Pasar al contenido principal

DESARROLLO DE CODIGOS Y ALGORITMOS PARALELOS DE ALTAS PRESTACIONES PARA LA MEJORA DE LA EFICIENCIA EN LOS SECTORES EOLICO, SOLARTERMICO Y EDIFICACION

ENE2014-60577-R

•
Nombre agencia financiadora Ministerio de Economía y Competitividad
Acrónimo agencia financiadora MINECO
Programa Programa Estatal de I+D+i Orientado a los Retos de la Sociedad
Convocatoria Proyectos I+D+i Retos de Investigación
Año convocatoria 2014
Plan Estatal 2013-2016 Plan Estatal de Investigación Científica y Técnica y de Innovación
Unidad de gestión Secretaría de Estado de Educación, Formación Profesional y Universidades
Centro beneficiario UNIVERSITAT POLITECNICA DE CATALUNYA
Identificador persistente http://dx.doi.org/10.13039/501100003329

Resultados relacionados

Resultados totales (Incluyendo duplicados): 20
Encontrada(s) 1 página(s)

UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Parallel mesh multiplication and adaptation technique for turbulent flow simulation using unstructured meshes

  • Antepara Zambrano, Óscar|||0000-0002-4596-0289
  • Lehmkuhl Barba, Oriol|||0000-0002-2670-1871
  • Borrell Pol, Ricard
  • Oliet Casasayas, Carles|||0000-0003-2170-5299
  • Oliva Llena, Asensio|||0000-0002-2805-4794
In this paper, we report the development of a parallel mesh multiplication code to subdivide a base mesh (containing tetrahedral, pyramids, prisms or hexahedral) into a finer mesh without user intervention, taking care of the quality of refined meshes generated by the algorithm. First, a coarse mesh is generated with an unstructured mesh generator and subdivided to the level of resolution needed for the simulation. On the refinement process of tetrahedral elements, geometrical properties are taking into account in order to preserve the shape quality of the subdivided elements. Then, the mesh is conformed on the solid surfaces to the original geometry, since linear subdivision ignores surface curvatures, and interior points of the mesh are adapted using Radial Basis functions, due to the surface correction. Moreover, a smoothing algorithm for tetrahedral mesh is applied to improve the mesh quality according to element geometry metric or based in minimize numerical errors in the CFD solution computed with unstructured meshes. Finally, the applicability of this method is shown on the numerical simulation of the turbulent flow over an Ahmed car at Reh = 7.68 × 10e5., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

New parallel method for adjacent disconnected unstructured 3D meshes

  • Muela Castro, Jordi|||0000-0003-1583-8490
  • Martínez Valdivieso, Daniel
  • Lehmkuhl Barba, Oriol|||0000-0002-2670-1871
  • Pérez Segarra, Carlos David|||0000-0003-1007-3142
  • Oliva Llena, Asensio|||0000-0002-2805-4794
A new parallel method for simulations with non-overlapping disconnected mesh domains but adjacent boundaries is presented and studied. This technique allows simulations using 3D unstructured meshes that are independent., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Fine-scale dynamics in turbulent Rayleigh-Bénard convection

  • Dabbagh, Firas|||0000-0001-8440-507X
  • Trias Miquel, Francesc Xavier|||0000-0002-5966-0703
  • Gorobets, Andrei
  • Oliva Llena, Asensio|||0000-0002-2805-4794
Small-scale universal features captured normally in many turbulent flows have been reviewed in Rayleigh-Bénard convection. Namely, dynamical Lagrangian representations of the velocity gradient tensor invariants have displayed the so-called “tearing drop” spiraling feature within the bulk region. Furthermore, essential common aspects as the asymmetric alignment of vorticity with the vortex-stretching vector, and the preferential alignment between vorticity and the intermediate eigenvector of the rate-of-strain tensor have been observed in that scope. Further and in more appropriate implement, we have performed a similar dynamical study of flow topology pertained to the traceless part of the gradient velocity-times-temperature tensor. The new invariants are related basically with the small-scale dynamics associated with the evolution of thermal plumes and buoyant production. We find that a self-amplification of velocity derivatives developed at harder turbulent flow supports indirectly the local effects of vortex-stretching generation. Moreover, the averaged evolution of the new invariants have revealed a converging change towards two skew-symmetric origins at Rayleigh number equals to 1e8 but one zero-valued origin at 1e10., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Strategies for the heterogeneous execution of large-scale simulations on hybrid supercomputers

  • Álvarez Farré, Xavier|||0000-0002-1684-7658
  • Gorobets, Andrei
  • Trias Miquel, Francesc Xavier|||0000-0002-5966-0703
Massively-parallel devices of various architectures are being adopted by the newest supercomputers to overcome the actual power constraint in the context of the exascale challenge. This progress leads to an increasing hybridisation of HPC systems and makes the design of computing applications a rather complex problem. Therefore, the software e¿ciency and portability are of crucial importance. In this context of accelerated innovation, we developed the HPC2 (Heterogeneous Portable Code for HPC). It is a portable, algebra-based framework for heterogeneous computing with many potential applications in the ¿elds of computational physics and mathematics, such as modelling of incompressible turbulent ¿ows. In its application to CFD, the algorithm of the timeintegration phase relies on a reduced set of only three algebraic operations: the sparse matrix-vector product, the linear combination of vectors and the dot product. This algebraic approach combined with a multilevel MPI+OpenMP+OpenCL parallelisation naturally provides modularity and portability. In this work, we present the strategies for the e¿cient heterogeneous execution of large-scale simulations on hybrid supercomputers that are part of the HPC2 core. The performance is studied in detail for the major computing kernel, the sparse matrix-vector product, using a sparse matrix derived from a simulation on a hybrid unstructured mesh and up to 32 nodes of a hybrid CPU+GPU supercomputer.
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

On the extension of RANS/LES methods from incom-pressible to compressible transonic turbulent flows with SBLI

  • Pedro Costa, Juan Bautista
  • Baez Vidal, Aleix|||0000-0003-4227-1768
  • Lehmkuhl Barba, Oriol|||0000-0002-2670-1871
  • Pérez Segarra, Carlos David|||0000-0003-1007-3142
  • Oliva Llena, Asensio|||0000-0002-2805-4794
In the present work, a compressible extension of the incompressible RANS and LES models implemented in TermoFluids is carried out. The objective is to solve compressible turbulent flows, i.e. high-Reynolds and high-Mach flows, by means of turbulence modelling. Specifically, transonic turbulent flows where shock-boundary layer interaction (SBLI) is present are the main target due to its practical interest. With this idea in mind, several simulations of the Sajben’s transonic diffuser are presented., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Towards a direct numerical simulation of a lifted CH4-air diffusion flame

  • Ventosa Molina, Jordi|||0000-0002-8276-5001
  • Lehmkuhl Barba, Oriol|||0000-0002-2670-1871
  • Pérez Segarra, Carlos David|||0000-0003-1007-3142
  • Oliva Llena, Asensio|||0000-0002-2805-4794
The near field jet of a self-igniting methane-air diffusion jet flame is studied by means of Direct Numerical Simulation (DNS) and a Flamelet/Progress-Variable (FPV) model to include detailed chemistry. Hence, all flow and transport scales are resolved and chemical reactions are assumed to take place in thin layers. Key aspects in the creation of the flamelet database, prior to the DNS simulation, are considered. Inclusion of differential diffusion effects are discussed by analysing the S-shaped curve and autoignition delay times. Similarly, the influence on these parameters of two detailed mechanism, the GRI2.11 and GRI3.0, is assessed. The definition of the progress-variable used in the FPV is analysed and a new definition is proposed. Results of the near jet, up to 20 jet diameters, are shown for the two detailed mechanism., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

A complete analysis of the heat, air and moisture transfer on building performance

  • Souaihi, Oussama
  • Lopez, Joan
  • Capdevila Paramio, Roser|||0000-0003-2569-2544
  • Rigola Serrano, Joaquim|||0000-0002-6685-3677
  • Lehmkuhl Barba, Oriol|||0000-0002-2670-1871
  • Oliva Llena, Asensio|||0000-0002-2805-4794
The simulation of combined heat, air and moisture (HAM) is important to predict the indoor air quality and thermal comfort. Moreover, inappropriate levels of indoor humidity and temperature can contribute to a high movement of water vapor through the building walls, causing deterioration and reduction of the thermal insulation which leads to higher energy demand. The simulation of the buildings behavior can help to optimize the design of new or existing building, help better control the HVAC system and therefore results in energy efficient buildings. In this work, an in-house modular object-oriented tool (NEST) for the multiphysics simulation of buildings is presented. The whole building is modeled as a collection of basic elements (e.g., walls, rooms, openings, occupancy, HVAC system, solar radiation distributor, etc.). These elements can be modeled using different physical models and scales. A combined heat, air and moisture transfer model for the building envelopes and rooms have been implemented and validated with different benchmark cases. The in-house simulation tool has been used for the simulation of hygrothermal behavior of rooms inside different public buildings (residential apartments, hospital rooms, universities and school plants). The simulations allowed as the analysis of the humidity effect on thermal comfort and energy performance of the rooms., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Building proper invariants for subgrid-scale eddy-viscosity models

  • Trias Miquel, Francesc Xavier|||0000-0002-5966-0703
  • Folch, David
  • Gorobets, Andrei
  • Oliva Llena, Asensio|||0000-0002-2805-4794
Since direct simulations of the incompressible Navier-Stokes equations are limited to relatively low-Reynolds numbers, dynamically less complex mathematical formulations are necessary for coarse-grain simulations. Eddy-viscosity models for Large-Eddy Simulation is probably the most popular example thereof: they rely on differential operators that should be able to capture well different flow configurations (laminar and 2D flows, near-wall behavior, transitional regime...). Most of them are based on the combination of invariants of a symmetric second-order tensor that is derived from the gradient of the resolved velocity field. In the present work, they are presented in a framework where all the models are represented as a combination of elements of a 5D phase space of invariants. In this way, new models can be constructed by imposing appropriate restrictions in this space. The performance of the proposed models is successfully tested for a turbulent channel flow., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Effectiveness of adaptive mesh refinement strategies in the DNS of multiphase flows

  • Schillaci, Eugenio|||0000-0002-6690-6871
  • Antepara Zambrano, Óscar|||0000-0002-4596-0289
  • Lehmkuhl Barba, Oriol|||0000-0002-2670-1871
  • Balcázar Arciniega, Néstor|||0000-0003-0776-2086
  • Oliva Llena, Asensio|||0000-0002-2805-4794
In this work, a finite-volume Adaptive Mesh Refinement (AMR) method for two phase flows able to dynamically achive the required mesh size in specific parts of the domain has been developed. The Navier-Stokes equations for incompressible flows are coupled to a robust conservative level-set method for the capture of the interface between the fluids. The effectiveness of the method has been firstly validated by simulating some known test-cases, as the vortex flow and the rising bubbles. Finally, it has been applied to a possible industrial application, the simulation of a jet injector., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Memory aware poisson solver for peta-scale simulations with one FFT diagonalizable direction

  • Oyarzun Altamirano, Guillermo|||0000-0001-9524-3782
  • Borrell Pol, Ricard
  • Trias Miquel, Francesc Xavier|||0000-0002-5966-0703
  • Oliva Llena, Asensio|||0000-0002-2805-4794
Problems with some sort of divergence constraint are found in many disciplines: computational fluid dynamics, linear elasticity and electrostatics are examples thereof. Such a constraint leads to a Poisson equation which usually is one of the most computationally intensive parts of scientific simulation codes. In this work, we present a memory aware Poisson solver for problems with one Fourier diagonalizable direction. This diagonalization decomposes the original 3D system into a set of independent 2D subsystems. The proposed algorithm focuses on optimizing the memory allocations and transactions by taking into account redundancies on such 2D subsystems. Moreover, we also take advantage of the uniformity of the solver through the periodic direction for its vectorization. Additionally, our novel approach automatically optimizes the choice of the preconditioner used for the solution of each frequency subsystem and dynamically balances its parallel distribution. Altogether constitutes a highly efficient and robust HPC Poisson solver that has been successfully attested up to 16384 CPU-cores., Peer Reviewed
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

A complete analysis of the heat, air and moisture transfer on building performance

  • Souaihi, Oussama
  • Lopez, Joan
  • Capdevila Paramio, Roser
  • Rigola Serrano, Joaquim
  • Lehmkuhl Barba, Oriol
  • Oliva Llena, Asensio
The simulation of combined heat, air and moisture (HAM) is important to predict the indoor air quality and thermal comfort. Moreover, inappropriate levels of indoor humidity and temperature can contribute to a high movement of water vapor through the building walls, causing deterioration and reduction of the thermal insulation which leads to higher energy demand. The simulation of the buildings behavior can help to optimize the design of new or existing building, help better control the HVAC system and therefore results in energy efficient buildings. In this work, an in-house modular object-oriented tool (NEST) for the multiphysics simulation of buildings is presented. The whole building is modeled as a collection of basic elements (e.g., walls, rooms, openings, occupancy, HVAC system, solar radiation distributor, etc.). These elements can be modeled using different physical models and scales. A combined heat, air and moisture transfer model for the building envelopes and rooms have been implemented and validated with different benchmark cases. The in-house simulation tool has been used for the simulation of hygrothermal behavior of rooms inside different public buildings (residential apartments, hospital rooms, universities and school plants). The simulations allowed as the analysis of the humidity effect on thermal comfort and energy performance of the rooms., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

On the extension of RANS/LES methods from incom-pressible to compressible transonic turbulent flows with SBLI

  • Pedro Costa, Juan Bautista
  • Baez Vidal, Aleix
  • Lehmkuhl Barba, Oriol
  • Pérez Segarra, Carlos David
  • Oliva Llena, Asensio
In the present work, a compressible extension of the incompressible RANS and LES models implemented in TermoFluids is carried out. The objective is to solve compressible turbulent flows, i.e. high-Reynolds and high-Mach flows, by means of turbulence modelling. Specifically, transonic turbulent flows where shock-boundary layer interaction (SBLI) is present are the main target due to its practical interest. With this idea in mind, several simulations of the Sajben’s transonic diffuser are presented., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Parallel mesh multiplication and adaptation technique for turbulent flow simulation using unstructured meshes

  • Antepara Zambrano, Óscar
  • Lehmkuhl Barba, Oriol
  • Borrell Pol, Ricard
  • Oliet Casasayas, Carles
  • Oliva Llena, Asensio
In this paper, we report the development of a parallel mesh multiplication code to subdivide a base mesh (containing tetrahedral, pyramids, prisms or hexahedral) into a finer mesh without user intervention, taking care of the quality of refined meshes generated by the algorithm. First, a coarse mesh is generated with an unstructured mesh generator and subdivided to the level of resolution needed for the simulation. On the refinement process of tetrahedral elements, geometrical properties are taking into account in order to preserve the shape quality of the subdivided elements. Then, the mesh is conformed on the solid surfaces to the original geometry, since linear subdivision ignores surface curvatures, and interior points of the mesh are adapted using Radial Basis functions, due to the surface correction. Moreover, a smoothing algorithm for tetrahedral mesh is applied to improve the mesh quality according to element geometry metric or based in minimize numerical errors in the CFD solution computed with unstructured meshes. Finally, the applicability of this method is shown on the numerical simulation of the turbulent flow over an Ahmed car at Reh = 7.68 × 10e5., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Fine-scale dynamics in turbulent Rayleigh-Bénard convection

  • Dabbagh, Firas
  • Trias Miquel, Francesc Xavier
  • Gorobets, Andrei
  • Oliva Llena, Asensio
Small-scale universal features captured normally in many turbulent flows have been reviewed in Rayleigh-Bénard convection. Namely, dynamical Lagrangian representations of the velocity gradient tensor invariants have displayed the so-called “tearing drop” spiraling feature within the bulk region. Furthermore, essential common aspects as the asymmetric alignment of vorticity with the vortex-stretching vector, and the preferential alignment between vorticity and the intermediate eigenvector of the rate-of-strain tensor have been observed in that scope. Further and in more appropriate implement, we have performed a similar dynamical study of flow topology pertained to the traceless part of the gradient velocity-times-temperature tensor. The new invariants are related basically with the small-scale dynamics associated with the evolution of thermal plumes and buoyant production. We find that a self-amplification of velocity derivatives developed at harder turbulent flow supports indirectly the local effects of vortex-stretching generation. Moreover, the averaged evolution of the new invariants have revealed a converging change towards two skew-symmetric origins at Rayleigh number equals to 1e8 but one zero-valued origin at 1e10., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Building proper invariants for subgrid-scale eddy-viscosity models

  • Trias Miquel, Francesc Xavier
  • Folch, David
  • Gorobets, Andrei
  • Oliva Llena, Asensio
Since direct simulations of the incompressible Navier-Stokes equations are limited to relatively low-Reynolds numbers, dynamically less complex mathematical formulations are necessary for coarse-grain simulations. Eddy-viscosity models for Large-Eddy Simulation is probably the most popular example thereof: they rely on differential operators that should be able to capture well different flow configurations (laminar and 2D flows, near-wall behavior, transitional regime...). Most of them are based on the combination of invariants of a symmetric second-order tensor that is derived from the gradient of the resolved velocity field. In the present work, they are presented in a framework where all the models are represented as a combination of elements of a 5D phase space of invariants. In this way, new models can be constructed by imposing appropriate restrictions in this space. The performance of the proposed models is successfully tested for a turbulent channel flow., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Strategies for the heterogeneous execution of large-scale simulations on hybrid supercomputers

  • Álvarez Farré, Xavier
  • Gorobets, Andrei
  • Trias Miquel, Francesc Xavier
Massively-parallel devices of various architectures are being adopted by the newest supercomputers to overcome the actual power constraint in the context of the exascale challenge. This progress leads to an increasing hybridisation of HPC systems and makes the design of computing applications a rather complex problem. Therefore, the software e¿ciency and portability are of crucial importance. In this context of accelerated innovation, we developed the HPC2 (Heterogeneous Portable Code for HPC). It is a portable, algebra-based framework for heterogeneous computing with many potential applications in the ¿elds of computational physics and mathematics, such as modelling of incompressible turbulent ¿ows. In its application to CFD, the algorithm of the timeintegration phase relies on a reduced set of only three algebraic operations: the sparse matrix-vector product, the linear combination of vectors and the dot product. This algebraic approach combined with a multilevel MPI+OpenMP+OpenCL parallelisation naturally provides modularity and portability. In this work, we present the strategies for the e¿cient heterogeneous execution of large-scale simulations on hybrid supercomputers that are part of the HPC2 core. The performance is studied in detail for the major computing kernel, the sparse matrix-vector product, using a sparse matrix derived from a simulation on a hybrid unstructured mesh and up to 32 nodes of a hybrid CPU+GPU supercomputer., Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Effectiveness of adaptive mesh refinement strategies in the DNS of multiphase flows

  • Schillaci, Eugenio
  • Antepara Zambrano, Óscar
  • Lehmkuhl Barba, Oriol
  • Balcázar Arciniega, Néstor
  • Oliva Llena, Asensio
In this work, a finite-volume Adaptive Mesh Refinement (AMR) method for two phase flows able to dynamically achive the required mesh size in specific parts of the domain has been developed. The Navier-Stokes equations for incompressible flows are coupled to a robust conservative level-set method for the capture of the interface between the fluids. The effectiveness of the method has been firstly validated by simulating some known test-cases, as the vortex flow and the rising bubbles. Finally, it has been applied to a possible industrial application, the simulation of a jet injector., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

New parallel method for adjacent disconnected unstructured 3D meshes

  • Muela Castro, Jordi
  • Martínez Valdivieso, Daniel
  • Lehmkuhl Barba, Oriol
  • Pérez Segarra, Carlos David
  • Oliva Llena, Asensio
A new parallel method for simulations with non-overlapping disconnected mesh domains but adjacent boundaries is presented and studied. This technique allows simulations using 3D unstructured meshes that are independent., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Towards a direct numerical simulation of a lifted CH4-air diffusion flame

  • Ventosa Molina, Jordi
  • Lehmkuhl Barba, Oriol
  • Pérez Segarra, Carlos David
  • Oliva Llena, Asensio
The near field jet of a self-igniting methane-air diffusion jet flame is studied by means of Direct Numerical Simulation (DNS) and a Flamelet/Progress-Variable (FPV) model to include detailed chemistry. Hence, all flow and transport scales are resolved and chemical reactions are assumed to take place in thin layers. Key aspects in the creation of the flamelet database, prior to the DNS simulation, are considered. Inclusion of differential diffusion effects are discussed by analysing the S-shaped curve and autoignition delay times. Similarly, the influence on these parameters of two detailed mechanism, the GRI2.11 and GRI3.0, is assessed. The definition of the progress-variable used in the FPV is analysed and a new definition is proposed. Results of the near jet, up to 20 jet diameters, are shown for the two detailed mechanism., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Memory aware poisson solver for peta-scale simulations with one FFT diagonalizable direction

  • Oyarzun Altamirano, Guillermo
  • Borrell Pol, Ricard
  • Trias Miquel, Francesc Xavier
  • Oliva Llena, Asensio
Problems with some sort of divergence constraint are found in many disciplines: computational fluid dynamics, linear elasticity and electrostatics are examples thereof. Such a constraint leads to a Poisson equation which usually is one of the most computationally intensive parts of scientific simulation codes. In this work, we present a memory aware Poisson solver for problems with one Fourier diagonalizable direction. This diagonalization decomposes the original 3D system into a set of independent 2D subsystems. The proposed algorithm focuses on optimizing the memory allocations and transactions by taking into account redundancies on such 2D subsystems. Moreover, we also take advantage of the uniformity of the solver through the periodic direction for its vectorization. Additionally, our novel approach automatically optimizes the choice of the preconditioner used for the solution of each frequency subsystem and dynamically balances its parallel distribution. Altogether constitutes a highly efficient and robust HPC Poisson solver that has been successfully attested up to 16384 CPU-cores., Peer Reviewed, Postprint (author's final draft)