21549 research outputs found
Sort by
Moteur à Flux Axial à Aimants permanents et à Concentration de Flux : Réalisation, Mesure de Paramètres Statiques et Modèle Numérique 3D
Get PDFGreen energy or sources with zero-carbon emissions plays a leading role in the worldwide electrification. Electric machinery field is a subject of attention leading to the green technology. These machines contribute mainly in the context of industrial applications as electric or hybrid cars. Avoiding the issue and price of rare-earth magnets, potential researches contribute particularly to non-rare-earth machines. For that reasons, this thesis studies the ferrite motor through the ‘Spoke Type Axial Flux Permanent Magnet’ (STAFPM) topology. One from the main interesting properties of this motor is the capability of the no-load magnetic flux concentration in its airgap. The properties of this motor with ferromagnetic poles in the rotor are not so-well known. The present thesis focuses on studying its performances. Achieving this goal, consists firstly by doing a review on the analytical and numerical sizing approaches applied for radial and axial flux machines based on the magnetic field models and as well as, a review on the experimental test benchmark for salient-pole machines which provide us the identification and computation of the electromechanical parameters. Finally, a revision takes place on the 3D magnetic field modeling for sizing purposes. Thus the first section is partially devoted to review the application of the 3D finite difference method for axial machines which is a main objective in this thesis. Afterwards, using a 1D analytical model, a performance comparison takes place between a single stator-single rotor STAFPM motor and a reference Surface Mounted Axial Flux Permanent Magnet (SMAFPM) motor. This comparison is made on the electromagnetic torque and at unified parameters. Consequently, new STAFPM prototype is sized by a consideration of some magnetic constraints. This prototype takes place on a test bench. An original method of the experimental identification of the parameters of the lumped parameter electromechanical model is developed. This method is based on the static torque measurement as a function of the rotor position. A 3D numerical magnetic field model of the STAFPM motor is proposed. This tool studies the no-load magnetic field as well as the armature reaction fields. An original flux calculation method in the framework of the magnetic scalar potential formulation is developed. This method allows the quick calculation of the parameters of the lumped electromechanical model. The calculated parameters are compared to the experimentally identified parameters
Modelling challenges of stationary combustion in inert porous media
Get PDFThanks to strong heat recirculation, submerged combustion within porous media presents unique technological features such as broadened flammability limits and extended power range. The associated possibility to burn ultra-lean mixtures with minimal CO/NOx emissions makes porous media combustion a potential alternative in the industry, for instance in domestic heat generation or clean aviation where low pollutant emissions and robust operability are of paramount importance. However, even though this combustion mode has been studied for decades, there remains many open questions regarding the intertwined flame structure and the validity of associated low-order modelling. To date, volume-averaged models are mostly based upon ad hoc hypotheses and still present large discrepancies with experiments. Aiming to chal- lenge and strengthen these models, the present work presents analytical and numerical studies of the volume-averaged equations, followed by 3D direct pore-level simulations of methane-air and hydrogen-air combustion. Chapters 1 and 2 provide a critical review of concepts associated to flows and flames within porous media, with a focus on non-adiabatic combustion and macroscopic effective characteriza- tion. A classification of gaseous flames in terms of the thermal Péclet number is proposed, and the upscaling procedure on the pointwise equations is presented. Chapter 3 presents asymptotic results based on the volume-averaged equations, and the proposed theoretical framework un- veils the first fully-explicit formulae for flame speed in infinite and finite-length porous burners. Multi-layered burners are also considered theoretically for the first time, and the important con- cept of contact resistance between two stacked porous plates is underlined. Chapter 4 proposes a general classification of porous media combustion in three distinct regimes for increasing inter- phase heat transfer, only based on two reduced parameters, in order to reconcile the literature frameworks of local thermal equilibrium (LTE) and non-equilibrium (LTNE). Chapter 5, 6 and 7 present 3D pore-level direct numerical simulations of flames within porous media using complex kinetics, for various structural topologies and pore sizes. As a major technical hurdle encountered during the thesis, the meshing workflow from X-ray tomography to conformal computational mesh is given for practical use in the community. These DNS unveil the internal flame structure of methane-air and hydrogen-air flames within typical porous burners, and it is shown that when the pore size is larger than the flame thickness, sharp and locally- anchored flame fronts are observed. These local discontinuities related to the strongly non-linear reaction rates are shown to be in direct violation of the classical volume-averaged hypotheses. This demonstrates that new volume-averaged models are required, and accordingly a closure for reaction rates based upon phenomenology and observations in the 3D DNS is proposed. Eventually, the pore-level specificities of hydrogen combustion at pore scale are described
Study and development of an AI assistant for future Moon and Mars stations
Get PDFFollowing the Global Exploration Roadmap (GER) defined by the International Space Exploration Coordination Group
(ISECG), the Spaceship FR team from CNES, the French Space Agency, wishes to contribute to the development of
technologies extending human reach toward space, notably for the development of Moon and Mars bases. The
operation and sustainability of such structures in stressful isolation conditions constitute a high level technological and human challenge. To relieve the high mental load of the astronauts, the solution could be an artificial intelligence
assistant that would supervise the automation of the base as well as monitor and maintain the mental health of the crew
through a cognitive approach of the human-computer interaction. AI4U is the system at the crossroad between
computer sciences and human factors, aspiring to take on the task. Besides, its organisation and automation skills, this
artificial intelligence interacts with the astronauts with an intuitive and natural interface. It can support their work and leisure activities with an empathic approach, recognising their current mental state by using face and voice recognition. This paper will present the different steps of the development of AI4U, the challenges encountered and the next steps until its usage in a Moon outpost
Cohesive strength and fracture toughness of atmospheric ice
Get PDFThis paper aims at defining the key mechanical properties of atmospheric ice in order to improve the design of
mechanical de-icing systems. Based on ice fracture mechanisms, the parameters of interest are the cohesive
strength of the ice and its fracture toughness. A hybrid experimental/numerical vibrating method is used to
measure those critical values. Parameters such as temperature and precipitation rate influence the ice density
and ice samples are thus defined with respect to this parameter. First, the cohesive strength of ice is measured
over the entire range of ice density and a polynomial expression of the cohesive strength of ice is given as a
function of this density. Then the fracture toughness is measured for a smaller range of density and an average
critical value is given. Finally, the influence of the properties computed is discussed to assess the conditions of
atmospheric ice mechanical removal and the challenges for the design of mechanical ice protection systems. The
study tends to show that as its density decreases, ice is more difficult to remove mechanically
Classification of flight phases based on pilots’ visual scanning strategies
Get PDFEye movements analysis has great potential for understanding operator behaviour in many safety-critical domains, including aviation. In addition to traditional eye-tracking measures on pilots’ visual behavior, it seems promising to incorporate machine learning approaches to classify pilots’ visual scanning patterns. However, given the multitude of pattern measures, it is unclear which are better suited as predictors. In this study we analyzed the visual behaviour of eight pilots, flying different flight phases in a moving-base flight simulator. With this limited dataset we present a methodological approach to train linear Support Vector Machine models, using different combinations of the attention ratio and scanning pattern features. The results show that the overall accuracy to classify the pilots’ visual behaviour in different flight phases, improves from 51.6% up to 64.1% when combining the attention ratio and instrument scanning sequence in the classification model
Exergy analysis of unsteady flow around an adiabatic cylinder in vortex shedding condition
Get PDFThis paper shows the application of the aerodynamics
exergy analysis preliminary formulation developed for
unsteady flows for adiabatic and fixed bodies under low
subsonic flow conditions (no shock-waves). The
numerical implementation for URANS CFD cases of
the presented formulation is explained. The test case is a
circular cylinder in vortex shedding regime at Reynolds
number 140, at low subsonic flow, with no movement
with respect to the reference frame. The validation of
the developed formulation is done by comparing the
drag coefficient computed by the presented unsteady
exergy formulation versus the near field drag coefficient
taken as reference
Surface modification of 5083 Aluminum-Magnesium induced by marine microorganisms
Get PDFThe influence of microorganisms from a salt marsh in the surface modification of 5083 aluminium alloy (Al-Mg) in seawater was evaluated. An immersion test performed for 50 days in biotic and abiotic conditions, with electrochemical monitoring and surface/cross-section characterization by SEM/EDX and TEM after exposure, showed that microorganisms induced the formation of a homogenous layer on the Al-Mg surface. This layer, which proved to be composed of a double-structure: a dense, amorphous inner layer and a more porous outer layer, was demonstrated to influence the corrosion resistance of the Al-Mg alloy in seawater
Tim e-domain Wave Propagation in Rigid Porous Media using Equivalent Fluid Model with a Quadratic Nonlinearity
Get PDFThe acoustic properties of rigid porous media can be described by the equivalent fluid model (EFM) in the frequency
domain, involving complex-valued functions. These physical quantities can be irrational, which leads to fractional derivatives in the
time domain. Besides, this model is built with a constant flow resistivity, which is known to grow linearly with the flow velocity in the
Forchheimer regime. Hence, a correction on the EFM is made according to the Darcy-Forchheimer law, leading to a more general model with an additional nonlinear term. Here, an approach is presented to formulate the EFM equations with the Forchheimer’s correction in the time domain, where the fractional derivatives described by causal convolution are approximated by additional differential equations.
It results in a nonlinear system on which an energy-based analysis is performed to ensure its stability under suitable conditions
Efficient implementation of non-linear flow law using neural network into the Abaqus Explicit FEM code
Get PDFMachine learning techniques are increasingly used to predict material behavior in scientific applications and offer a significant advantage over conventional numerical methods. In this work, an Artificial Neural Network (ANN) model is used in a finite element formulation to define the flow law of a metallic material as a function of plastic strain , plastic strain rate \mdot{\varepsilon}^p and temperature . First, we present the general structure of the neural network, its operation and focus on the ability of the network to deduce, without prior learning, the derivatives of the flow law with respect to the model inputs. In order to validate the robustness and accuracy of the proposed model, we compare and analyze the performance of several network architectures with respect to the analytical formulation of a Johnson-Cook behavior law for a 42CrMo4 steel. In a second part, after having selected an Artificial Neural Network architecture with hidden layers, we present the implementation of this model in the Abaqus Explicit computational code in the form of a VUHARD subroutine. The predictive capability of the proposed model is then demonstrated during the numerical simulation of two test cases: the necking of a circular bar and a Taylor impact test. The results obtained show a very high capability of the ANN to replace the analytical formulation of a Johnson-Cook behavior law in a finite element code, while remaining competitive in terms of numerical simulation time compared to a classical approach
Upscaling optimal topology multimaterials structures using Deep Neural Networks
Get PDFThe problem of Topology Optimization aims to solve the question of the optimal material distribution subjected to known boundary and load conditions subject to a target volume fraction. In this study, we present a machine learning framework to tackle the problem of Multi Material Topology upscaling, i.e., the prediction of a higher resolution topology with just the low- resolution input. A Convolutional Deep Neural network was trained with a data set generated from an iterative code found in the existing literature. The network architecture implemented in this study is a modified version of SRGAN which has proven capabilities in upscaling complex real-world images. In this study, the perceptual loss function was used as the loss function in order to not penalize the network for its predictions that are off by a couple of pixels while simultaneously rewarding the network for its outputs that yield accurate compliance. The paper aims to present a novel approach to the problem of Topology Upscaling of 2x by mapping local features of the low-resolution topology to their higher resolution counterparts