Spelling suggestions: "subject:"partition off units method"" "subject:"partition oof units method""
1 |
Shape and topology optimization with parametric level set method and partition of unity method. / CUHK electronic theses & dissertations collectionJanuary 2010 (has links)
First of all, the PDE form of the classical level set function phi is parameterized with an analytical form of Radial Basis Function (RBF), which is real-valued and continuously differentiable. Such that the upwind scheme, extension velocity and reinitialization algorithms in solving the discrete Hamilton-Jacobi equation can be waived in the numerical process, the whole framework is transformed into a standard mathematical programming problem in which the linear objective function can be directly optimized by a gradient algorithm - shape sensitivity. The minimization of the mean compliance is studied and presented to demonstrate the advantages of the parametrical method. / Parametrization substantially reduces the complexity of the original discrete PDE level set method. However, the result shows that the high number of RBF knots leads to dense coefficient matrices. Thus, it induces numerical instabilities, slow convergence and less accuracy in the process. Consequently, we then study the distribution of knots density for faster computation. By updating the movement of the knot, the knot moves towards the position where the change is directly determined by the shape sensitivity. In such case, we may use lesser number of knots to describe the properties of the system while the smoothness of the implicit function is satisfied. The sensitivity study is evaluated carefully and discussed in detail. Results show a significant improvement in the computational speed and stability. / The study found significant improvement obtained in the structural optimization with the parametric level set method, both the stability and efficiency were given as the benefits of using the method of the parametrization. / Traditional structural optimization approaches can be referred to as sizing optimization, since their design variables are the proportions of the structure or material. A major restriction in the sizing problem is that the shape and the topology of the structure are fixed a priori. Undoubtedly, changes in shape (e.g., curved boundary) and topology (e.g., holes in a member) could produce more significant improvement in dynamic performance than modifications in size alone. A recent development of shape and topology optimization based on the implicit moving boundaries with the use of the renowned level set method is regarded as one of the most sophisticated methods in handling the change of the structural topology. In this thesis, we study the parametrization of the classical level set method for the structural optimization and the associated computational methodology. / Usually, a large-scale model will lead to bulk coefficient matrices in the RBF optimization and the linear function normally require O (N3) flops and O (N2) memory while processing. It is becoming impractical to solve as N goes over 10,000. In fact, the dense system equation matrix frequently leads to the numerical instabilities and the failure of the optimization. Finally, we introduce the method of Partition of Unity (POU) to deal with this problem. POU is often used in 3D reconstruction of implicit surfaces from scattered point sets. It breaks the global domain into smaller overlapping subdomains such that the implicit functions can be more efficiently interpolated. Meanwhile, the global solution is obtained by blending all the local solutions with a set of weighting functions. The algorithm of POU is presented here, and we analyze and discuss the numerical results accordingly. / Ho, Hon Shan. / Adviser: Michael Y. Wang. / Source: Dissertation Abstracts International, Volume: 73-03, Section: B, page: . / Thesis (Ph.D.)--Chinese University of Hong Kong, 2010. / Includes bibliographical references (leaves 106-119). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [201-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstract also in Chinese.
|
2 |
Development of the partition of unity finite element method for the numerical simulation of interior sound field / Développement de la partition de l'unité méthode des éléments finis pour la simulation numérique de champ sonore intérieurYang, Mingming 29 June 2016 (has links)
Dans ce travail, nous avons introduit le concept sous-jacent de PUFEM et la formulation de base lié à l'équation de Helmholtz dans un domaine borné. Le processus d'enrichissement de l'onde plane de variables PUFEM a été montré et expliqué en détail. L'idée principale est d'inclure une connaissance a priori sur le comportement local de la solution dans l'espace des éléments finis en utilisant un ensemble de fonctions d'onde qui sont des solutions aux équations aux dérivées partielles. Dans cette étude, l'utilisation des ondes planes se propageant dans différentes directions a été favorisée car elle conduit à des algorithmes de calcul efficaces. En outre, nous avons montré que le nombre de directions d'ondes planes dépend de la taille de l'élément PUFEM et la fréquence des ondes à la fois en 2D et 3D. Les approches de sélection de ces ondes planes sont également illustrés. Pour les problèmes 3D, nous avons étudié deux systèmes de distribution des directions d'ondes planes qui sont la méthode du cube discrétisé et la méthode de la force de Coulomb. Il a été montré que celle-ci permet d'obtenir des directions d'onde espacées de façon uniforme et permet d'obtenir un nombre arbitraire d'ondes planes attachées à chaque noeud de l'élément de PUFEM, ce qui rend le procédé plus souple.Dans le chapitre 3, nous avons étudié la simulation numérique des ondes se propageant dans deux dimensions en utilisant PUFEM. La principale priorité de ce chapitre est de venir avec un schéma d'intégration exacte (EIS), résultant en un algorithme d'intégration rapide pour le calcul de matrices de coefficients de système avec une grande précision. L'élément 2D PUFEM a ensuite été utilisé pour résoudre un problème de transmission acoustique impliquant des matériaux poreux. Les résultats ont été vérifiés et validés par la comparaison avec des solutions analytiques. Les comparaisons entre le régime exact d'intégration (EIS) et en quadrature de Gauss ont montré le gain substantiel offert par l'EIE en termes de temps CPU.Une 3D exacte Schéma d'intégration a été présenté dans le chapitre 4, afin d'accélérer et de calculer avec précision (jusqu'à la précision de la machine) des intégrales très oscillatoires découlant des coefficients de la matrice de PUFEM associés à l'équation 3D Helmholtz. Grâce à des tests de convergence, un critère de sélection du nombre d'ondes planes a été proposé. Il a été montré que ce nombre ne pousse que quadratiquement avec la fréquence qui donne lieu à une réduction drastique du nombre total de degrés de libertés par rapport au FEM classique. Le procédé a été vérifié pour deux exemples numériques. Dans les deux cas, le procédé est représenté à converger vers la solution exacte. Pour le problème de la cavité avec une source de monopôle située à l'intérieur, nous avons testé deux modèles numériques pour évaluer leur performance relative. Dans ce scénario, où la solution exacte est singulière, le nombre de directions d'onde doit être choisie suffisamment élevée pour faire en sorte que les résultats ont convergé.Dans le dernier chapitre, nous avons étudié les performances numériques du PUFEM pour résoudre des champs sonores intérieurs 3D et des problèmes de transmission d'ondes dans lequel des matériaux absorbants sont présents. Dans le cas particulier d'un matériau réagissant localement modélisé par une impédance de surface. Un des critères d'estimation d'erreur numérique est proposé en considérant simplement une impédance purement imaginaire qui est connu pour produire des solutions à valeur réelle. Sur la base de cette estimation d'erreur, il a été démontré que le PUFEM peut parvenir à des solutions précises tout en conservant un coût de calcul très faible, et seulement environ 2 degrés de liberté par longueur d'onde ont été jugées suffisantes. Nous avons également étendu la PUFEM pour résoudre les problèmes de transmission des ondes entre l'air et un matériau poreux modélisé comme un fluide homogène équivalent. / In this work, we have introduced the underlying concept of PUFEM and the basic formulation related to the Helmholtz equation in a bounded domain. The plane wave enrichment process of PUFEM variables was shown and explained in detail. The main idea is to include a priori knowledge about the local behavior of the solution into the finite element space by using a set of wave functions that are solutions to the partial differential equations. In this study, the use of plane waves propagating in various directions was favored as it leads to efficient computing algorithms. In addition, we showed that the number of plane wave directions depends on the size of the PUFEM element and the wave frequency both in 2D and 3D. The selection approaches for these plane waves were also illustrated. For 3D problems, we have investigated two distribution schemes of plane wave directions which are the discretized cube method and the Coulomb force method. It has been shown that the latter allows to get uniformly spaced wave directions and enables us to acquire an arbitrary number of plane waves attached to each node of the PUFEM element, making the method more flexible.In Chapter 3, we investigated the numerical simulation of propagating waves in two dimensions using PUFEM. The main priority of this chapter is to come up with an Exact Integration Scheme (EIS), resulting in a fast integration algorithm for computing system coefficient matrices with high accuracy. The 2D PUFEM element was then employed to solve an acoustic transmission problem involving porous materials. Results have been verified and validated through the comparison with analytical solutions. Comparisons between the Exact Integration Scheme (EIS) and Gaussian quadrature showed the substantial gain offered by the EIS in terms of CPU time.A 3D Exact Integration Scheme was presented in Chapter 4, in order to accelerate and compute accurately (up to machine precision) of highly oscillatory integrals arising from the PUFEM matrix coefficients associated with the 3D Helmholtz equation. Through convergence tests, a criteria for selecting the number of plane waves was proposed. It was shown that this number only grows quadratically with the frequency thus giving rise to a drastic reduction in the total number of degrees of freedoms in comparison to classical FEM. The method has been verified for two numerical examples. In both cases, the method is shown to converge to the exact solution. For the cavity problem with a monopole source located inside, we tested two numerical models to assess their relative performance. In this scenario where the exact solution is singular, the number of wave directions has to be chosen sufficiently high to ensure that results have converged. In the last Chapter, we have investigated the numerical performances of the PUFEM for solving 3D interior sound fields and wave transmission problems in which absorbing materials are present. For the specific case of a locally reacting material modeled by a surface impedance. A numerical error estimation criteria is proposed by simply considering a purely imaginary impedance which is known to produce real-valued solutions. Based on this error estimate, it has been shown that the PUFEM can achieve accurate solutions while maintaining a very low computational cost, and only around 2 degrees of freedom per wavelength were found to be sufficient. We also extended the PUFEM for solving wave transmission problems between the air and a porous material modeled as an equivalent homogeneous fluid. A simple 1D problem was tested (standing wave tube) and the PUFEM solutions were found to be around 1% error which is sufficient for engineering purposes.
|
Page generated in 0.1051 seconds