We present high-fidelity reacting simulations of a supersonic cavity flame-holder configuration. The focus of this work is on flame stabilization brought about by varying the location of fuel ...injection in a cavity stabilized supersonic flow of air. Central to our approach is a compressible multi-species reacting flow solver that uses adaptive-mesh-refinement (AMR), enabling the resolution of flame, shock-waves, boundary-layers, and small-scale structures in the computational domain. Our analysis indicates that fuel injection closer to the ramp at the aft end of the cavity allows for greater mixing and lower peak temperatures compared to fuel injection upstream that is closer to the backward facing step of the cavity. This difference is mainly due to greater turbulent fluctuations generated from the shear-layer towards the cavity ramp, thereby enhancing the mixing of fuel and air. A low frequency oscillatory behaviour in heat-release and pressure was also observed for the upstream injection case while a much higher-frequency phenomena was observed in the near-ramp injection case. By identifying the important physical determinants of the combustion processes, this study illustrates a promising pathway to design and optimize direct fuel injection strategies in supersonic cavity flame-holders that can improve flame stability, combustion efficiency, and reduce emissions.
In this work, we use machine learning (ML) techniques to develop presumed probability density function (PDF) models for large eddy simulations (LES) of reacting flows. The joint sub-filter PDF of ...mixture fraction and progress variable is modeled using various ML algorithms and commonly used analytical models. The ML algorithms evaluated in the work are representative of three major classes of ML techniques: traditional ensemble methods (random forests), deep learning (deep neural network (DNN)s), and generative learning (conditional variational autoencoder (CVAE)). The first two algorithms are supervised learning algorithms, and the third is an unsupervised learning algorithm. Data from direct numerical simulation (DNS) of the low-swirl burner1 are used to develop training data for sub-filter PDF models. Models are evaluated on predictions of the sub-filter PDF as well as predictions of the filtered reaction rate of the progress variable, computed through an integral of the product of the sub-filter PDF and the conditional means of the reaction rate. This a-priori modeling study demonstrates that deep learning models for presumed PDF modeling are three times more accurate than analytical β–β PDF models and linear regression models. These models are as accurate as random forest models while using five times fewer trainable parameters and being 25 times faster for inference. In this work, conditional unsupervised learning did not present additional advantages beyond supervised learning with a feed-forward neural network. We illustrate how models generalize to other regions of the flow and develop criteria based on the Jensen–Shannon divergence to quantify the performance of a model on new data.
Thermodynamics of Life Henry, Marc
Substantia (Firenze),
03/2021, Letnik:
5, Številka:
1
Journal Article
Recenzirano
Odprti dostop
Biology is currently plagued by several fossil concepts that may be responsible for the current stagnation in medicine. Through a careful screening of the origins of thermodynamics, such fossils ...concepts have been identified: assumption that heat is a form of energy, assimilation of entropy to disorder, assimilation of death to states of maximum entropy, assimilation of ATP to the energy currency of living cells, non-recognition of entropy as a state function of the whole universe, belief that free energies are another kind of energy, self-referencing in the definition of life, ignorance of basic principles of quantum physics and more particularly of the importance of intrinsic spin, confusion between three different forms of reversibility, non-recognition that irreversibility is at the heart of living systems. After stowing of these concepts in the cabinet of useless and nasty notions, a fresh new look is proposed showing how life is deep-rooted trough the entropy concept in quantum physics on the one hand and in cosmology on the other hand. This suggests that life is not an emergent property of matter, but rather that it has always been a fundamental property of a universe filled with particles and fields. It is further proposed to dismiss the first (energy = heat + work) and third laws (entropy decreases to zero at zero Kelvin) of thermodynamics, retaining only the clear Boltzmann's definition of entropy in terms of multiplicity of microstates Ω, S = kB×Ln Ω, and the second law in its most general form applicable to any kind of macrostates: ∆Suniv ≥ 0. On this ground, clear definitions are proposed for life/death, healthiness/illness and for thermodynamic coupling. The whole unfolding of life in the universe: Big Bang → Light → Hydrogen → Stars → Atoms → Water → Planets → Metabolism → Lipids → RNA's → Viruses → Ribosome → Proteins → Bacteria → Eukaryote → Sex → Plants → Animals → Humans → Computers → Internet, may then be interpreted as a simple consequence of a single principle: ∆Suniv ≥ 0. We thus strongly urge biologists and physicians to change and adapt their ideas and vocabulary to the proposed reformulation for a better understanding of what is life and as a consequence for better health for living beings.
There are very few things that modern science does not yet understand. One of them is consciousness; the other is water. Our main idea was then that if consciousness and water are still deep ...mysteries for science, may be it could be because these apparently two different problems are in fact deeply entangled. Shading light on one of them may thus well have the wanted effect of illuminating the other. Here we explore the idea that it exists a mirror relationship between an immaterial pair acting on an immaterial world formed by consciousness and information on the one hand and a pair acting on a material world formed by electromagnetic radiations and water on the other hand. By using group theoretical arguments applied to the famous Maxwell’s equations, it is formally deduced that the external visible world les not a 4D space-time continuum (named M4 hereafter), but rather a 5D-space-time-scale hyper-surface (named C5 hereafter) embedded in the 6D-continuum of consciousness (named V6 hereafter) identified with the vacuum state of quantum physics or the ether (dynamical background) of general relativity. The new fifth degree of freedom in C5 is associated to the ability of living things to grow from small sizes to large sizes by leaving invariant its identity at all scales. The sixth degree of freedom in V6 is associated to the possibility of living either as a virtual non-observable entity or as a non-virtual observable entity. In both cases, life is associated to the ability of managing information stored in the intrinsic quantum structure of the V6-ether or in the water shells surrounding any living cell in the C5-hypersurface. Memory capacities and associated bandwidths may be quantitatively evaluated directly from theory and compared to what is observed experimentally giving confidence in the proposed approach. It follows directly from the modeling that space, time and mass should be considered as creations of consciousness in the form of persistent bits field, bringing strong support to eastern ways of philosophical thinking based were Tao or Vacuity are the only non-dual realities. For western minds, the modeling has the great advantage of explaining clearly what is life and consciousness in a mathematical frame unifying physics, chemistry and biology.
Despite major recent advances in C-H activation, discrimination between two similar, unactivated C-H positions is beyond the scope of current chemocatalytic methods. Here we demonstrate that ...integration of regioselective halogenase enzymes with Pd-catalysed cross-coupling chemistry, in one-pot reactions, successfully addresses this problem for the indole heterocycle. The resultant 'chemobio-transformation' delivers a range of functionally diverse arylated products that are impossible to access using separate enzymatic or chemocatalytic C-H activation, under mild, aqueous conditions. This use of different biocatalysts to select different C-H positions contrasts with the prevailing substrate-control approach to the area, and presents opportunities for new pathways in C-H activation chemistry. The issues of enzyme and transition metal compatibility are overcome through membrane compartmentalization, with the optimized process requiring no intermediate work-up or purification steps.
Once upon a time, there was a Russian guy named Dmitri Mendeleev (1834-1907), who became Doctor of Science in 1865 at Saint Petersburg State University for his dissertation «On the Combinations of ...Water with Alcohol», a most important mixture in eastern European countries. After becoming a teacher in this university (1867), he planned to prepare a book for his teaching and after a dream, he has envisioned the complete arrangement of the elements as a 2D-table based on the “magic” integer 8. On 6 March 1869, he made a formal presentation to the Russian Chemical Society, describing elements according to both atomic weight and valence. The role of the magic number 8 was elucidated between 1923 and 1930 through the development of quantum mechanics and new magic electron count numbers were discovered: 2 (He), 10 (Ne), 18 (Ar), 36 (Kr), 54 (Xe), 86 (Rn) and 118 (Uuo). In October 1957, a most important paper was published showing that the three most abundant elements in the universe was in that order: H, He and O.
La conception de l’enquête qu’Éric Chauvier déploie dans son oeuvre, c’est tout au moins l’idée que j’essaie de défendre dans cet article, n’est pas tant située quelque part entre anthropologie et ...littérature, qu’elle caractérise une forme d’anthropologie qui s’appuie sur la littérature, à la fois comme fiction et comme récit, parce qu’elle est habitée par une recherche continuelle sur la façon d’approcher ce qui échappe constamment à l’énonciation, l’altérité. Éric Chauvier a besoin de mobiliser d’autres méthodes que les formules standardisées et les protocoles de recherche savants en raison de leur impuissance à saisir l’expérience humaine dans sa polyphonie et dans son inconstance, dans sa fragilité et sa précarité aussi quand l’enquête s’attaque à rendre compte de « vies minuscules ».
Keywords : Enquête fictionnelle, épistémologie de l’ordinaire, dispositifs littéraires, exigence de réel, vies minuscules
Intitulé « Savoirs pluriels. Enjeux épistémologiques et politiques », ce dossier de SociologieS vise à contribuer aux débats qui se poursuivront à l’occasion de la tenue du XXIIe Congrès de ...l’Association internationale des sociologues de langue française (AISLF), à Ottawa en juillet 2024, sur le thème « Sciences, savoirs et sociétés ». Il interroge les enjeux d’actualité entourant la multiplication des acteurs et des espaces émergents de production des connaissances sur le social, les nouvel...
PeleC: An adaptive mesh refinement solver for compressible reacting flows Henry de Frahan, Marc T; Rood, Jon S; Day, Marc S ...
International journal of high performance computing applications/The international journal of high performance computing applications,
03/2023, Letnik:
37, Številka:
2
Journal Article
Recenzirano
Odprti dostop
Reacting flow simulations for combustion applications require extensive computing capabilities. Leveraging the AMReX library, the Pele suite of combustion simulation tools targets the largest ...supercomputers available and future exascale machines. We introduce PeleC, the compressible solver in the Pele suite, and detail its capabilities, including complex geometry representation, chemistry integration, and discretization. We present a comparison of development efforts using both OpenACC and AMReX’s C++ performance portability framework for execution on multiple GPU architectures. We discuss relevant details that have allowed PeleC to achieve high performance and scalability. PeleC’s performance characteristics are measured through relevant simulations on multiple supercomputers. The success of PeleC’s design for exascale is exhibited through demonstration of a 160 billion cell simulation and weak scaling onto 100% of Summit, an NVIDIA-based GPU supercomputer at Oak Ridge National Laboratory. Our results provide confidence that PeleC will enable future combustion science simulations with unprecedented fidelity.
Although the Discontinuous Galerkin (dg) method has seen widespread use for compressible flow problems in a single fluid with constant material properties, it has yet to be implemented in a ...consistent fashion for compressible multiphase flows with shocks and interfaces. Specifically, it is challenging to design a scheme that meets the following requirements: conservation, high-order accuracy in smooth regions and non-oscillatory behavior at discontinuities (in particular, material interfaces). Following the interface-capturing approach of Abgrall 1, we model flows of multiple fluid components or phases using a single equation of state with variable material properties; discontinuities in these properties correspond to interfaces. To represent compressible phenomena in solids, liquids, and gases, we present our analysis for equations of state belonging to the Mie–Grüneisen family. Within the dg framework, we propose a conservative, high-order accurate, and non-oscillatory limiting procedure, verified with simple multifluid and multiphase problems. We show analytically that two key elements are required to prevent spurious pressure oscillations at interfaces and maintain conservation: (i) the transport equation(s) describing the material properties must be solved in a non-conservative weak form, and (ii) the suitable variables must be limited (density, momentum, pressure, and appropriate properties entering the equation of state), coupled with a consistent reconstruction of the energy. Further, we introduce a physics-based discontinuity sensor to apply limiting in a solution-adaptive fashion. We verify this approach with one- and two-dimensional problems with shocks and interfaces, including high pressure and density ratios, for fluids obeying different equations of state to illustrate the robustness and versatility of the method. The algorithm is implemented on parallel graphics processing units (gpu) to achieve high speedup.