ECFD workshop, 7th edition, 2024

Description

• Event from 22th of January to 2nd of February 2024
• Location: Hôtel Club de la Plage, Merville-Franceville, near Caen (14)
• Two types of sessions:
  • common technical presentations: roadmaps, specific points
  • mini-workshops. Potential workshops are listed below
• Free of charge
• More than 70 participants from academics, HPC center/experts and industry.

• Objectives
  • Bring together experts in high-performance computing, applied mathematics and multi-physics CFDs
  • Identify the technological barriers of exaflopic CFD via numerical experiments
  • Identify industrial needs and challenges in high-performance computing
  • Propose action plans to add to the development roadmaps of the CFD codes
• Organizers
  • Guillaume Balarac (LEGI), Simon Mendez (IMAG), Pierre Bénard, Vincent Moureau, Léa Viovenel (CORIA).

Agenda

Thematics / Mini-workshops

These mini-workshops may change and cover more or less topics. This page will be adapted according to your feedback.

To come...

Projects

Hackathon GENCI - P. Begou, LEGI

The GENCI Hackathon will be devoted to porting two CFD codes to the Mi250 GPUs of the Adastra supercomputer deployed by GENCI at CINES.

For the YALES2 code the goal is to obtain a first reference version giving the expected results then, if possible, to start its optimization to gain performance. The approach is OpenACC based with the objective of an implementation as least intrusive as possible in the existing code and which remains portable with the work done on the Nvidia GPUs of the Jean-Zay supercomputer at IDRIS.

The porting of the AVBP code is more advanced with a prototype already functional on Adastra but "hard-coded". The objective is to rationalize this first implementation, to integrate the latest developments in the code, to centralize memory management (host and device), to work on porting the Lagrangian part of the code and, of course, to improve the global performance.

This Hackathon is supported by GENCI, HPE, AMD and CINES with the presence on site of several development experts on AMD GPUS.

Mesh adaptation - R. Letournel, Safran

M1: ASMR for reheat chamber applications - Paul Pouech (CERFACS), Thibault Duranton, Luis Carbajal Carrasco (Safran)

Combustion in reheat chambers feature a wide range of lenght scales. Mesh refinement is thus mandatory to capture the flow characteristics within a reasonnable CPU cost for LES computations using the AVBP code. The purpose of this project is to consolidate mesh refinement criteria and strategy in an academic reference case. The retained workflow is supported by the Lemmings code that calls the Tékigô wrapper for the mesh adaptations. During the ECFD7, the convergence time needed to have significant distribution of quantities of interest was analysed. An optimum runtime, based on a characteristic flow time-scale, was thus identified and led to a reduced running time for each adaptation step. As a second step, discussions with the ECFD7 participants led to the identification of interesting refinement criteria, namely the flame sensor or the mach rms for instance. Parametric analysis showed the robustness of the workflow based on a ponderation of different criteria. Finally, in order to facilitate the use of the workflow, efforts were made to improve the user experience by making it more human readable.

M2: Parallel remeshing - B. Andrieu, C. Benazet, K. Hoogveld, B. Maugars, E. Quémerais (ONERA)

Mesh adaptation is a crucial tool in order to automate industrial RANS numerical simulations. To meet this need, we need to carry out mesh adaptation as quickly as possible by setting up an efficient, parallel solution. To this end, we have explored two avenues: a parallel edge-splitting algorithm that has recently been initiated in the ParaDiGM library, and a solution based on the refine library for adapting meshes with MPI implementation. On the one hand, we fixed several bugs in our split operator, and validated it on test cases of increasing complexity with a node-centered solver. In addition, we've added interfaces to refine so as to avoid using files, and call directly in library mode. We also investigated geometric projection issues during the mesh adaptation procedure, notably by looking at solutions such as EGADS, which offers a simplified API for CAD interrogation. We finally implemented metric gradation (in serial), metric intersection and complexity computations. All the ingredients we've tested give us a clearer picture of the entire mesh adaptation process.

M3: Anisotropic mesh refinement - R. Barbera (LEGI & Safran), G. Ghigliotti, G. Balarac (LEGI), R. Letournel (Safran)

Mesh adaptation is now a key feature for simulations of complex industrial flows. For transient flows such as multiphase and/or reactive flows, where regions of interest are strongly moving in space, dynamic mesh adaptation appears as the most suitable strategy. This strategy is now widely used in YALES2 based on isotropic mesh definition. The purpose of this project is to adapt this strategy to an anisotropic framework to reduce the overall simulation costs (in term of memory consumption, cpu cost and time to solution). In order to be able to handle multiphase flows, the main objective of the project is to study the conditions for accurately describing the dynamics of the level-set function with an anisotropic mesh. Accuracy is mainly assessed in terms of interface position and mass conservation. The inaccuracy of mass conservation is mainly due to interpolation errors after the adaptation step. Furthermore, inaccuracy in interface position may be due to misalignment between the anisotropic mesh elements and the interface normal. The first methodological corrections have been proposed, as an adaptation of the level-set reinitialization algorithm to the anisotropic mesh.

Numerics - S. Mendez, IMAG & G. Balarac, LEGI

N1: Treatment of boundary conditions for high-order schemes - M. Bernard & G. Balarac (LEGI), G. Lartigue (Total Energies)

In the context of Finite Volumes Method, spacial accuracy of a numerical scheme depends on ability to evaluate accurately fluxes through interface of each control volume (CV). Such accurate evaluation is not straightforward, especially when dealing with distorted grids. This project follows the work of [1] where fluxes use pointwise quantities, which are reconstructed from integrated quantities advanced in time. During the workshop, task force was dedicated to the treatment of **inlet** boundary conditions (BC) and **non-planar walls**. For inlet BC, the key resides in the spatial integration of convective flux over discrete faces of the CV touching the boundary. Such treatment lead to exact integration for linear inlet profile and large error reduction on other profiles. Concerning non-planar walls, the strategy adopted consists in the enforcement of the BC on each discrete face, by modifying the normal component of the wall gradient in order to evaluate accurately the diffusive flux. Again, a large reduction of this error has been observed.

[1] A framework to perform high-order deconvolution for finite-volume method on simplicial meshes, , Bernard et. al., IJNMF 2020

N2: Implementation of linearised implicit time integration in ALE solver - T. Berthelon, G. Balarac (LEGI)

An linearised implicit time integration has recently been developed in the incompressible solver of YALES2. This new integration scheme allows to use larger time-step that the ones constraints by classic stability criteria inherent to explicit time integration method. This allows to reduce the restitution time of Large Eddy Simulations [1]. The objective of this project was to implement this new time integration in the ale solver in order to be able to reduce restitution time of moving mesh configuration.

The developments were validated on a scalar advection case and on a rotor-stator interaction case. Although the results seem to be in line with the explicit integration methods, the validation of the temporal convergence to 2nd order remains to be shown.

[1] Toward the use of LES for industrial complex geometries. Part II: Reduce the time-to-solution by using a linearised implicit time advancement, Berthelon et al., JoT, 2023

N4: Non-uniform outlet pressure and coupling with CWIPI - J. B. Lagaert (LMO), Y. Lakrifi, T. Berthelon, G.Balarac (LEGI) , R. Letournel (Safran)

In simulations, artificial boundaries need to be introduced due to the limited size of computational domains. At these boundaries, flow variables need to be calculated in a way that will not induce any perturbation of the interior solution. During ECFD#7, a generic outlet boundary condition defined from non-uniform pressure has been implemented in Yales2. This non-uniform pressure can de determined from a traction model (null or advected from the interior domain, for example). This non-uniform pressure can also be deducted through a coupling between two simulations. In this case a coupling via CWIPI is performed where the velocity and the pressure are exchanged at the common boundary to define the inlet and outlet conditions, respectively.

N5: Optimization of the RBC solver - F. Rojas, S. Mendez (IMAG)

In the study of blood diseases, the mechanical behaviour of Red Blood Cells (RBCs) is one of the most relevant effects to take into account in the numerical models but also in experimental setups. Our system of interest is the thin gap of a rheometer where RBC suspensions are placed to explore their properties. To interpret the experimental data, the simulations of large suspensions of RBC are required to determine the blood’s microstructure (spatial arrangement of cells) and its rheological properties.

Currently, YALES2BIO’s RBC solver is capable to manage thousands of cells, but in order to approach closer to the experimental scales, we propose the characterisation and optimisation of its performance to reduce the computational requirements and increase the RBC’s number and domain sizes in our simulations. During the workshop a parametric study was carried out to obtain the strong and weak scaling. Studying the increase in the volume fraction allowed us to quantify how the cost of the simulation increases rapidly with the RBC’s number and identify which routines have the biggest impact on the performance. One conclusion is that the cost is spread of several routines, which makes code optimization more cumbersome. However, the amount of RBCs and RBC nodes duplicated over processors is identified as a key factor for performance. Indeed, as RBCs may interact with several partitions, it is duplicated as much as needed based on criteria of boundaing box intersections. However, the current criteria have been shown to be too loose.

In order to limit the amount of work during the RBC processing, stricter criteria were introduced to avoid unnecessary calculations at the level of the nodes with a small gain in performance. On the other hand, much better results were obtained using cartesian partitioning to optimise the bounding box of each processor, reducing the involved RBC operations: this demonstrates that the performances of the RBC solver may be optimized by a stricter selection of RBC duplicates over processors.

We thank Ghislain Lartigue and Renaud Mercier for helpful discussions.

N6: Electrodeformation of red blood cells, extension to 3D and improved accuracy at membrane - A. Spadotto, S. Mendez (IMAG), M. Bernard (LEGI)

The Leaky Dielectric Model is a popular framework to describe electric stresses over micro-scale membranes. We have adopted it to simulate the effect of a DC electric field on a red blood cell using the YALES2BIO solver. The goal of the project is to reproduce the electric charging process of the membrane, as well as the resulting stresses, which may yield to electrodeformation of the cell. From the point of view of the implementation, the grid is represented by a 2D surface mesh embedded in a 3D eulerian grid. The need to make variables stored on the surface interact with quantities stored on the Eulerian grid calls for a proper bidirectional 2D-membrane/3D-grid dynamic connectivity. The advancement of theis task during this ECFD has led to the first 3D simulation of a charging fixed spherical shell. Moreover, the estimation of grid variables on elements cut by the membrane has been improved thanks to a High-Order extrapolation. The latter has been successfully tested on 2D configurations. The project opens the way for a series of validation tests. In particular, future work will demand treatment of instabilities emerging in symmetrical configurations.

N7: Optimisation Dorothy - M. Roperch, G. Pinon (LOMC), B. Gaston (CRIANN), P. Benard (CORIA)

Dorothy is a Lagrangian code using the particle vortex method. This method must have a homogeneous distribution of particles in space. To achieve this, at regular intervals during the simulation a Cartesian grid with new particles is created. The weights of the old particles are interpolated for each of the new particles. Before ECFD7, all the processors knew the general grid and the new particles. The aim of ECFD was to parallelize this module to avoid memory problem. To do this, each processor creates a grid corresponding to the particles it knows. They then exchange data on the supperposition zones. This solves the issue because the quantity of new particles known is smaller. During ECFD7, a trial on a ring vortex case was successfully carried out to test domain communications and supperposition. The next step will be to implement this new method in the Dorothy code.

Turbulence - P. Benard, CORIA & L. Bricteux, UMONS

T1: Wall Law for immersed boundaries – P. Bénez (CORIA), M. Cailler (Safran), S. Meynet (GDTech), J. Carmona (CORIA), Y. Bechane (CORIA)

Conservative Lagrangian Immersed Boundaries (CLIB) are now a useful way to take into account complex geometries in YALES2. In order to study highly turbulent configurations, it appears necessary to implement wall law models adapted to this method. If we consider a non-moving immersed body, developing wall-law models in a conservative immersed boundary formalism presents numerous challenges related to the diffuse interface property of the solid and the continuous formulation of the penalty force. During the ECFD, a new formulation of the penalty force has been established to ensure the imposition of the wall shear stress across the immersed solid interface. A strategy based on the use of two near-wall level sets was implemented to estimate the wall shear stress from the LES fluid velocity field at a distance D from the solid interface. At the end of the ECFD, turbulent flat plate cases were set up to start the validation of the strategy implemented for a logarithmic wall law. Future works will focus on validating this strategy for fixed solids.

T2: Turbulence injection Compressible flows – P. Tene Hedje (UMONS), J. Carmona (CORIA), Y. Bechane (CORIA), L. Bricteux (UMONS)

Turbulence injection for compressible flows remains a real challenge. Indeed, In these types of flow, the acoustic waves must also be controlled on boundaries. In addition, the non-reflective formulation of the Navier-Stokes characteristic Boundary Conditions (NSCBC) generally used in compressible solvers produce spurious pressure oscillations when applied to turbulent flows, making turbulence injection difficult for such applications. During the ECFD, two turbulence injection approaches were investigated and applied within the framework of the Explicit compressible solver (ECS) of YALES2. The first involved modifying the NSCBC formulation to inject turbulence from the inlet of the domain. To this end, the vortical-flow characteristic boundary condition [1] was implemented in ECS and the first validations were performed. The second was to use AL to generate a turbulence grid in the flow [2]. Future works will focus on further validating these approaches.

[1] Guézennec et al., Acoustically nonreflecting and reflecting boundary conditions for vortcity injection in compressible solvers, AIAA journal, 47(7), 1709-1722, 2009.

[2] Houtin-Mongrolle et al., Actuator line method applied to grid turbulence generation for large-Eddy simulations, Journal of Turbulence, 21(8), 407-433, (2020).

T3: Aero-servo-elastic simulations of wind turbines including atmospheric effects – E. Muller (SGRE), U. Vigny (UMONS), P. Benard (CORIA), F. Houtin-Mongrolle (SGRE)

Aero-servo-elastic engineering solvers used in the industry (i.e., BHawC) for structural response and power assessments are unsuited for wake simulations, as aerodynamic loads are usually derived from a BEM-like method. To tackle this, the YALES2 library was coupled (P11-ECFD3) to BHawC, the Siemens Gamesa Renewable Energy (SGRE) in-house certification code for wind turbines. This allowed the investigation of neutral atmospheric conditions. This project aims to include stable and unstable atmospheric conditions into this coupling based on the development done in T4-ECFD7. Therefore, this project is divided into three work packages: Work package 1: Adjustment and refactoring of the existing coupling library between YALES2 and BHawC. Work package 2: Rethink how turbulence is injected into the simulation (recycling in SGRE setup) to consider thermal and Coriolis effects. Work package 3: Adapt how the blade forces are computed in the coupling to consider the resulting density fluctuations.

T4: Atmospheric solver – U. Vigny (UMONS), L. Voivenel (CORIA), S. Zeoli (UMONS), P. Benard (CORIA)

Wind turbines, bigger and bigger, are now influenced by atmospheric flows. An atmospheric solver has already been developed in YALES2 to represents some of its effects (Coriolis, veer, thermal stratification). In this continuum, the project has been divided into two work-packages. - Work-package 1: The use of the Variable density solver (VDS). Before ECFD7, thermal stratification was taken into account using the Boussinesq buoyancy approximation within the incompressible solver framework. Now, VDS can be used, taking into account all thermal effect. Results are promissing. - Work-package 2: Wall law velocity filtering. Wall law are using velocity at the first grid node to compute wall shear stress. Before ECFD7, atmospheric wall law were using the local velocity, leading sometimes to convergence errors. Now a gather-scatter filter can be used to average velocity (and temperature) at first grid node.

T5: Implementation of the RVMs-WALE model in YALES2 – L. Bricteux (UMONS), P. Benard (CORIA), Y. Bechane (CORIA)

This study focused on the implementation of an advanced multiscale variational subgrid-scale model, incorporating scaling based on the WALE (Wall-Adapting Local Eddy-viscosity) model within YALES2. This model has demonstrated efficiency across various flow configurations, and it is anticipated that its multiscale nature can enhance the spectral selectivity. The aim is to ensure that its dissipative effects specifically target the smallest scales near the cut-off point.

Additionally, collaborative work with G. Balarac aimed to enhance the mesh adaptation strategy for wall-bounded flows with massive boundary layer detachment and vortical wake. This new strategy based on vortex detection was developed during the ECFD6 and ECFD4 workshops. We have now shown that this strategy is effective. Flow simulations around a hemisphere at Reynolds number Re=55K have been conducted, and we anticipate publishing the results soon."

Two Phase Flow - M. Cailler, Safran & V. Moureau, CORIA

P1: Level set reinitialization at the contact line for boiling flows - H. Lam, M. Benard, G. Ghigliotti (LEGI)

In Savinien Pertant's PhD thesis (2022), DNS of nucleate boiling at the bubble scale were performed, but suffered some lack of accuracy in the imposition of the contact angle. Indeed, the contact angle was not well respected, with a difference of around 10 degrees between the desired angle and the angle measured on the solution. This lack of accuracy, that contrast with the accurate imposition obtained in the spray solver (SPS), is due to fluctuations of the contact line. This behavior that was traced back to the modifications of the level set reinitialisation needed to take correctly into account the triple line, and for which the solution applied in 2022 was to revert to the standard Janodet reinitialisation. S. Pertant tested, at the very end of its PhD, a correction which nullifies the temperature transport term at the first node from the wall of the contact line. This correction was introduced to overcome an instability of the code when the contact line velocity on the substrate changes direction, from receding to advancing. It turned out at the ECFD7 that this correction proves to be very efficient to stabilise the contact line for contact angles between 50 to 90 degrees even in the case of the use of the level set reinitialization. We were able to simulate nucleate boiling with a smooth contact line at the triple line and a precision in the angle of the contact angle of around +-0.5 degrees. More work remains to be able to run DNS of nucleate boiling for extreme contact angles (<50° and >100°). Moreover, longer runs will be needed to further confirm these results.

P2: Compatibility of Boiling solver with PCS and MPH structure - H. Lam, M. Benard, G. Ghigliotti (LEGI)

The boiling solver does not work since the introduction of MPH data structure and PCS solver in March 2022. An investigation work was carried out to understand the changes made between the previous and the new version of the different solvers. A simple test case was created to show potential differences between the working version of the code and the new one. Several problems were spotted: the order of level set declaration became important as it is the first one declared which is advected. Sign convention was chosen differently for the mass transfer rate. The temporal discretization of the level set was different. A test case with no flow and at an imposed mass transfer rate (i.e., no coupling of the level set with the temperature field) was run successfully and the results of the commit prior to the March 2022 modifications were retrieved. More work is needed to find the origin of the differences between the two solvers when the temperature field is solved and coupled with the level set and the velocity field. New common test cases for the two solvers will have to be implemented in order to cross-validate the results and avoid such cases happening again (i.e., cross-fertilization).

P3: Blood platelets adhesion model - C. Raveleau, S. Mendez, F. Nicoud (IMAG)

Medical devices in contact with blood (e.g. artificial valves) are used to treat various cardiovascular diseases, but their thrombogenicity remains the main unresolved issue in their development. A numerical model of blood platelets is being constructed to help to understand the effect of microstructuration on the thrombogenicity of artificial surface. The Force Coupling Method (FCM) was previously implemented and allows the modelisation of ellipsoidal particle and their interaction with the surrounding fluid. During the workshop, the particle model was extended to include adhesive and repulsive interactions with walls or with other particles. The adhesive bonds are modeled with springs forming when the distance between a node of a particle surface and a node of the wall or another particle is smaller than a given threshold. The stiffness of the bond is increased after a given formation time to mimic the 2-step adhesion process of platelets to von Willebrand Factor. A Lennard-Jones potential was used to model the collision of particles. Future work will aim at generalizing these implementations for an arbitrary number of particles (currently only working for 2 particles) and ensuring the interactions are unaltered by the crossing of a periodic boundary.

P5: Towards even more efficient particle algorithms - M. Helal (CORIA & Safran), M. Cailler (Safran)

Lagrangian particles are widely used in the YALES2 plateform to model: liquid spray, granular flow, two-phase flows with SPH approach or solids in IB method. Though important developments to handle efficiently high number of particles in massively parallel simulations, the growing use of particles in Yales2 make necessary to re-evaluate and optimize the performances of Lagrangian particles algorithms handling. Objective of this project was twofold: analyze and improve the performance and robustness of the newly developed SPH solver of YALES2 and improve the performance of the Lagrangian particle relocation (identification of connectivity between Lagrangian and Eulerian grid) during the Dynamic Mesh Adaptation. Regarding the first subject, profiling tools have been used to identify the hot-spots and bottle-necks in the SPH solver. Optimizations including code factorization, removal of string comparison allows to reduce the computational cost by a factor 3. Moreover, robustification of the solver was achieved. In the second sub-project, a new implicit 4th-level decomposition has been introduced. This implicit decomposition consists in contiguous coloring of sub-el_grp in element group. The availability of smallest group of elements has been used to improve the local particle relocation algorithm that mainly relies on bounding-box comparison. This new relocation algorithm has been tested for various number of sub-el_grp on a representative case of gear lubrication showing a decrease by a factor 3 to 5 of the relocation algorithm. Perspective is to extend the use of sub-el_grp to the interpolation algorithm.

P8: Wall liquid film numerical model - N. Gasnier (EM2C & Safran), J. Leparoux (Safran), J. Carmona (CORIA)

Wall liquid films are likely to be formed when fuel sprays impact the walls of aeronautical fuel injection systems. Such phenomenon may have a significant influence on the whole combustion process, however the small scales involved prevent from performing high fidelity simulations of film flows in the context of industrial geometries. Therefore, a low order model is required to model the dynamics of thin liquid flows under the action of spray droplets and of a turbulent gas shear. During ECFD7, a liquid film numerical model accounting for the influence of surface tension as well as gas shear, and based on the 2-dimensional Shallow Water Equations was implemented in Yales2. This model was then coupled to an algorithm ensuring a proper transition between fully resolved liquid structures (levelset) and film model during liquid droplet impacts on a solid wall.

P9: Casting simulation for the study of ceramic core displacement - S. Sirot, R. Mercier, M. Cailler (Safran), S. Meynet (GDTech)

Ceramic core displacement and deformation during the casting process is a major source of cooled blades manufacturing scrap. A possible source of core deformation may be the fluidic forces due to the filling of the mold with the liquid alloy. Predictive numerical simulations of the casting process would be an essential asset to increase the efficiency of the conception and industrial processes. During the workshop, a numerical methodology to simulate the filling process was drawn, with several modelling levels (with or without surface tension and slipping-wall conditions), in order to estimate the relevance of each of these models. Numerical results were then compared to available experimental results. Numerical deformation of the core was approximated as a beam flexion. Despite this post-processing approximation, the correlation between experimental measurements and numerical simulations is satisfying. The evolution of the core displacement with the inlet velocity of the fluid also has the same behaviour in the experiments and in the simulation. Future work will aim at including the dynamic contact angles in the simulations, in order to evaluate the relevance of this finer modelling, as well as correlating simulations with experiments on cases more representative of the industrial process.

P10: Velocity regularization for Euler-Lagrange conversion - I. El Yamani (CORIA/Safran), M. Cailler (Safran), L. Voivenel, J. Carmona (CORIA)

The Euler Lagrange multi scale approach aims to reduce the computational costs when simulating two phase flow. To reduce the cost even more, more droplets have to be converted in the Lagrangian formalism where droplets are seen as point forces. It implies that droplets can not always check the hypothesis of the LPP (Lagrangian Particle Point) formalism which is that the diameter of the particle has to be much smaller than the cell size. This hypothesis allows to have a good approximation of the undisturbed velocity for the Lagrangian particle. If the hypothesis is not checked when a Eulerian droplet is converted into a Lagrangian particle a residual velocity field can exists and therefore the velocity given to the particle is impacted by itself. This project aims to filter the gaseous velocity field through a gaussian filtering to remove the contribution of the Eulerian droplet to better approximate the undisturbed velocity.

Combustion - K. Bioche, CORIA & R. Mercier, Safran

C1: Plasma discharge models for reacting system - S. Wang, B. Kruljevic, B. Fiorina (EM2C), Y. Bechane (CORIA)

To reduce the expensive computational cost of Plasma-Assisted Combustion (PAC) full 3D simulations, the EM2C laboratory has developed phenomenological approaches to model Nanosecond Repetitively Pulsed (NRP) plasma discharges in reacting flows (Castela 2016 & Blanchard 2023). As part of previous works and ECFDs, both models were implemented and validated in the Low-Mach number framework (YALES2-VDS). While they were also implemented in the Compressible framework (YALES2-ECS), the validation against existing measurements or computations remained. During the workshop, numerical simulations of pin-to-pin configurations were performed with different numerical schemes and reactive mixtures to validate both models in ECS. The energy deposition was relatively well-validated through 2D simulations in the conditions of Castela et al. CNF 2016 and Rusterholtz et al. JPhysD 2013. A glimpse of baroclinic instabilities was observed through 3D simulations in the conditions of Castela et al. PROCI 2017.

C4: Developement of an automated virtual scheme generator for CFD - T. Luu, M. Hustache, N. Darabiha, B. Fiorina (EM2C)

In reactive CFD simulations, a non-negligible part of the time cost is spent in the resolution of the chemical system. Simplified chemistry models aim to reduce the number of transported species while still ensuring a correct representation of the phenomena of interest. Among them, the virtual chemistry method consists of using “virtual” species and reactions to reproduce detailed chemistry results through a mechanism of drastically smaller size. These “virtual” species and reactions are optimized to target quantities of interest such as temperature, laminar flame speed or pollutants. In practice, the optimization is done using a learning database composed of representative canonical reactive configurations computed with detailed chemistry. The objective of this project was to develop a tool to easily generate virtual schemes. The tool, named VISION (Virtual Scheme optimizatION), is currently able to both generate a user-defined database of wide reactive configurations and optimize a given scheme structure using either CANTERA or REGATH.

C5: Partially-Stirred reactor model for MILD combustion - E. Stendardo, L. Bricteux (UMONS), M. Laignel, K. Bioche (CORIA), J. Blondeau (VUB)

MILD combustion produces intense turbulence and extensive reaction zones, necessitating costly mesh refinement over large areas. Practical mesh lacks precision, leading to sub-grid heterogeneity and turbulent fluctuations. A Partially Stirred Reactor model was implemented to address turbulence-combustion interaction. This model multiplies the source term by a limiter factor, allowing modelling of residence time in the inner cell reactive structure. Testing various limiter formulations based on mixing and chemical timescales revealed increased computational costs. Future work aims to reduce costs by utilizing the model only where necessary. This ongoing research seeks to optimize performance while minimizing computational overhead for efficient application in engineering scenarios.

C6: Static Mesh Adaptation for Hydrogen High pressure combustion using GPUs - G. Hexilar, C. Brunet, R. Mari, S. Richard (Safran), P. Pouech, Q. Douasbin, G. Staffelbach (Cerfacs)

This research project focuses on advancing the understanding of hydrogen combustion under high-pressure conditions (up to 10bars), employing an automated workflow coupled with static mesh adaptation to tailor computational simulations to specific requirements. The study aims to enhance the accuracy and efficiency of combustion models by utilizing both Central Processing Units (CPUs) and Graphics Processing Units (GPUs). The automated workflow streamlines the simulation process, optimizing resource utilization and minimizing manual intervention. Static mesh adaptation further refines the computational mesh based on evolving combustion dynamics, ensuring accurate representation of high-pressure hydrogen combustion phenomena. By leveraging the parallel processing capabilities of GPUs alongside traditional CPUs, the research team aims to achieve significant computational speedup. This innovative approach not only contributes to fundamental insights into high-pressure hydrogen combustion but also establishes a robust framework for scalable and efficient simulations in complex reactive flow scenarios.

C7: High fidelity simulation of a cone calorimeter - A. Grenouilloux, K. Bioche (CORIA), N. Dellinger (ONERA), R. Letournel (Safran)

A methodology to simulate the decomposition of a composite sample in a calorimeter cone has been proposed. The configuration consists in the imposition of an incident radiative flux that heats the test coupon until it degrades. During test campaigns, the composite degradation can lead to the auto-ignition of the outgassed species, a phenomenon that needs to be predicted by the simulation. The variety of physical phenomena encountered, as well as the different characteristic time-scales, require the implementation of a coupled simulation. Hence, the proposed methodology relies on the coupling between two solvers of the massively parallel library YALES2 (fluid, radiation) and the MoDeThec solver developed at ONERA (solid degradation). First, a set of elementary validation tests to characterize the composite’s properties has been performed, showing good agreement with experimental data. A reduced three-equation kinetic scheme for the ignition delay has been derived, which aligns with experimental observations. Additionally, the framework for the coupled simulation involving the three solvers has been implemented.

User Experience & Data - L. Korzeczek, GDTech

U1: Refactoring the YALES2 tools - J. Leparoux, M. Cailler (Safran), L. Voivenel, J. Carmona, I. El Yamani (CORIA), S. Meynet, L. Korzeczek (GDTech)

The YALES2 distribution "tools" are becoming difficult to read and are a mixture of several types of tools. This leaves developers and end users unaware of what exists and how to use it (duplicated functions or tools) and makes it impossible to propose generic data analysis tools (FFT, confidence intervals, ...) that can be easily applied to YALES2 data structures. The main efforts have been concentrated on promoting a new architecture for YALES2 distribution tools with an object-oriented structure, including a refactored version of the main readers of YALES2 data. Several tutorials using Jupyter notebooks have been published for demonstration and explanation. A new CLI is now available under the name 'y2tools'. More work is needed before this structure can be pushed to the master trunk.

U2: Improved USEX for Multi-Scale Eulerian-Lagrangian simulation - L. Voivenel, J. Carmona, I. El Yamani (CORIA) J. Leparoux, M. Cailler (Safran)

The multi-scale Eulerian-Lagrangian approach has now reached a certain maturity and is being used to simulate fuel spray atomization. Post-treatments of these multi-scale simulations require the development of specific tools that track liquid structures either described in an Eulerian or Lagrangian way. In this project, we implemented a strategy to register in a post-treatment particle-set all Eulerian droplets crossing an arbitrarily shaped surface (described with an interior-boundary). The strategy is based on artificial Eulerian droplet advancement (using a Lagrangian representation) and verification of the new Eulerian droplet position compared to the surface of interest. We used this strategy to build a new post-treatment that allows to track both Eulerian and Lagrangian structures and build particle size or velocity distributions.

U3: Evaluate technological debt - P. Pouech, T. Marzlin, A. Dauptain (CERFACS)

The tool dedicated to cleaning and extracting meta data from the codebase (Tucan) is now able to interpret the most common #Ifdef statements for C and Fortran. Is has be tested successfully on the codebases of Yales2 , Neko and some elements of the Chemkin II package, for Fortran language. It has also included more tests, especially on C++ samples and large files. The call-graph aggregator “Marauder’s map” also evolved a lot. It was used as a refactoring monitoring tool for the UX project U1 “Refactoring Yales2 tools”, on a codebase mixing Python and Fortran files, providing feedback to both U3 and U1 teams. With this experience, we plan to add at least two additional complexity metrics in the months to come : the single component #Ifdef footprint , and the Custom Structures footprint : declarations vs usage in the code.

U4: CWIPI 1.0 porting - N. Dellinger, B. Andrieu, K. Hoogveld, E. Quémerais (ONERA), A. Grenouilloux (CORIA), R. Letournel (Safran)

Coupling is a cornerstone of numerical simulation, especially for addressing multi-physics problems using highly-specialized solvers for each phenomenon. The CWIPI library, developed at ONERA for coupling codes in a massively parallel environment, has been used in YALES2 for many years for internal and external coupling. Significant developments have been carried out in recent years to improve the performance and usability of CWIPI, resulting in the release of version 1 in july 2023. This version features a completely revised API to overcome the limitations of version 0.12 and offer more possibilities to users. The goal of this project was to support users in their transition to version 1. A training course based on Jupyter Notebooks was first organized. Assistance was then provided to successfully port MoDeTheC's and YALES2's internal couplings to the new version. Some fixes were made in CWIPI along the way, and will be reported in a new patched version.

U5: Integration of YALES2 in PRESTO supervisor - A. Pushkarev (GE Vernova), G. Balarac (LEGI)

A Graphical User Interface (GUI) exists at GE Vernova (Hydro) provides a user-friendly solution to perform numerical simulations for typical geometries of hydroelectric turbine in varied operation regimes. Previously, we implemented an interface for YALES2 code as alternative of CFX solver for this GUI client. Actual project is dedicated to implementation of the automatic mesh generation process for the runner section of the turbine using only section profile files of geometry such as blade profiles, meridional channel section, guide vane profile, etc... The algorithm should be able to generate a new *.msh mesh file once geometry profiles are updated as well as to setup standard named sections of the numerical domain.

U6: Optimization of YALES2 compilation time - R. Mercier (Safran), G. Lartigue (Total Energy)

YALES2 Compilation time are very different depending on the compilation options but also depending on the machine where the code is built. In particular, compiling on cluster always takes more time than on local station because of the slower filesystems. On these machines, the compilation time scales with the number of .f90 files to compile. 1) During this project we proposed a detailed compilation timing system activated with Y2_COMPILE_TIMER = TRUE and plotted using the tool y2_compilation_gantt.py. 2) This new profiling tool allowed us to spot some very large modules including a lot a dependencies. To enhance the tasks parallelization, the expl_comp_numerics module has been spliced in several modules. An important gain in compilation time has been obtained. 3) Another optimization of the compilation time have been developed using an on-the-fly modification of the .f90 tree. The use of _h.f90 is dynamically removed and replaced by including equivalent .defs files allowing the divide by 2 the number of .f90 files to compile. Some demonstrations have been performed but this compilation mode still have to be cross-check with different compilation option to ensure its robustness. At the end of the project, the sum of all optimization allowed us to divide by 2 the compilation time.