Massively Parallel Computational Fluid Dynamics With Large Eddy Simulation in Complex Geometries

2011 ◽  
Author(s):  
Andrew Duggleby ◽  
Joshua L. Camp ◽  
Yuval Doron ◽  
Paul F. Fischer
Keyword(s):  
Computer Aided Design ◽  
High Performance ◽  
Complex Geometry ◽  
Spectral Element ◽  
Computational Time ◽  
Model And Simulation ◽  
Large Eddy ◽  
Order Of Magnitude ◽  
Grid Points ◽  
Pipe Geometry

To perform complex geometry large eddy simulations in an industrially relevant timeframe, one must reduce the total time to half a day (overnight simulation). Total time includes the time of developing the mesh from the computer-aided design (CAD) model and simulation time. For reducing CAD-to-mesh time, automatic meshing algorithms can generate valid but often non-efficient meshes with often up to an order of magnitude more grid points than a custom-based mesh. These algorithms are acceptable only if paired with high-performance computing (HPC) platforms comprising thousands to millions of cores to significantly reduce computational time. Efficient use of these tools calls for codes that can scale to high processor counts and that can efficiently transport resolved scales over the long distances and times made feasible by HPC. The rapid convergence of high-order discretizations makes them particularly attractive in this context. In this paper we test the combination of automatic hexahedral meshing with a spectral element code for incompressible and low-Mach-number flows, called Nek5000, that has scaled to P >262,000 cores and sustains >70% parallel efficiency with only ≈7000 points/core. For our tests, a simple pipe geometry is used as a basis for comparing with previous fully resolved direct numerical simulations.

NUMERICAL STUDY OF DETONATION PROPAGATION IN VISCOUS TURBULENT FLOW IN A DUCT

2020 ◽  
Author(s):  
V. A. SABELNIKOV ◽  
◽  
V. V. VLASENKO ◽  
S. BAKHNE ◽  
S. S. MOLEV ◽  
...  
Keyword(s):  
Complex Geometry ◽  
Numerical Study ◽  
Navier Stokes ◽  
Detonation Waves ◽  
Detonation Propagation ◽  
Eddy Simulation ◽  
Detonation Structure ◽  
Classical Studies ◽  
Large Eddy ◽  
Physical Effects

Gasdynamics of detonation waves was widely studied within last hundred years - analytically, experimentally, and numerically. The majority of classical studies of the XX century were concentrated on inviscid aspects of detonation structure and propagation. There was a widespread opinion that detonation is such a fast phenomenon that viscous e¨ects should have insigni¦cant in§uence on its propagation. When the era of calculations based on the Reynolds-averaged Navier- Stokes (RANS) and large eddy simulation approaches came into effect, researchers pounced on practical problems with complex geometry and with the interaction of many physical effects. There is only a limited number of works studying the in§uence of viscosity on detonation propagation in supersonic §ows in ducts (i. e., in the presence of boundary layers).

scholarly journals Computational Complexity Reduction of Neural Networks of Brain Tumor Image Segmentation by Introducing Fermi–Dirac Correction Functions

Entropy ◽  
2021 ◽  
Vol 23 (2) ◽  
pp. 223
Author(s):  
Yen-Ling Tai ◽  
Shin-Jhe Huang ◽  
Chien-Chang Chen ◽  
Henry Horng-Shing Lu
Keyword(s):  
Neural Networks ◽  
Deep Learning ◽  
Computational Complexity ◽  
High Performance ◽  
Low Cost ◽  
Structural Complexity ◽  
Correction Function ◽  
Computational Time ◽  
Learning Methods ◽  
Band Theory

Nowadays, deep learning methods with high structural complexity and flexibility inevitably lean on the computational capability of the hardware. A platform with high-performance GPUs and large amounts of memory could support neural networks having large numbers of layers and kernels. However, naively pursuing high-cost hardware would probably drag the technical development of deep learning methods. In the article, we thus establish a new preprocessing method to reduce the computational complexity of the neural networks. Inspired by the band theory of solids in physics, we map the image space into a noninteraction physical system isomorphically and then treat image voxels as particle-like clusters. Then, we reconstruct the Fermi–Dirac distribution to be a correction function for the normalization of the voxel intensity and as a filter of insignificant cluster components. The filtered clusters at the circumstance can delineate the morphological heterogeneity of the image voxels. We used the BraTS 2019 datasets and the dimensional fusion U-net for the algorithmic validation, and the proposed Fermi–Dirac correction function exhibited comparable performance to other employed preprocessing methods. By comparing to the conventional z-score normalization function and the Gamma correction function, the proposed algorithm can save at least 38% of computational time cost under a low-cost hardware architecture. Even though the correction function of global histogram equalization has the lowest computational time among the employed correction functions, the proposed Fermi–Dirac correction function exhibits better capabilities of image augmentation and segmentation.

scholarly journals A Low Mach Number IMEX Flux Splitting for the Level Set Ghost Fluid Method

2021 ◽  
Author(s):  
Jonas Zeifang ◽  
Andrea Beck
Keyword(s):  
Level Set ◽  
Phase Interface ◽  
Time Discretization ◽  
Spectral Element ◽  
Computational Time ◽  
Minor Role ◽  
Ghost Fluid Method ◽  
Specific Level ◽  
Flux Splitting ◽  
Mach Numbers

AbstractConsidering droplet phenomena at low Mach numbers, large differences in the magnitude of the occurring characteristic waves are presented. As acoustic phenomena often play a minor role in such applications, classical explicit schemes which resolve these waves suffer from a very restrictive timestep restriction. In this work, a novel scheme based on a specific level set ghost fluid method and an implicit-explicit (IMEX) flux splitting is proposed to overcome this timestep restriction. A fully implicit narrow band around the sharp phase interface is combined with a splitting of the convective and acoustic phenomena away from the interface. In this part of the domain, the IMEX Runge-Kutta time discretization and the high order discontinuous Galerkin spectral element method are applied to achieve high accuracies in the bulk phases. It is shown that for low Mach numbers a significant gain in computational time can be achieved compared to a fully explicit method. Applications to typical droplet dynamic phenomena validate the proposed method and illustrate its capabilities.

scholarly journals A Modified KNN Algorithm for High-Performance Computing on FPGA of Real-Time m-QAM Demodulators

Electronics ◽  
2021 ◽  
Vol 10 (5) ◽  
pp. 627
Author(s):  
David Marquez-Viloria ◽  
Luis Castano-Londono ◽  
Neil Guerrero-Gonzalez
Keyword(s):  
Real Time ◽  
High Performance ◽  
Interference Mitigation ◽  
Parallel Implementation ◽  
Computational Time ◽  
Successful Implementation ◽  
Interchannel Interference ◽  
The Difference ◽  
High Level ◽  
Performance Computing

A methodology for scalable and concurrent real-time implementation of highly recurrent algorithms is presented and experimentally validated using the AWS-FPGA. This paper presents a parallel implementation of a KNN algorithm focused on the m-QAM demodulators using high-level synthesis for fast prototyping, parameterization, and scalability of the design. The proposed design shows the successful implementation of the KNN algorithm for interchannel interference mitigation in a 3 × 16 Gbaud 16-QAM Nyquist WDM system. Additionally, we present a modified version of the KNN algorithm in which comparisons among data symbols are reduced by identifying the closest neighbor using the rule of the 8-connected clusters used for image processing. Real-time implementation of the modified KNN on a Xilinx Virtex UltraScale+ VU9P AWS-FPGA board was compared with the results obtained in previous work using the same data from the same experimental setup but offline DSP using Matlab. The results show that the difference is negligible below FEC limit. Additionally, the modified KNN shows a reduction of operations from 43 percent to 75 percent, depending on the symbol’s position in the constellation, achieving a reduction 47.25% reduction in total computational time for 100 K input symbols processed on 20 parallel cores compared to the KNN algorithm.

scholarly journals A Peristaltic Micropump Based on the Fast Electrochemical Actuator: Design, Fabrication, and Preliminary Testing

Actuators ◽  
2021 ◽  
Vol 10 (3) ◽  
pp. 62
Author(s):  
Ilia Uvarov ◽  
Pavel Shlepakov ◽  
Artem Melenev ◽  
Kechun Ma ◽  
Vitaly Svetovoy ◽  
...  
Keyword(s):  
High Performance ◽  
Drug Delivery Systems ◽  
Main Part ◽  
Fabrication Technology ◽  
Fabrication Procedure ◽  
Order Of Magnitude ◽  
Peristaltic Micropump ◽  
Biomedical Field ◽  
Lift Off ◽  
Actuator Design

Microfluidic devices providing an accurate delivery of fluids at required rates are of considerable interest, especially for the biomedical field. The progress is limited by the lack of micropumps, which are compact, have high performance, and are compatible with standard microfabrication. This paper describes a micropump based on a new driving principle. The pump contains three membrane actuators operating peristaltically. The actuators are driven by nanobubbles of hydrogen and oxygen, which are generated in the chamber by a series of short voltage pulses of alternating polarity applied to the electrodes. This process guaranties the response time of the actuators to be much shorter than that of any other electrochemical device. The main part of the pump has a size of about 3 mm, which is an order of magnitude smaller in comparison with conventional micropumps. The pump is fabricated in glass and silicon wafers using standard cleanroom processes. The channels are formed in SU-8 photoresist and the membrane is made of SiNx. The channels are sealed by two processes of bonding between SU-8 and SiNx. Functionality of the channels and membranes is demonstrated. A defect of electrodes related to the lift-off fabrication procedure did not allow a demonstration of the pumping process although a flow rate of 1.5 µl/min and dosage accuracy of 0.25 nl are expected. The working characteristics of the pump make it attractive for the use in portable drug delivery systems, but the fabrication technology must be improved.

scholarly journals The Implementation of Spectral Element Method in a CAE System for the Solution of Elasticity Problems on Hybrid Curvilinear Meshes

2017 ◽  
Vol 2017 ◽  
pp. 1-7 ◽  
Author(s):  
Dmitriy Konovalov ◽  
Anatoly Vershinin ◽  
Konstantin Zingerman ◽  
Vladimir Levin
Keyword(s):  
Mathematical Simulation ◽  
High Performance ◽  
Spectral Element Method ◽  
New Technologies ◽  
Spectral Element ◽  
High Tech ◽  
Engineering Analysis ◽  
Elasticity Problems ◽  
Cae System ◽  
Element Method

Modern high-performance computing systems allow us to explore and implement new technologies and mathematical modeling algorithms into industrial software systems of engineering analysis. For a long time the finite element method (FEM) was considered as the basic approach to mathematical simulation of elasticity theory problems; it provided the problems solution within an engineering error. However, modern high-tech equipment allows us to implement design solutions with a high enough accuracy, which requires more sophisticated approaches within the mathematical simulation of elasticity problems in industrial packages of engineering analysis. One of such approaches is the spectral element method (SEM). The implementation of SEM in a CAE system for the solution of elasticity problems is considered. An important feature of the proposed variant of SEM implementation is a support of hybrid curvilinear meshes. The main advantages of SEM over the FEM are discussed. The shape functions for different classes of spectral elements are written. Some results of computations are given for model problems that have analytical solutions. The results show the better accuracy of SEM in comparison with FEM for the same meshes.

Large eddy simulation of the differentially heated cubic cavity flow by the spectral element method

2013 ◽  
Vol 86 ◽  
pp. 210-227 ◽  
Author(s):  
Christoph Bosshard ◽  
Abdelouahab Dehbi ◽  
Michel Deville ◽  
Emmanuel Leriche ◽  
Riccardo Puragliesi ◽  
...  
Keyword(s):  
Large Eddy Simulation ◽  
Spectral Element Method ◽  
Cavity Flow ◽  
Spectral Element ◽  
Eddy Simulation ◽  
Large Eddy ◽  
Element Method

scholarly journals Highly efficient self-powered perovskite photodiode with an electron-blocking hole-transport NiOx layer

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Amir Muhammad Afzal ◽  
In-Gon Bae ◽  
Yushika Aggarwal ◽  
Jaewoo Park ◽  
Hye-Ryeon Jeong ◽  
...  
Keyword(s):  
Energy Harvesting ◽  
Bias Voltage ◽  
High Performance ◽  
Hole Transport ◽  
Transport Layer ◽  
Hole Transport Layer ◽  
Zero Bias ◽  
Decay Times ◽  
Order Of Magnitude ◽  
Self Powered

AbstractHybrid organic–inorganic perovskite materials provide noteworthy compact systems that could offer ground-breaking architectures for dynamic operations and advanced engineering in high-performance energy-harvesting optoelectronic devices. Here, we demonstrate a highly effective self-powered perovskite-based photodiode with an electron-blocking hole-transport layer (NiOx). A high value of responsivity (R = 360 mA W−1) with good detectivity (D = 2.1 × 1011 Jones) and external quantum efficiency (EQE = 76.5%) is achieved due to the excellent interface quality and suppression of the dark current at zero bias voltage owing to the NiOx layer, providing outcomes one order of magnitude higher than values currently in the literature. Meanwhile, the value of R is progressively increased to 428 mA W−1 with D = 3.6 × 1011 Jones and EQE = 77% at a bias voltage of − 1.0 V. With a diode model, we also attained a high value of the built-in potential with the NiOx layer, which is a direct signature of the improvement of the charge-selecting characteristics of the NiOx layer. We also observed fast rise and decay times of approximately 0.9 and 1.8 ms, respectively, at zero bias voltage. Hence, these astonishing results based on the perovskite active layer together with the charge-selective NiOx layer provide a platform on which to realise high-performance self-powered photodiode as well as energy-harvesting devices in the field of optoelectronics.

An Improved Assembling Algorithm in Boundary Elements With Galerkin Weighting Applied to Three-Dimensional Stokes Flows

2017 ◽  
Vol 140 (1) ◽  
Author(s):  
Sofia Sarraf ◽  
Ezequiel López ◽  
Laura Battaglia ◽  
Gustavo Ríos Rodríguez ◽  
Jorge D'Elía
Keyword(s):  
Boundary Element Method ◽  
Linear System ◽  
Boundary Element ◽  
Numerical Solutions ◽  
Three Dimensional ◽  
Boundary Integral ◽  
Computational Time ◽  
Creeping Flows ◽  
Order Of Magnitude ◽  
Element Method

In the boundary element method (BEM), the Galerkin weighting technique allows to obtain numerical solutions of a boundary integral equation (BIE), giving the Galerkin boundary element method (GBEM). In three-dimensional (3D) spatial domains, the nested double surface integration of GBEM leads to a significantly larger computational time for assembling the linear system than with the standard collocation method. In practice, the computational time is roughly an order of magnitude larger, thus limiting the use of GBEM in 3D engineering problems. The standard approach for reducing the computational time of the linear system assembling is to skip integrations whenever possible. In this work, a modified assembling algorithm for the element matrices in GBEM is proposed for solving integral kernels that depend on the exterior unit normal. This algorithm is based on kernels symmetries at the element level and not on the flow nor in the mesh. It is applied to a BIE that models external creeping flows around 3D closed bodies using second-order kernels, and it is implemented using OpenMP. For these BIEs, the modified algorithm is on average 32% faster than the original one.

scholarly journals The Influence of Exposure Energy Density on Porosity and Microhardness of the SLM Additive Manufactured Elements

Materials ◽  
2018 ◽  
Vol 11 (11) ◽  
pp. 2304 ◽  
Author(s):  
Janusz Kluczyński ◽  
Lucjan Śnieżek ◽  
Krzysztof Grzelak ◽  
Janusz Mierzyński
Keyword(s):  
Computer Aided Design ◽  
Complex Geometry ◽  
Final Analysis ◽  
Machine Building ◽  
Layer By Layer ◽  
Laser Melting ◽  
Measurement Results ◽  
Computer Aided ◽  
Exposure Energy ◽  
Aided Design

Selective laser melting (SLM) is an additive manufacturing technique. It allows elements with very complex geometry to be produced using metallic powders. A geometry of manufacturing elements is based only on 3D computer-aided design (CAD) data. The metal powder is melted selectively layer by layer using an ytterbium laser. This paper contains the results of porosity and microhardness analysis made on specimens manufactured during a specially prepared process. Final analysis helped to discover connections between changing hatching distance, exposure speed and porosity. There were no significant differences in microhardness and porosity measurement results in the planes perpendicular and parallel to the machine building platform surface.

Sign in / Sign up

Export Citation Format

Share Document