anisotropy in the interfacial energy can be large enough to give rise to corners in the Wulff shapes in two dimensions. In particles of finite sizes, the corners become rounded, and their shapes tend towards the Wulff shape with increasing particle size.

Size-dependent and strain rate effects in quantum dot nanostructures with molecular dynamic simulations

Size and strain rate effects are among several factors which play an important role in determining the response of nanostructures, such as their deformations, to the mechanical loadings. The mechanical deformations in nanostructure systems at finite temperatures are intrinsically dynamic processes. Most of the recent works in this context have been focused on nanowires [1, 2], but very little attention has been paid to such low dimensional nanostructures as quantum dots (QDs). In this contribution, molecular dynamics (MD) simulations with an embedded atom potential method(EAM) are carried out to analyse the size and strain rate effects in the silicon (Si) QDs, as an example. We consider various geometries of QDs such as spherical, cylindrical and cubic. We choose Si QDs as an example due to their major applications in solar cells and biosensing. The analysis has also been focused on the variation in the deformation mechanisms with the size and strain rate for Si QD embedded in a matrix of SiO2 [3] (other cases include SiN and SiC matrices).It is observed that the mechanical properties are the functions of the QD size, shape and strain rate as it is in the case for nanowires [2]. We also present the comparative study resulted from the application of different EAM potentials in particular, the Stillinger-Weber (SW) potential, the Tersoff potentials and the environment-dependent interatomic potential (EDIP) [1]. Finally, based on the stabilized structural properties we compute electronic bandstructures of our nanostructures using an envelope function approach and its finite element implementation.