Ruiz, E.; Roca, X.; Sarrate, J.; Montenegro, R.; Escobar, J.M. Advances in engineering software Vol. 80, p. 12-24 DOI: 10.1016/j.advengsoft.2014.09.021 Data de publicació: 2015-02-01 Article en revista
In this work, we present a simultaneous untangling and smoothing technique for quadrilateral and hexahedral meshes. The algorithm iteratively improves a quadrilateral or hexahedral mesh by minimizing an objective function defined in terms of a regularized algebraic distortion measure of the elements. We propose several techniques to improve the robustness and the computational efficiency of the optimization algorithm. In addition, we have adopted an object-oriented paradigm to create a common framework to smooth meshes composed by any type of elements, and using different minimization techniques. Finally, we present several examples to show that the proposed technique obtains valid meshes composed by high-quality quadrilaterals and hexahedra, even when the initial meshes contain a large number of tangled elements.
In electric circuit theory, it is of great interest to compute the effective resistance between any pair of vertices of a network, as well as the Kirchhoff Index. During the last decade these parametres have been applied in Organic Chemistry as natural structural indexes different from the usual ones in order to achieve an improvement in the discrimination between different molecules which have similar structural behaviours. Moreover, a wide range of generalized Kirchhoff Indexes for some networks have been introduced. The objective of the present work is to obtain the Kirchhoff Index for a wide range of composite networks known as cluster networks
In electric circuit theory, it is of great interest to compute the effective resistance between any pair of vertices of a network, as well as the Kirchhoff Index. During the last decade these parametres have been applied in Organic Chemistry as natural structural indexes different from the usual ones in order to achieve an improvement in the discrimination between different molecules which have similar structural behaviours. Moreover, a wide range of generalized Kirchhoff Indexes for some networks have been introduced. The objective of the present work is to obtain the Kirchhoff Index for a wide range of composite networks known as cluster networks.
The multi-sweeping method is one of the most used algorithms to generate hexahedral meshes for extrusion volumes. In this method the geometry is decomposed in sub-volumes by means of projecting nodes along the sweep direction and imprinting faces. However, the quality of the final mesh is determined by the location of inner nodes created during the decomposition process and by the robustness of the imprinting process.
In this work we present two original contributions to increase the quality of the decomposition process. On the one hand, to improve the robustness of the imprints we introduce the new concept of computational domain for extrusion geometries. Since the computational domain is a planar representation of the sweep levels, we improve several geometric operations involved in the imprinting process. On the other hand, we propose a three-stage procedure to improve the location of the inner nodes created during the decomposition process. First, inner nodes are projected towards source surfaces. Second, the nodes are projected back towards target surfaces. Third, the final position of inner nodes is computed as a weighted average of the projections from source and target surfaces.
The submapping method is one of the most widely used techniques to generate structured quadrilateral meshes. This method splits the geometry into pieces logically equivalent to a quadrilateral. Then, it meshes each piece keeping the mesh compatibility between them by solving an integer linear problem. The main limitation of submapping algorithms is that it can only be applied to geometries in which the angle between two consecutive edges is, approximately, an integer multiple of p=2. In addition, special procedures are required in order to apply it to multiply connected domains. This article presents two original modifications to mitigate these shortcomings. Finally, it presents several numerical examples that show the applicability of the developed algorithms.
A simple explicit, characteristic-based finite element method for the numerical simulation of the dispersion of the thermal outflow in Huelva estuary is considered in this paper. The derivation involves a local Taylor expansion of the convection-diffusion equation. The numerical model used was originally designed to simulate the dispersion of pollutants in the open sea and has now been reconfigured to simulate the dispersion of the thermal outflow of some power plants which may be constructed in Huelva estuary in the near future. Numerical results obtained are presented. (C) 2006 Elsevier Ltd. and Civil-Comp Ltd. All rights reserved.
Many environmental processes can be modelled as transient convection–diffusion–reaction problems. This is the case, for instance, of the operation of activated-carbon filters. For industrial applications there is a growing demand for 3D simulations, so efficient linear solvers are a major concern. We have compared the numerical performance of two families of incomplete Cholesky factorizations as preconditioners of conjugate gradient iterations: drop-tolerance and prescribed-memory strategies. Numerical examples show that the former are computationally more efficient, but the latter may be preferable due to their predictable memory requirements.