FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2103306373> ?p ?o ?g. }

• W2103306373 abstract "Large scale wave propagation simulation is currently achievable in reasonable turnaround times by using distributed computing in multiple cpu clusters. However, if one needs to perform many such simulations, as is the case in optimization, tomography, or seismic imaging, then the resources required are still prohibitive. Model order reduction of large dynamical systems has been successfully used in several application domains to paliate that problem and in this paper we explore one of its manifestations, Proper Orthogonal Decomposition, for wave propagation. We describe the method and show how it can be easily interfaced with two different high fidelity simulators. We exemplify its use on several problems of increasing complexity and size. AMS subject classifications. 65M99 1. Introduction. There are many applications that require the repeated trans ient simu- lation of acoustic, elastic or electromagnetic wave propag ation. To name a few: structural analysis, blast on structures, vibrations of Navy vessels, sonar, design of piezoelectric trans- ducers for medical ultrasound, medical imaging and therapeutics uses of ultrasound, earth seismic imaging for the oil industry and earthquake seismology, optimization driven by sim- ulation for material identification and optimal design. As s uch, any significant improvement in the performance of numerical simulators would be very important. Model Order Reduction (MOR) refers to a collection of techniques to reduce the number of degrees of freedom of the very large scale dynamical systems that result after space dis- cretization of time-dependent partial differential equat ions. Some of these techniques have been successfully employed in the simulation of VLSI circuits, computational fluid mechan- ics, real-time control, heat conduction and other problems (1, 3, 5, 7). Not much has been done for wave propagation, although it does not seem that there are fundamental difficulties for its application (2). However, since none of these techniques are trivial to inter face with existing large scale high fidelity codes, it is important to be able to select wisel y the correct approach in order to minimize development costs. At this time we have centered our attention on the class of methods that go by the name of Proper Orthogonal Decomposition (POD). We start from the premise that it is possible to run a few full simulations with in the domain of interest. POD uses snapshots from these simulations to form an orthogonal basis for the solution space. This can be thought of as a problem-dependent modal decomposition, as opposite to the use of artificial basis functions (Fourier expansions, wavelet s). By using truncated Singular Value Decompositions it is possible to reduce even further the siz e of this basis without sacrificing accuracy and also to prevent the introduction of high freque ncy noise. The dynamic behav- ior of a new problem is calculated by solving projected collocation equations for the time dependent coefficients of a linear combination of the natura l basis functions. A different class of methods, tailored to problems where even a few high fidelity sim- ulations are not an option, is based on Krylov subspace machinery for large-scale matrix computations (5). These methods generate reduced-order models that are in a certain sense optimal, directly from the large-scale data matrices descr ibing the given linear system. In-" @default.
• W2103306373 created "2016-06-24" @default.
• W2103306373 creator A5031771708 @default.
• W2103306373 creator A5037112266 @default.
• W2103306373 date "2008-01-01" @default.
• W2103306373 modified "2023-09-27" @default.
• W2103306373 title "FAST WAVE PROPAGATION BY MODEL ORDER REDUCTION" @default.
• W2103306373 cites W1563685198 @default.
• W2103306373 cites W1968512837 @default.
• W2103306373 cites W1987166016 @default.
• W2103306373 cites W2022445566 @default.
• W2103306373 cites W2028213769 @default.
• W2103306373 cites W2054445592 @default.
• W2103306373 cites W2092398714 @default.
• W2103306373 cites W2161344224 @default.
• W2103306373 cites W2330012333 @default.
• W2103306373 hasPublicationYear "2008" @default.
• W2103306373 type Work @default.
• W2103306373 sameAs 2103306373 @default.
• W2103306373 citedByCount "3" @default.
• W2103306373 countsByYear W21033063732014 @default.
• W2103306373 countsByYear W21033063732021 @default.
• W2103306373 crossrefType "journal-article" @default.
• W2103306373 hasAuthorship W2103306373A5031771708 @default.
• W2103306373 hasAuthorship W2103306373A5037112266 @default.
• W2103306373 hasConcept C111335779 @default.
• W2103306373 hasConcept C113775141 @default.
• W2103306373 hasConcept C11413529 @default.
• W2103306373 hasConcept C118496448 @default.
• W2103306373 hasConcept C121332964 @default.
• W2103306373 hasConcept C2524010 @default.
• W2103306373 hasConcept C2779277453 @default.
• W2103306373 hasConcept C33923547 @default.
• W2103306373 hasConcept C41008148 @default.
• W2103306373 hasConcept C459310 @default.
• W2103306373 hasConcept C57493831 @default.
• W2103306373 hasConcept C79675319 @default.
• W2103306373 hasConcept C8058405 @default.
• W2103306373 hasConceptScore W2103306373C111335779 @default.
• W2103306373 hasConceptScore W2103306373C113775141 @default.
• W2103306373 hasConceptScore W2103306373C11413529 @default.
• W2103306373 hasConceptScore W2103306373C118496448 @default.
• W2103306373 hasConceptScore W2103306373C121332964 @default.
• W2103306373 hasConceptScore W2103306373C2524010 @default.
• W2103306373 hasConceptScore W2103306373C2779277453 @default.
• W2103306373 hasConceptScore W2103306373C33923547 @default.
• W2103306373 hasConceptScore W2103306373C41008148 @default.
• W2103306373 hasConceptScore W2103306373C459310 @default.
• W2103306373 hasConceptScore W2103306373C57493831 @default.
• W2103306373 hasConceptScore W2103306373C79675319 @default.
• W2103306373 hasConceptScore W2103306373C8058405 @default.
• W2103306373 hasLocation W21033063731 @default.
• W2103306373 hasOpenAccess W2103306373 @default.
• W2103306373 hasPrimaryLocation W21033063731 @default.
• W2103306373 hasRelatedWork W186089322 @default.
• W2103306373 hasRelatedWork W2003226931 @default.
• W2103306373 hasRelatedWork W2023545416 @default.
• W2103306373 hasRelatedWork W2027063973 @default.
• W2103306373 hasRelatedWork W2143171460 @default.
• W2103306373 hasRelatedWork W2181561723 @default.
• W2103306373 hasRelatedWork W2242305229 @default.
• W2103306373 hasRelatedWork W2786538540 @default.
• W2103306373 hasRelatedWork W2911174873 @default.
• W2103306373 hasRelatedWork W3029130126 @default.
• W2103306373 hasRelatedWork W3035079212 @default.
• W2103306373 hasRelatedWork W3099874806 @default.
• W2103306373 hasRelatedWork W3106711351 @default.
• W2103306373 hasRelatedWork W3169283880 @default.
• W2103306373 hasRelatedWork W3188467770 @default.
• W2103306373 hasRelatedWork W3212229994 @default.
• W2103306373 hasRelatedWork W329794753 @default.
• W2103306373 hasRelatedWork W630447320 @default.
• W2103306373 hasRelatedWork W786442925 @default.
• W2103306373 hasRelatedWork W1506760142 @default.
• W2103306373 isParatext "false" @default.
• W2103306373 isRetracted "false" @default.
• W2103306373 magId "2103306373" @default.
• W2103306373 workType "article" @default.