504 research outputs found
Carl Bode papers
No full textCarl Bode (1911-1993) was an author, professor of English at the University of Maryland, and officer of several literary and cultural organizations. His scholarly interests included Emerson, Thoreau, and H. L. Mencken. His books included Mencken, the first full biography to be published after Mencken's death; Maryland, a 350-year history of the state; The Man Behind You, a volume of poetry; and The Anatomy of American Popular Culture. He received a Ford Fellowship in 1952-1953 and a Guggenheim award in 1954-1955. Bode also founded the national American Studies Association. His papers consist of correspondence, drafts of publications, documentation from editing projects, and records of participation in political campaigns. Correspondence relates to Joseph Tydings, Walter R. Harding, and Wilson Follett. There is an unprocessed addendum to the collection, consisting of books on American literature and Maryland; correspondence; course materials; financial records; personnel-related materials; photographs; tapes; publications; and work papers
Bode Analysis of Uncertain Multivariable Systems
No full textBode plots are crucial for frequency domain analysis of SISO systems. The aim of this paper is to develop a complete approach for Bode plots of multivariable uncertain systems for both the magnitude and phase. The magnitude is based on the singular values. The phase is based on the phase spread of the numerical range. An IQC-based approach is pursued to provide both the magnitude and phase. A simulation example shows that the presented approach allows the generation of multivariable Bode plots of multivariable uncertain systems
Best Paper Award at ISAV 2023
No full textMathis Bode, Jens Henrik Göbbert, Jonathan Windgassen and their collaborators from Argonne National Laboratory (USA) have won the Best Paper Award for their paper “Scaling Computational Fluid Dynamics: In Situ Visualization of NekRS using SENSEI”. It was presented at the “ISAV 2023: In Situ Infrastructures for Enabling Extreme-scale Analysis and Visualization” workshop, which took place in conjunction with the SC23 on 13 November 2023 in Denver, Colorado, USA.The team describes in their paper a novel pipeline for in situ and in transit visualization and analysis utilizing SENSEI, ADIOS2, and ParaView over Python. The aim is to solve the dilemma having to choose between data accuracy or decreasing the resolution for Computational Fluid Dynamics on GPU-powered HPC systems. Their approach makes more regular data snapshots directly from memory and thus bypasses the pitfalls of checkpointing. The application NekRS is a GPU-centric thermal-fluid simulation, which showcases diverse in situ and in transit strategies. Experiments on the Polaris and JUWELS Booster supercomputers were conducted to demonstrate real-world implications, which offered crucial insights how efficient data management can be achieved without compromising accuracy
A conservative cell-based unsplit Volume of Fluid advection scheme for three-dimensional atomization simulations
No full textThe Volume of Fluid (VoF) method is widely used for capturing the interface motion in multiphase flow simulations. In particular, unsplit geometrical advection schemes have proved well-suited for flows with complex topologies. In the cell-based approach, the computational cell is followed along its Lagrangian trajectories and provides a conceptually simple framework for advancing the interface. Thus, this is a semi-Lagrangian method, and enforcing conservation is difficult. This paper presents a cell-based three-dimensional (3D) unsplit advection scheme that is conservative. The method relies on the Hybrid Lagrangian Eulerian Method (HyLEM) of Le Chenadec and Pitsch [3] but additionally ensures discrete conservation by introducing a correction of the projected cell, which is inspired by the 3D flux-based method of Owkes and Desjardins [14] and the two-dimensional cell-based method of Comminal et al. [4]. While the projected cell vertices are evaluated as in HyLEM, additional vertices are introduced to modify the projected cell faces. The positions of those are obtained from conservative flux volumes. The proposed method provides the same accuracy as the method of Owkes and Desjardins [14], but is more efficient. The proposed method is tested in various benchmark cases and applied in an atomization case.</p
Numerical modeling of single droplet flash boiling behavior of e-fuels considering internal and external vaporization
No full textIn recent years, significant efforts have been made to develop e-fuels from renewable electricity and carbon sources for enabling highly efficient and advanced propulsion systems. Compared to conventional fuels, such fuels can have very different thermo-physical properties depending on their molecular structure. Particularly, fuels with high vapor pressures are highly susceptible to flash boiling depending on boundary conditions, which can significantly alter the spray formation and mixing behavior. Thus, it becomes imperative to develop a fundamental understanding of the underlying physics associated with the flash boiling of these fuels in a single droplet configuration. In this work, oxymethylene ethers (OMEx) are chosen as a generic example to study the flashing behavior of newly developed e-fuels. This study employs the Lagrangian Particle Tracking (LPT) approach considering both internal and external vaporization of flash boiling single droplets. The internal vaporization model includes several sub-models that compute bubble number density, bubble growth rate, and droplet bursting criterion. External vaporization is modeled considering heat transfer from the droplet interior to the droplet surface and from the surrounding gas to the droplet surface. The study reveals that the formation and subsequent growth of vapor bubble nuclei is the primary source causing the transition of the metastable liquid phase into the stable state. We found that for moderate to high superheating degree, the bubble growth characteristics indicate three distinct growth phases: (1) surface tension-controlled, (2) transition, and (3) inertia-controlled, whereas, for low superheating degree, only two of these were present, namely (1) surface tension-controlled, and (2) transition phase. It is also observed that the chain length of OMEx has significant impact on bubble dynamics. OME4 is found to have a larger critical nucleus, a longer time delay in bubble growth, and a slower growth rate compared with dimethyl ether (DME). Furthermore, a quantitative analysis shows that droplets burst earlier with increasing superheating degree. In addition, it is found that the system pressure has a negligible influence on the initiation of the bursting process, except when the superheating degree is very low
Influence of adversarial training on super-resolution turbulence reconstruction
No full textSupervised super-resolution deep convolutional neural networks (CNNs) have
gained significant attention for their potential in reconstructing velocity and
scalar fields in turbulent flows. Despite their popularity, CNNs currently lack
the ability to accurately produce high-frequency and small-scale features, and
tests of their generalizability to out-of-sample flows are not widespread.
Generative adversarial networks (GANs), which consist of two distinct neural
networks (NNs), a generator and discriminator, are a promising alternative,
allowing for both semi-supervised and unsupervised training. The difference in
the flow fields produced by these two NN architectures has not been thoroughly
investigated, and a comprehensive understanding of the discriminator's role has
yet to be developed. This study assesses the effectiveness of the unsupervised
adversarial training in GANs for turbulence reconstruction in forced
homogeneous isotropic turbulence. GAN-based architectures are found to
outperform supervised CNNs for turbulent flow reconstruction for in-sample
cases. The reconstruction accuracy of both architectures diminishes for
out-of-sample cases, though the GAN's discriminator network significantly
improves the generator's out-of-sample robustness using either an additional
unsupervised training step with large eddy simulation input fields and a
dynamic selection of the most suitable upsampling factor. These enhance the
generator's ability to reconstruct small-scale gradients, turbulence
intermittency, and velocity-gradient probability density functions. The
extrapolation capability of the GAN-based model is demonstrated for
out-of-sample flows at higher Reynolds numbers. Based on these findings,
incorporating discriminator-based training is recommended to enhance the
reconstruction capability of super-resolution CNNs
Fast Identification of Bound Structures in Large N-body Simulations
No full textWe present an algorithm that is designed to allow the efficient identification and preliminary dynamical analysis of thousands of structures and substructures in large N-body simulations. First, we utilize a refined density gradient system (based on denmax) to identify the structures and then apply an iterative approximate method to identify unbound particles, allowing fast calculation of bound substructures. After producing a catalogue of separate energetically bound substructures, we check to see which of these are energetically bound to adjacent substructures. For such bound complex subhaloes, we combine components and check if additional free particles are also bound to the union, repeating the process iteratively until no further changes are found. Thus, our subhaloes can contain more than one density maximum, but the scheme is stable: starting with a small smoothing length initially produces small structures that must be combined later and starting with a large smoothing length produces large structures within which sub-substructure is found. We apply this algorithm to three simulations. Two that are using the TPM algorithm by Bode, Ostriker & Xu and one on a simulated halo by Diemand, Moore & Stadel. For all these haloes, we find about 5–8 per cent of the mass in substructures
Applying Physics-Informed Enhanced Super-Resolution Generative Adversarial Networks to Large-Eddy Simulations of ECN Spray C
No full textLarge-eddy simulation (LES) is an important tool to understand and analyze sprays, such as those found in engines. Subfilter models are crucial for the accuracy of spray-LES, thereby signifying the importance of their development for predictive spray-LES. Recently, new subfilter models based on physics-informed generative adversarial networks (GANs) were developed, known as physics-informed enhanced super-resolution GANs (PIESRGANs). These models were successfully applied to the Spray A case defined by the Engine Combustion Network (ECN). This work presents technical details of this novel method, which are relevant for the modeling of spray combustion, and applies PIESRGANs to the ECN Spray C case. The results are validated against experimental data, and computational challenges and advantages are particularly emphasized compared to classical simulation approaches
Book review: Omojola, Bode (ed.) (2017). Music and Social Dynamics in Nigeria. Religion and Society in Africa, Volume 3.
No full textBook title: Music and Social Dynamics in Nigeria. Religion and Society in Africa, Volume 3.
Author: Bode Omojola (ed.)
New York: Peter Lang. ISBN (print): 978-143-3134-01-2, ISBN (e-book): 978-145-3918-52–4. viii, 226 pp. <https://www.peterlang.com/view/title/23171> $102.85
A heterogeneidade enunciativa em memes do “Bode Gaiato”
No full textThis work aims to analyze forms of heterogeneity shown in Bode Gaiato memes. We are mainly based on studies by Jacqueline Authier-Revuz (1982, 1990), in which the author, based on conceptions of decentralization of the subject of Freudian-Lacan psychoanalysis and Bakhtin’s notion of dialogism, discusses the effects of presence of several voices in the discourse of a speaker, showing how every discourse is crossed by multiple voices, so that one can not believe in the existence of a homogeneous discourse. For this, after reviewing theoretical studies of Authier-Revuz (1982, 1990), Bakhtin (2002), Benveniste (1989), Dunker (2010) and Flores; Teixeira (2005), we have collected 17 texts published in the profile of Bode Gaiato in one of the most accessed social networks over the internet, Facebook. These texts were then analyzed in the light of the aforementioned theorists. It was then verified that there is a very significant interlacing of social voices for the construction of the saying in the analyzed memes, so that to recognize them, referring them to the discourses they evoke, is relevant to the understanding of these texts. It is thus proved that every discourse is intrinsically heterogeneous and that the subject does not constitute himself alone.Este trabalho tem como objetivo analisar formas de heterogeneidade mostrada em memes do Bode Gaiato. Fundamentamo-nos, sobretudo, nos estudos de Jacqueline Authier-Revuz (1982, 1990), nos quais a autora, apoiando-se em concepções de descentramento do sujeito da psicanálise freudo-lacaniana e na noção de dialogismo de Bakhtin, discute os efeitos da presença de diversas vozes no discurso de um locutor, mostrando como todo discurso é atravessado por múltiplas vozes, de modo que não se pode acreditar na existência de discurso homogêneo. Para tanto, após a revisão de estudos teóricos de Authier-Revuz (1982, 1990), Bakhtin (2002), Benveniste (1989), Dunker (2010) e Flores; Teixeira (2005), coletamos 17 textos publicados no perfil do Bode Gaiato, numa das mais acessadas redes sociais, o Facebook. Esses textos foram, então, analisados à luz dos teóricos supracitados. Foi constatado que há, nos memes analisados, um entrecruzamento de vozes sociais bastante significativo para a construção do dizer, de modo que reconhecer essas vozes, remetendo-as aos discursos que evocam, é relevante para a compreensão desses textos. Comprova-se, assim, que todo discurso é intrinsecamente heterogêneo e que o sujeito não se constitui sozinho
- …
