Martin David

Pau, 64000 - France martin.david@univ-perp.fr

CV:



Introduction

I am currently an Associate Professor at the LAMPS laboratory, University of Perpignan Via Domitia, France. My research focuses on turbulence modelling. With expertise in Computational Fluid Dynamics (CFD), my background includes extensive work on turbulent flows for low-carbon energy systems. I previously served postdoctoral researcher at the Inria Institute of Bordeaux under the supervision of Rémi Manceau and studying self-adaptive approaches for hybrid RANS/LES. I was also a temporary teaching and research assistant at the PROMES laboratory (University of Perpignan Via Domitia), where I investigated and developed turbulence models for Large Eddy Simulation (LES), specifically targeting highly anisothermal flows found in solar receivers. My Ph.D., supervised by Prof. Françoise Bataille and Assoc. Prof. Adrien Toutant, was dedicated to the simulation and modeling of flows in gas-pressurized solar receivers, employing a multi-fidelity approach.

I am eager to dedicate my future work to the challenge of certified CFD, namely user-independent CFD with quantified uncertainties that is suitable to replace experimental tests in certification processes. I am particularly interested in deepening the self-adaptive approaches, investigating uncertainty quantification in CFD, and performing multi-fidelity CFD-based optimization. In addition, I am willing to explore the capabilities of machine learning to complement traditional turbulence modeling with physics-informed approaches and I am eager to investigate multiphase flows and fluid-structure interactions.



Research interests

  • Computational Fluid Dynamics (Incompressible, Low-Mach)
  • DNS, T-LES, LES, hybrid RANS/LES, RANS
  • Forced/Natural convection
  • Turbulence/heat fluxes modeling
  • Self-adaptive computations
  • Uncertainty quantification
  • Multi-fidelity approaches
  • Education

  • PhD - Univ. of Perpignan Via Domitia, PROMES - CNRS (2018 - 2021)
  • Modelisation and simulation of flows in gas-pressurized solar receivers.

  • Engineering school ENSGTI (2015 - 2018)
  • Specialized in industrial energy systems


    Experience

    Associate Professor

    University of Perpignan Via Domitia, LAMPS laboratory

    Toward certified CFD.

    Oct. 2022 - 2024

    Postdoctoral student

    National Institute for Research in Computer Science and Control

    Investigation of self-adaptive strategy based on physical criteria for hybrid RANS/LES.

    Oct. 2022 - 2024

    Temporary teaching and research assistant

    University of Perpignan Via Domitia - PROMES - CNRS

    Development of turbulence model of Thermal-LES in the operating conditions of gas-pressurized solar receivers.

    Sept. 2021 - Sept. 2022

    Ph.D. Student

    University of Perpignan Via Domitia - PROMES - CNRS

    Modelisation and simulation of flows in gas-pressurized solar receivers.

    Oct. 2018 - Sept. 2021

    Engineering Internship

    EOLINK

    Optimizing the placement of windturbines in offshore windfarms.

    Apr. 2018 - Sept. 2018

    Research Internship

    University of La Réunion - PIMENT

    Numerical modeling of passive smoke extraction in buildings.

    June 2017 - Sept. 2017

    Research

    My work is generally devoted to the modeling of turbulent flows, the simulation and the analysis of turbulence/heat transfer couplings, and development of self-adaptive approaches. Some of my research works are presented below:

    Self-adaptive strategy

    Postdoctoral student

    In continuous hybrid RANS/LES approaches, the location of the switching between the RANS and LES modes is based on the mesh and has a significant impact on the results. The development of a self-adaptive strategy to mitigate the influence of the user's meshing choices on the results is then of primary interest. In this work, a physical criterion is used to determine where to switch between RANS and LES and provides the cell sizes in the LES region. The proposed method is applied to the backward-facing step with the Hybrid Temporal LES (HTLES) model. Notice that the approach is non-intrusive and applicable to any other hybrid approach. The algorithm of the method is illustrated below.

    ...
    Algorithm used in the self-adaptive approach.

    The proposed method rapidly converges and, when associated with the HTLES model, significantly improves the RANS results. The LES region, and thus the generation of resolved structures, is automatically restricted to the region of interest as can be seen in the following animation:



    Multi-fidelity approach for solar receivers flows

    PhD

    The flows encountered in receivers of CSP involve characteristic size and time of different orders of magnitudes and strong interactions between the dynamics and the heat transfer. For these reasons, approaches at different levels are needed. In this work, a multi-fidelity approach has been conducted to produce a thorough analysis of these challenging flows.

    ...
    Investigation of solar receiver flows with multi-fidelity approach.

    The DNS produces high-fidelity results with a very thin level of description, deepening the flow comprehension and providing reference data for the lower-fidelity computations. However, these simulations are very costly and can only be performed at the local scale: the DNS of a channel flow at the turbulence level encountered in solar receivers requires up to 4.5 million hours on a supercomputer. The T-LES performed in the studied conditions needs about 20000 hours, namely 200 times less than the DNS. The geometry studied with the T-LES was not more complex than the one studied in DNS but the LES offered the possibility of investigating a large number of operating conditions. A heat transfer correlation has then been developed based on the results of 70 T-LES, enabling heat transfer computation in the entire solar receiver.



    Study of the coupling between dynamics and heat transfers using Direct Numerical Simulations

    PhD

    Solar receiver flows are very complex since the heating fluid is strongly turbulent and asymmetrically heated at high-temperature levels. To deepen the understanding of these flows, and ultimately improve the receiver performance by increasing the heat transfer and reducing the wall friction, DNS have been performed in different fluid heating configurations.

    Vizualization of the coupling between dynamics and heat transfers at the wall. From David et al. (2023)

    The above figure highlights the strong correlation between the wall heat fluxes and the wall frictions. Low heat fluxes and wall frictions are associated with elongated patterns reflecting the turbulent structures in the near-wall region. The ratio between the two quantities shows that, generally, the biggest values of the wall heat flux to wall friction ratio involve patterns elongated in the transverse direction. Further investigations point out a range of normalized sizes of the pattern to promote aiming to increase this ratio. Notice that the patterns are smaller at the cold wall (right pictures) due to the higher turbulence level.



    Development of turbulence models of T-LES in the operating conditions of solar receivers

    Temporal teaching and research assistant

    This work deals with Thermal Large-Eddy Simulation (T-LES) of anisothermal turbulent channel flow in the working conditions of solar receivers used in concentrated solar power towers. The flow is characterized by high-temperature levels and strong heat fluxes. The hot and cold friction Reynolds numbers of the simulations are respectively 630 and 970.

    Assessment of T-LES on the turbulent heat fluxes of the hot and cold walls. The labels "nomodel" corresponds to the simulation performed without model, "A03-A03" corresponds to the simulation performed with the AMD model for both the eddy-viscosity and the eddy-diffusivity, "A03-As03" corresponds to the simulation performed with the AMD model for the eddy-viscosity and the scalar AMD model for the eddy-diffusivity. The two last labels ("Ac03-Ac03" and "A06+B05-A06+B04 2L") correspond to newly proposed models, respectively an adaptation of the classical AMD closure to compressible flows and a two-layer mixed model.

    The results stress that the proposed models perform better on the prediction of the streamwise turbulent heat flux. Specifically, the two-layer mixed model is in good agreement with DNS results. The influence of the mesh variations on the proposed two-layer mixed model is displayed below. The AAA, BAB, and CAC meshes consist of 3.7, 2.3, and 1.4 million cells, respectively. The results point out the weak impact of the mesh on the performance of the two-layer mixed model.

    Influence of the mesh on the prediction of the normalized temperature profile, assement of the two layer mixed model.


    Development of a heat transfer correlation for gas-pressurized solar receivers

    PhD

    Heat transfer correlations are widely utilized in industry because of their simplicity of use. However, their applicability domain is often very restrictive. We developed a heat transfer correlation dedicated to the operating conditions of solar receivers. The development of this correlation relied on the execution of numerous T-LES across a wide range of configurations, encompassing scenarios with both symmetric and asymmetric fluid heating. As shown in the following figure, the proposed correlation significantly improves the heat transfer prediction when compared to existing correlations (particularly in asymetric heating conditions) and provides an estimate of less than 10% error in all studied configurations.

    Assessment of three correlations of the literature and the proposed one in the operating conditions of solar receivers. "PC" stands for Proposed Correlation. From David et al. (2021)


    Error propagation of flow parameters on wall heat transfers in solar receivers

    PhD

    Uncertainties may skew the understanding of the results of experiments or numerical simulations. The impact of the flow parameters uncertainties of measurement on wall heat flux is investigated in anisothermal turbulent channel flow. The proposed heat transfer correlation dedicated to gas-pressurized solar receivers of concentrated solar tower power is used. The Guide to the expression of Uncertainty in Measurement (GUM) is applied and provides an analytical expression of the uncertainty propagation. Assuming the quasi-normality and quasi-linearity of the studied function, this method provides approximate results. The results obtained following the methodology described in the GUM are then compared to the direct computation of the wall heat flux with altered temperatures.

    Errors in the wall flux depending on the bulk-to-wall temperature ratio and the error of the measurements of the wall temperatures in the case of symmetric heating conditions. The left graph shows the reference results, and the right graph exposes the estimation produced by the GUM. The lines indicate the isovalues of error 2%, 4%, 8%, 12%, 16%, 20%, 30%, 40%, 60%, 80%, and 100%. From David et al. (2021)

    The figure shows that the wall heat flux is strongly impacted by the wall temperature error of measurement. An under-estimation of the wall temperature leads to slightly higher uncertainty on the wall heat flux than an overestimation. The model provides an accurate prediction of the error on the entire validity domain of the correlation.


    Supervision activities

    I participated to the supervision of three students at different stage of their education, with research topics related to multi-phase flows and artificial intelligence (AI). They are listed below.

    • Mélanie Dreina, PhD student - Compact schemes & high-performance computing - application to flow simulation.
    • Yanis Zatout, PhD student - Simulation and learning of strongly anisothermal turbulent flows.
    • Bouziane Boudraa, PhD student - Large eddy simulations of a turbulent flow with hybrid nanofluid subjected to asymmetric heating.
    • Yazid Boucetta, Master 2 internship - Turbulence modeling using artificial intelligence.

    Publications

    All my publications are visible on Google Scholar, HAL Science Ouverte, and Research Gate.

    International journals
    1. M. David, M. Mehta, and R. Manceau, On the feasibility of a self-adaptive strategy for hybrid RANS/LES based on physical criteria and its inital testing on low Reynolds number backward-facing step flow,, Flow, Turbulence and Combustion, InPress, 2024.
    2. B. Boudraa, M. David, A. Toutant, F. Bataille, and R. Bessaih, Large eddy simulations of a turbulent flow with hybrid nanofluid subjected to symmetric and asymmetric heating, International Journal of Heat and Fluid Flow, vol. 107, p. 109338, Jul. 2024.
    3. M. David, A. Toutant, et F. Bataille, Thermal Large-Eddy Simulations methods to model highly anisothermal and turbulent flows, Physics of Fluids, vol 35, p. 035106, Mar. 2023.
    4. M. David, A. Toutant, et F. Bataille, Study of asymmetrically heated flows passing through gas-pressurized solar receivers using Direct Numerical Simulations, International Journal of Heat and Mass Transfer, vol. 201, p. 123577, Feb. 2023.
    5. M. David, A. Toutant, and F. Bataille, Impact of asymmetrical heating on the uncertainty propagation of flow parameters on wall heat transfers in solar receivers, Applied Thermal Engineering, vol. 199, p. 117547, Nov. 2021.
    6. M. David, A. Toutant, and F. Bataille, Direct simulations and subgrid modeling of turbulent channel flows asymmetrically heated from both walls, Physics of Fluids, vol. 33, no. 8, p. 085111, Aug. 2021.
    7. M. David, A. Toutant, and F. Bataille, Investigation of thermal large-eddy simulation approaches in a highly turbulent channel flow submitted to strong asymmetric heating, Physics of Fluids, vol. 33, no. 4, p. 045104, Apr. 2021.
    8. M. David, A. Toutant, and F. Bataille, Numerical development of heat transfer correlation in asymmetrically heated turbulent channel flow , International Journal of Heat and Mass Transfer, vol. 164, p. 120599, Jan. 2021.
    International conferences
    1. M. David, M. Mehta, and R. Manceau, Towards self-adaptivity in hybrid RANS/LES based on physical criteria presented at the 10th international symposium on Turbulence, Heat and Mass Transfer, Rome, Italy, Sep. 2023.
    2. A.Toutant, M. David, Y. Zatout, F. Bataille, L. Mathelin, O. Semeraro, Thermal large eddy simulations for high temperature solar receivers presented at the 17th International Heat Transfer Conference, Cape Town, South Africa, Aug. 2023.
    3. M. David, A. Toutant, and F. Bataille, Assessment and comparison of Large Eddy Simulations in asymmetrically heated and highly turbulent channel flows presented at the 13th International ERCOFTAC Symposium on Engineering Turbulence Modelling and Measurements, Rhodes, Greece, Sep. 2021.
    4. M. David, A. Toutant, and F. Bataille, Sensitivity analysis of heat transfers in an asymmetrically heated turbulent channel flow presented at the 13th International Conference on Computational Heat Mass and Momentum Transfer, Paris, France, May 2021.
    5. M. David, A. Toutant, and F. Bataille, Study of anisothermal channel flow physics with direct numerical simulations presented at the 28th Franco-Polish Mechanics Seminars, Perpignan, France, Oct 2021.
    National journals
    1. M. David, A. Toutant, and F. Bataille, Analyse de sensibilité des flux de chaleur pariétaux en canal asymétriquement chauffé au moyen d’une corrélation permettant d’estimer les transferts de chaleur, Entropie, vol. 2, no. 2, Nov. 2021.
    2. M. David, A. Toutant, and F. Bataille, Evaluation de modèles de simulation des grandes échelles en canal plan turbulent chauffé asymétriquement : de la comparaison des grandeurs intégrales à l’analyse des corrélations, Entropie, vol. 1, no. 4, Nov. 2020.
    National conferences
    1. M. David, A. Toutant, and F. Bataille, Tests et améliorations de modèles mixtes de Simulation des Grandes Échelles d’un écoulement à fort nombre de Reynolds en canal asymétriquement chauffé, presented at the SFT, Valencienne, France, Jun. 2022.
    2. M. David, A. Toutant, and F. Bataille, Développement et analyse d’une corrélation pour estimer les transferts de chaleur en situation de fort chauffage asymétrique d’un écoulement en canal., presented at the SFT, Belfort, France, Jun. 2021.
    3. M. David, A. Toutant, and F. Bataille, Tests a posteriori de modèles de sous-mailles dans un écoulement en canal plan à haut nombre de Reynolds et soumis à un fort flux de chaleur, presented at the SFT, Belfort, France, Jun. 2020. (Cancelled due to Covid)
    Research reports
    1. M. David, Démarche multi-échelle pour les récepteurs solaires : de la physique des couplages température/turbulence au développement de corrélation, Rapport d’activité GENCI, Principaux résultats scientifiques, 2021
    Thesis
      M. David, A. Toutant, and F. Bataille, Simulation et modélisation des écoulements dans les récepteurs solaires à gaz sous-pression, PhD thesis, University of Perpignan Via Domitia, 2021
    Public participation
    1. M. David, R. Manceau, Demonstrator of self-adaptive hybrid RANS/LES based on physical criteria: application to the backward-facing step case, External seminar - IHCantabria - invited by Prof. J. Lopez Lara, Nov. 2023
    2. M. David, A. Toutant, and F. Bataille, Modélisation des écoulements dans les récepteurs solaires par Simulation des Grandes Echelles, Internal seminar - PROMES laboratory , 2018
    3. M. David, A. Toutant, and F. Bataille, Analyse du couplage entre la dynamique et les transferts de chaleur par Simulation Numérique Directe, Internal seminar - PROMES laboratory , 2021
    Submitted publications

    Teachings

    Since the beginning of my doctoral thesis, I taught more than 400 hours divided into lectures, tutorials, practical exercices, and projects. My experience includes teachings from L1 to M2, in engineering schools (Sup'EnR and ENSGTI), universities (Perpignan and Pau), and University Institute of Technology (Perpignan). They are mainly related to

    • CFD,
    • Thermodyamics,
    • Vibrational Analysis,
    • Fluid mechanics,
    • Numerical methods,
    • Heat exchangers,
    • Radiative transfers,
    • Renewable energies in general,
    • Wind power.

    ...             ...             ...             ...             ...

    Awards / Visibility

    During my PhD and temporary teaching and research assistant appointements, I obtained a total of 26 millions of computational hours on National (DARI) and European (PRACE) project's call. The project and the associated results were selected to be part of the GENCI annual report (see chapter "Principaux résultats" here).