This is an extensive description of my background (always under construction). Multi-language (English, French and Portuguese) compact versions are available on demand. This page was last modified in October 2024.
PROFESSIONAL EXPERIENCES
Research Engineer – IFP Energie Nouvelles
Rueil-Malmaison, France
Research in the field of wind turbines aerodynamics.
Developer of Blade Element Momentum (BEM) and vortex-based modules based in C++ and CUDA for modelling vertical and horizontal wind turbines. Implementation of new features, maintenance and support to users. Investigation and proposal of methods and algorithms to improve the physical representability and the computational performance based in state-of-the-art literature.
Computational Fluid Dynamics Engineer – CS Group
Toulouse, France
Development and implementation of new features to the physics module of the Computational Fluid Dynamics (CFD) software ProLB, based on the Lattice-Boltzmann method (LBM). Implementation and improvement of academic and industrial benchmarks for code validation, participation in validation campaigns. Interaction with scientific and industrial project partners. Object oriented programming in C++ and Python.
Postdoctoral Researcher – CERFACS
Toulouse, France
Implementation and validation of new features in the high-order fluids dynamics solver Jaguar based on the spectral-differences method in the framework of a European project. DNS (Direct Numerical Simulations) and LES (Large-eddy simulations) simulations of aerodynamic benchmarks.
Postdoctoral Researcher – ISAE-SUPAERO
Toulouse, France
Research on the use of deep-learning techniques for the modeling of direct and reverse (source localization) of acoustic waves in complex media. Use of pyTorch framework on GPU’s, database generation using LBM framework Palabos and finitive-volumes framework OpenFOAM.
Produced open-source libraries and repositories:
– acoustic beamforming library: beamlib
– multiphysical convolutional neural network reproducibility study: cnn-multiphysics-repro
– convolutional neural network for beamforming maps deconvolution: deconvbfnet
Simulation Engineer – Zelin
Toulouse, France
Simulation Engineer in engineering consultant company specialized in Fluid Mechanics. Extensive study of the literature on the on-road aerodynamic conditions for automotive industry clients aiming at the correct estimation of the fuel consumption and pollutant emissions. Simulation of automotive aeroacoustics benchmarks (full-vehicle, rear-mirror).
Doctoral Researcher – PPRIME Institute
Poitiers, France
Study of airframe noise. Numerical evaluation of obstacle noise using hybrid aeroacoustical approach: DNS incompressible flow (incompact3d solver with Immersed Boundaries Method – IBM) and Curle’s analogy. Shape optimization using Particle Swarm Optimization (PSO) and regression analysis using R. Simulations performed in HPC facility Mésocentre SPIN Calcul, France. Experimental testing of the sound emission and flow characteristics of prisms with different cross-section using hot-wire anemometry and microphone recordings, measurements are performed in the anechoic wind-tunnel BETI at Poitiers, France.
External Aerodynamic Intern – Stellantis
(former Groupe PSA)
Vélizy-Villacoublay, France
Numerical study of passive drag reduction devices in automotive wheel housing using an LBM commercial solver. Model modifications in CATIA. Pre (geometry cleaning, surface and volume meshing) to post-processing steps performed within the PowerFLOW suite.
Research intern – PPRIME Institute
Poitiers, France
Conception and preliminary testing of an experimental setup for the generation of intermittent aeroacoustical sources in an anechoic wind tunnel. Application of acoustic antenna methods (beamforming), spectral analysis and wavelet transform. Performed under the OpenLab Fluidics cooperation program with PSA Peugeot Citröen.
ACADEMIC BACKGROUND
Engineering Degree – Aeronautical Engineer
Federal University of Uberlândia (UFU)
Uberlândia, Brazil
5 years program, focused on airplanes conception. Principal topics: Thermodynamic, Fluids and Solid Mechanics, Material Science, Fabrication and Aeronautics.
Final thesis: Numerical and experimental analysis of flow around pick-up vehicles. Complete model design and CAD conception in CATIA; RANS simulations using Star-CCM+. Hot-wire measurements and tufts flow visualization.
Master Degree – “Transports Aéronautiques et Terrestres”
École Nationale Supérieure de Mécanique et d’Aérotechnique (ISAE-ENSMA)
Chasseneuil-du-Poitou, France
French Engineering School specialized in Aeronautics, participated in 2 of the 3 years program, major in Aerodynamics. Principal topics: Compressible Aerodynamics, Combustion, Thermal Modeling, Turbulence, Turbomachinery, Aeroacoustics, Numerical methods for Aerodynamics.
PUBLICATIONS & CONFERENCES
2023
Margnat, F., Gonçalves da Silva Pinto, W. J. & Noûs, C. (2023). Cylinder aeroacoustics: experimental study of the influence of cross-section shape on spanwise coherence length. Acta Acustica, 7. – 10.1051/aacus/2022061 – paper – supplementary material
abstract
New data and review of the spanwise coherence length is provided for flows over cylinders of different cross-sections: circular of diameter d, and rectangular of sectional aspect ratios (breadth (b) to height (d) ratio AR = b/d) of 1, 2 and 3. In the present measurements, the body has both d and spanwise length of 70d fixed, and the Reynolds number (based on d) range 6000–27,000 is covered. Two-point data are obtained from two hot-wire probes, one fixed in the symmetry plane and the other moving on the corresponding spanwise axis. Their position in a cross plane are deduced from preliminary measurement of the mean flow with a single probe, allowing fair comparisons between the different geometries and the introduction of uncertainty bars on coherence length values. At all tested regimes, a very good agreement is noticed between velocity-based and pressure-based coherence experimental data. Coherence length definitions are revisited, and the aeroacoustically consistent, integral length definition is selected, allowing fair synthesis of literature data into a single chart and empirical functions. Definitions for coherence decay models (e.g. Gaussian or Laplacian) are also adapted so that coherence length and coherence integral shall be equivalent. This preliminary work on coherence data and its spanwise integration enables transparent regressions and model selection. Generally, the Gaussian model is relevant for the lift peak, while the coherence exhibits a Laplacian decay at harmonics. On average, at peak Strouhal number, the coherence length for the circular and square cylinders is of 5d while it is of the order of 15d for the rectangular sections. It is concluded that the flow over those latter geometries is still a two-dimensional dynamics at the tone frequency. These values are almost preserved over the tested Reynolds number range. Coherence length value at harmonics is extensively documented. Spanwise coherence length is also discussed as an ingredient of acoustic efficiency.
2022
Pinto, W., Alguacil, A. & Bauerheim, M. (2022). On the reproducibility of fully convolutional neural networks for modeling time–space-evolving physical systems. Data-Centric Engineering, 3, E19. – 10.1017/dce.2022.18 – paper – repository
abstract
Reproducibility of a deep-learning fully convolutional neural network is evaluated by training several times the same network on identical conditions (database, hyperparameters, and hardware) with nondeterministic graphics processing unit operations. The network is trained to model three typical time–space-evolving physical systems in two dimensions: heat, Burgers’, and wave equations. The behavior of the networks is evaluated on both recursive and nonrecursive tasks. Significant changes in models’ properties (weights and feature fields) are observed. When tested on various benchmarks, these models systematically return estimations with a high level of deviation, especially for the recurrent analysis which strongly amplifies variability due to the nondeterminism. Trainings performed with double floating-point precision provide slightly better estimations and a significant reduction of the variability of both the network parameters and its testing error range.
2021
Pinto, W. G., Michaël Bauerheim, M. & Hélène Parisot-Dupuis, H. Deconvoluting acoustic beamforming maps with a deep neural network. INTER-NOISE and NOISE-CON Congress and Conference
Proceedings (2021) 263(1):5397–5408 – https://doi.org/10.3397/in-2021-3084 – repository
abstract
Localization and quantification of noise sources is an important scientific and industrial problem, the use of phased arrays of microphones being the standard techniques in many applications. Non-physical artifacts appears on the output due to the nature of the method, thus, a supplementary step known as deconvolution is often performed. The use of data-driven machine learning can be a candidate to solve such problem. Neural networks can be extremely advantageous since no hypothesis concerning the environment or the characteristics of the sources are necessary, different from classical deconvolution techniques. Information on the acoustic propagation is implicitly extracted from pairs of source-output maps. On this work, a convolutional neural network is trained to deconvolute the beamforming map obtained from synthetic data simulating the response of an array of microphones. Quality of the estimation and the computational cost are compared to those of classical deconvolution methods (DAMAS, CLEAN-SC). Constraints associated with the size of the dataset used for training the neural network are also investigated and presented.
Alguacil, A., Pinto, W. G., Bauerheim, M., Jacob, M. C. & Moreau, S. Effects of Boundary Conditions in Fully Convolutional Networks for Learning Spatio-Temporal Dynamics. Machine Learning and Knowledge Discovery in Databases. Applied Data Science Track (2021) Springer International Publishing p. 102-117 – https://doi.org/10.1007/978-3-030-86517-7_7 – paper
abstract
Accurate modeling of boundary conditions is crucial in computational physics. The ever increasing use of neural networks as surrogates for physics-related problems calls for an improved understanding of boundary condition treatment, and its influence on the network accuracy. In this paper, several strategies to impose boundary conditions (namely padding, improved spatial context, and explicit encoding of physical boundaries) are investigated in the context of fully convolutional networks applied to recurrent tasks. These strategies are evaluated on two spatio-temporal evolving problems modeled by partial differential equations: the 2D propagation of acoustic waves (hyperbolic PDE) and the heat equation (parabolic PDE). Results reveal a high sensitivity of both accuracy and stability on the boundary implementation in such recurrent tasks. It is then demonstrated that the choice of the optimal padding strategy is directly linked to the data semantics. Furthermore, the inclusion of additional input spatial context or explicit physics-based rules allows a better handling of boundaries in particular for large number of recurrences, resulting in more robust and stable neural networks, while facilitating the design and versatility of such types of networks. (Datasets, code and supplementary material are available at
gitlab.isae-supaero.fr/a.alguacil/boundary_conditions_fcn_dyn).
Pinto, W. G. & Margnat, F. Influence of cylinder breadth and shape on the onset of flow unsteadiness and the aeolian tone level. Computers & Fluids (2021), 105067 – https://doi.org/10.1016/j.compfluid.2021.105067 – accepted manuscript
abstract
The Reynolds number for the onset of flow unsteadiness is determined for several canonical geometries (triangles, rectangles, ellipses and lozenges) at different sectional breadth (L) to height (d) ratios (aspect ratio AR=L∕d), for more than 70 shapes. The flow is modeled using a direct Navier–Stokes incompressible two-dimensional solver and the shape is defined by an Immersed Boundary Method. The employed procedure takes the fluctuation of the velocity in the wake as the criterion to define the unsteadiness and a binary search to find the transition. This procedure yields critical Reynolds number Rec values in agreement with available data in the literature. When AR approaches zero, the five shapes lead to almost the same value of 31, which corresponds to Rec for a flat plate normal to the flow. It is then found that Rec grows exponentially with the aspect ratio, the influence of the cross section shape being accounted for by a single regression parameter. For all aspect ratios, the ellipse exhibits the highest Rec, and the front-pointing triangle the lowest, the three other geometries laying in between those two. The physics of the influence of cross-section shape on Rec is analysed, considering its link with recirculation length in particular. An exploitation of the results is outlined for the analysis of recent aeroacoustic shape optimizations at fixed Re=150, through correlation between the lift fluctuation at this regime with the distance to the onset of unsteadiness it corresponds to.
Pinto, W. J. G. da S., Margnat, F. & Noûs, C. Influence of the length of a cylinder on its aeolian tone level: measurement and modelling 14th World Congress on Computational Mechanics (WCCM ECCOMAS Congress 2020), 11-15 Jan 2021, virtual, 2021 – paper
abstract
The tonal noise generated by the flow over elongated cylinders is measured using microphones in the anechoic wind tunnel BETI of Institut Pprime. The effect of the length to diameter ratio of the cylinder is assessed by varying the diameter from 6 mm to 20 mm within the constant width (750mm), open jet, test section. The velocity is varied from 10 to 40 m/s, leading to a Reynolds number range of about 4,000 – 53,000. A proper normalization is needed to obtain a good collapse of experimental data from nearly 10 studies on the same evolution in 3 steps: for very short cylinders, the tone level does not depend on the length; for semi-long cylinders, it follows the fourth power of the length to diameter ratio; for very long cylinder, the length’s influence vanishes in the form of a sound level asymptote. This sigmoidal evolution is noticed for both the circular section and the square section cylinders, and questions the classical modelling of the aeroacoustic process using compact, coherent segments of cylinder associated with a coherence length to account for spanwise phase loss. Literature data from numerical simulation or experiment using end-plates are included in the analysis too.
2020
Pinto, W. J. G. da S., Margnat, F. & Noûs, C. Experimental study of the influence of the shape on the spanwise coherence length of the flow over a bluff body e-Forum Acousticum FA2020, 20-24 Dec 2020, virtual, 2020 – paper
abstract
The coherence length is a key ingredient in the efficiency of the aeolian tone generated by the flow over a bluff body, and is often used to extrapolate noise simulations of a short segment of cylinder for comparison with measurements using longer bodies. In the present study, the influence of the cylinder cross-section on the coherence length is studied experimentally in BETI wind tunnel, by hot-wire anemometry of the flow over a circular, a square and two rectangular cylinders, at Reynolds number from 6,700 to 27,000. The coherence length estimation method is first validated for the circular cylinder case, in agreement with values obtained from wall-pressure measurement in the literature. The mean flow is then scanned for the rectangular sections, and the locations of the probes for the spanwise study are based on its topology, such as the position of maximum velocity. This ensures fair comparisons between the geometries, while, finally, having no qualitative influence on the coherence length value. For the square cylinder, the coherence length at the peak frequency is close to, though a little higher than, that of the circular cylinder. For the rectangular cylinder, coherence lengths up to 20 diameters are measured at the peak frequency. Considering the length of the model and the size of the test-section, these high values leads to the conclusion that the shear-layer flapping is a 2D dynamics at these regimes.
Pinto, W. J. G. da S., Margnat, F. & Noûs, C. Aeolian tone level as a consequence of distance to the onset of unsteadiness: influence of bluff-body shape on the critical Reynolds number e-Forum Acousticum FA2020, 20-24 Dec 2020, virtual, 2020 – paper
abstract
In order to interpret recent results of shape optimization for the aeolian tone problem in 2D, laminar flows, the Reynolds number corresponding to the onset of unsteadiness in a flow over a bluff body is estimated for a set of shapes : five canonical shapes (the ellipse, the rectangle, the losenge and two triangles) are studied, as well as the influence of their aspect ratio that is varied from 0.2 to 2. The critical Reynolds number (Rec) is obtained by binary search within direct numerical solutions of the Navier-Stokes equations, using Incompact3D solver with an immersed boundary method. A criterion is designed on purpose to discriminate between steady and unsteady solutions. This methodology yields critical Reynolds number values in good agreement with literature results. For all the shapes, Rec increases exponentially with the aspect ratio. When the latter tends to zero, the evolution for all the shapes merge around Rec ≈ 31, which is that of a flat plate normal to the flow. The ranking of shapes using their Rec provides insight of the different levels of unsteadiness, and thus of radiated noise, that are noticed between bluff-body flows when analysed at a fixed Reynolds number.
da Silva Pinto, W. J. G. & Margnat, F. Shape optimization for the noise induced by the flow over compact bluff bodies. Computers & Fluids (2020), 198, 104400 – https://doi.org/10.1016/j.compfluid.2019.104400 – preprint
abstract
A shape optimization for tonal noise generated aerodynamically at low Mach number is performed for a cylinder with polygonal cross-section. Acoustic quantities are derived from a hybrid analytical formula, with aeroacoustic sources obtained from the incompressible solution of the direct Navier-Stokes equations in 2D at Re = 150; the solid domain is modelled by an Immersed Boundary Method. The optimization is done with the Particle Swarm Optimization (PSO) technique and performed in a cluster where each cost function evaluation is an independent flow simulation. The precision on the 4 main shape parameters is set to 0.001, consistently with the convergence criteria in time, grid and swarm. Optimal shapes for minimum drag and minimum acoustic power are relatively similar. A large range between the optimal shapes is obtained: factor 1.8 for drag and 20 dB for the acoustic power. The reduction of noise is associated with long and bluffer geometries, while the louder flows are associated with highly interacting shear layers obtained with back pointing triangles. The fluctuating lift is the major quantity to control noise at fixed length, while increasing the aspect ratio tends to reduce the noise for globally all geometries. An overall correlation between mean drag and fluctuating suction is also noticed.
2019
Pinto, W. J., Margnat, F. & Gervais, Y. Effect of cross-section on flow three-dimensionality for prismatic bodies and the associated noise emission 25th AIAA/CEAS Aeroacoustics Conference (Aeroacoustics 2019), 20-24 May 2019, Delft, The Netherlands, 2019 – https://doi.org/10.2514/6.2019-2531 – preprint
abstract
The influence of the geometry of prismatic bodies on their aerodynamic sound emission is studied. Curle’s formalism links the noise level to lift fluctuation and the spanwise organization of the flow coherent structures. Here, the acoustic field is measured using a set of microphones, and the flow spanwise coherence length is estimated from hot-wire velocity signals. An equivalent lift fluctuation is educed from the usual procedure of extrapolation of the sound level from 2D simulations to 3D long-span bodies. Circular and rectangular – of sectional aspect ratios (breadth-to-height ratio AR = b/d) of 1, 2 and 3 – cross sections are tested for a Reynolds number range from 6,000 to 27,000. The closure of the problem and the influence of the geometry are addressed by confronting the educed lift coefficients to literature values at Reynolds numbers in the same range, and in the 2D laminar regime. Microphone measurements show that all over the tested flow velocities, the acoustic efficiency of the cross sections respects the following hierarchy: circular < AR = 2 < square < AR = 3. The coherence length at the vortex shedding frequency is around 5d for the circular and square sections, while it is around 20d for the two rectangular sections. Educed lift fluctuation values are found to be lower than those previously reported, and of the same order than those obtained in the 2D laminar regime. The flow spanwise coherence length at the vortex shedding frequency is found to be inversely proportional to the breadth based, RMS, educed lift coefficient.
Pinto, W. J., Margnat, F. & Gervais, Y. Influence of cross-section on the aeolian tone: a numerical study in the laminar regime 25th AIAA/CEAS Aeroacoustics Conference (Aeroacoustics 2019), 20-24 May 2019, Delft, The Netherlands, 2019 – https://doi.org/10.2514/6.2019-2532 – preprint
abstract
Influence of the cross section on airframe noise is investigated numerically in 2D at low Reynolds number (Re = 150) by a direct Navier-Stokes (DNS) solver and an acoustic power formulation derived from an hybrid acoustic analogy. Flows around different canonical geometries (rectangle, triangle, ellipse and lozenge) and of different breadth-to-height ratio, called as aspect ratio (AR), are studied. The appearance of a wake organization with two rows of vortices is associated with a peak of RMS lift and frequency. Several relationships between flow integral, geometrical and topological quantities are proposed and discussed. The hormetic response of the RMS lift coefficient to the increase of AR is noticed for all geometries. The RMS lift coefficient, thus the noise, is found to correlate linearly to the vortex displacement normal to the flow and to the fourth power of the average drag.
Pinto, W. J. G. S. & Margnat, F. A shape optimization procedure for cylinders aeolian tone, Computers & Fluids (2019) – doi.org/10.1016/j.compfluid.2019.02.002
abstract
Shape optimisation for airframe noise is performed for a cylinder with polygonal crosssection. Acoustic quantities are derived from a hybrid approach, alimented by the incompressible solution of the direct Navier-Stokes equations in 2D; solid domain is defined by a Immersed Boundary Method. Optimisation is done with the Particle Swarm Optimisation (PSO) technique and performed in a cluster where each cost function evaluation is an independent flow simulation. The precision on the 4 main shape parameters is set to 0.001, consistently with the convergence criteria in time, grid and swarm. Optimal shapes for minimum drag and minimum acoustic power are relatively similar. A large range between the optimal shapes is obtained: factor 1.77 for drag and 20 dB for the acoustic power. The reduction of noise is associated with long and bluffer geometries, while the louder flows are associated with highly interacting shear layers obtained with back facing triangles. The fluctuating lift is the major quantity to control noise at fixed length, while the aspect ratio tends to reduce the noise for globally all geometries.
2018
Pinto, W. J. G. S. & Margnat, F. Shape optimisation for the noise induced by the flow over compact bluff bodies International Conference on Computational Fluid Dynamics (ICCFD10), 9-13 July 2018, Barcelona, Spain, 2018 – paper
abstract
Shape optimisation for airframe noise is performed for a cylinder with polygonal crosssection. Acoustic quantities are derived from a hybrid approach, alimented by the incompressible solution of the direct Navier-Stokes equations in 2D; solid domain is defined by a Immersed Boundary Method. Optimisation is done with the Particle Swarm Optimisation (PSO) technique and performed in a cluster where each cost function evaluation is an independent flow simulation. The precision on the 4 main shape parameters is set to 0.001, consistently with the convergence criteria in time, grid and swarm. Optimal shapes for minimum drag and minimum acoustic power are relatively similar. A large range between the optimal shapes is obtained: factor 1.77 for drag and 20 dB for the acoustic power. The reduction of noise is associated with long and bluffer geometries, while the louder flows are associated with highly interacting shear layers obtained with back facing triangles. The fluctuating lift is the major quantity to control noise at fixed length, while the aspect ratio tends to reduce the noise for globally all geometries.
Margnat, F. & Pinto, W. J. G. S. Contribution à l’Optimisation de Forme pour le Bruit d’Origine Aérodynamique 14ème Congrès Français d’Acoustique (CFA18), 23-27 April 2018, Le Havre, France, 2018
abstract - in French
Une procédure d’optimisation de forme est présentée. Elle est dédiée au bruit généré par les écoulements d’obstacle. La fonctionnelle coût est la puissance acoustique, qui est déduite directement des fluctuations d’effort aérodynamiques par une formule analytique en hypothèse de bruit tonal. Ces derniers sont estimés à partir de la résolution numérique directe de l’écoulement instationnaire en régime laminaire incompressible 2D sur un obstacle convexe symétrique sans incidence. La condition d’adhérence en paroi est assurée par une méthode de frontières immergées, qui permet d’utiliser le même maillage pour toutes les géométries. La forme de l’obstacle est définie à l’aide de 4 arcs de Bézier, dont les contraintes de continuité à l’ordre 2 conduisent à une paramétrisation à 4 degrés de liberté, pilotant le rapport d’aspect, la frontalité et les profilages amont et aval. L’optimisation est obtenue par un algorithme d’essaim particulaire (Particle Swarm Optimisation), qui offre un bon compromis entre le nombre d’évaluations de la fonctionnelle nécessaires pour trouver l’optimum et le risque d’optima locaux. Plusieurs optimisations sont présentées en augmentant progressivement la dimension de l’espace des paramètres afin d’affiner la perception de la surface de réponse et de la dynamique de l’essaim. On obtient notamment jusqu’à une différence d’efficacité en puissance d’au moins 15 dB entre les configurations extrêmes pour un rapport d’aspect fixé, et d’au moins 3.5 dB sous contrainte de surface ou de périmètre. Les premiers résultats montrent que les formes minimisant la fluctuation de portance sont relativement différentes de celles minimisant la trainée moyenne.
Margnat, F.; Cruz, R. V.; Pinto, W. J. G. S. & Lazure, H. Etude numérique d’un phénomène de sifflement de rétroviseur 14ème Congrès Français d’Acoustique (CFA18), 23-27 April 2018, Le Havre, France. 2018
abstract - in French
La problématique du sifflement généré par l’écoulement sur un rétroviseur automobile est abordée par une approche numérique. La modélisation consiste en la sélection d’un profil géométrique pour lequel la dynamique est majoritairement bidimensionnelle. Egalement, le choix d’un régime incompressible à bas nombre de Reynolds représente bien les aspects qualitatifs du phénomène tout en permettant de changer de configuration avec un temps de simulation non-prohibitif. Des tests réalisés pour des écoulements de cavité montrent que le mécanisme de sélection fréquentielle par verrouillage de phase est correctement décrit par la présente approche, sous hypothèse de nombre de Mach tendant vers zéro. Pour cette configuration modèle, l’obtention d’une rétroaction dans le cas du rétroviseur nécessite d’introduire une excitation des instabilités de couche cisaillée. Ceci est opéré par soufflage ponctuel au niveau de la paroi, dont la mise en œuvre est facilitée par l’utilisation d’une méthode de frontières immergées. Le sifflement apparait alors, à une fréquence d’environ 25 fois supérieure à celle de l’instabilité de sillage. Cette dernière apporte une modulation d’amplitude au sifflement, laquelle se traduit par l’apparition de pics secondaires autour de la fréquence de sifflement. L’analyse de l’écoulement sifflant et des conditions d’apparition du sifflement permettent de compléter la description des phénomènes observés lors des études expérimentales antérieures.
Pinto, W. J. G. S.; Margnat, F. & Gervais, Y. Poster session: Modélisation du Bruit d’origine Aérodynamique : Influence de la géométrie dans le bruit de forme Journées des Doctorants de l’Institut PPRIME (JDD2018), 22-23 Mars 2018, Poitiers, France, 2018
2015
Pinto, W. J. G. S.; Dourado, A.; Cavalini Jr., A. & Steffen Jr., W. Uncertainty Analysis in a Rotor System by Using the Fuzzy Logic Approach XXXVI Iberian Latin American Congress on Computational Methods in Engineering (CILAMCE 2015), 22-25 November 2015, Rio de Janeiro, Brazil, 2015
abstract
This paper is dedicated to the uncertainty analysis of the dynamic response of a simple supported rotor system composed by a horizontal flexible shaft, one rigid disc, and one additional ball bearing. A mathematical model of the rotor is derived from the Lagrange’s equation by using the Rayleigh-Ritz method. Some simplifications are adopted in order to obtain the responses of the rotor system considered. The inherent uncertainties of the bearing are modeled through the fuzzy logic technique. Triangular fuzzy numbers are used to represent the uncertain stiffness of the rotor’s bearing. A heuristic optimization algorithm is used in the α-level optimization to map the fuzzy inputs. A defuzzification process is carried out to estimate the stiffness of the bearing regarding the given uncertain interval. The analysis is confined to the frequency domain so that the envelopes of the rotor unbalance responses are evaluated.
RELEVANT UNDERGRADUATE PROJECTS
Preliminary design of a 7 passengers’ business turboprop plane
Responsible for CAD modeling and aerodynamic analysis. Multidisciplinary project accounting for aerodynamics, static and dynamic structural behavior, systems and flight control. Team of 12 students.
Conception of wind tunnel aeroelastic test section
Complete CAD conception of an aeroelastic test section with 2 degrees of freedom for flutter analysis of a NACA0012 profile.
Subsonic wind tunnel automation
Characterization of subsonic open wind tunnel. Conception in SolidWorks of an automate system with Pitot tube and aerodynamic balances.
Conceptual design of an agricultural aircraft
Modeling of an experimental agricultural aircraft in CATIA. Under a cooperation between UFU and Fábrica Brasileira de Aeronaves Ltda (FABE).
LANGUAGES
Portuguese: Native language
English: Proficient User – C1 (TOEIC : 970, march 2013)
French: Independent User – B2
German: Basic User – A2 (OnDaf : 52, january 2012)
COMPUTER & TECHNICAL SKILLS
Languages: Python, C/C++, Fortran, MATLAB, notions of R and VBA
CAD software: CATIA V5, SolidWorks, FreeCAD and DraftSight
Meshing: ICEM CFD, Gmsh, SnappyHexMesh (OpenFOAM) and notions of ANSA and HyperMesh
FEA: notions of ANSYS Mechanical, FEMAP and ABAQUS
CFD: OpenFOAM, ProLB, PowerFLOW suite, Star-CCM+, ANSYS Fluent and Palabos. Academic solvers: Incompact3D (PPRIME) and JAGUAR (CERFACS)
Miscellaneous: Microsoft Office (Word, Excel and Power Point) and LibreOffice; vector graphics editing (CorelDRAW and Inkscape) and Latex
