41 |
Combined application of structural geology, the mechanics of discrete media and the analysis of in situ stresses and displacements for the modelling of mechanical behaviour of fractured rock masses / Application combinée de la géologie structurale, de la mécanique des milieux discrets et de l’analyse de contraintes et déplacements in situ à la modélisation du comportement mécanique de massifs rocheux fracturéTran, Thi Thu Hang 22 April 2013 (has links)
Pour étudier le comportement mécanique des massifs rocheux, en prenant en compte le réseau des discontinuités au sein de la roche intacte, cette recherche a pour objectif la représentation du massif par des modèles géométriques basés sur des relevés de terrain et l'analyse de ces modèles par l'utilisation d'outils informatiques adaptés pour les milieux granulaires. Le premier chapitre fait l'état de l'art des roches fracturées, des méthodes numériques de la mécanique des roches et des approches du calcul de structure d'un tunnel. Ces études conduisent à la proposition d'une méthodologie depuis les recherches in situ jusqu'à la modélisation et l'analyse mécanique, présentée dans le deuxième chapitre. Le massif rocheux est d'abord représenté géométriquement par la distribution de ses discontinuités, et l'utilisation du logiciel RESOBLOK basé sur la méthode du Réseau de Fractures Discrètes. Les modèles mécaniques de massifs rocheux sont ensuite présentés à partir des données sur les études de l'histoire du massif, et des mesures faites sur site et en laboratoire. Les modèles numériques en 3D sont analysés par l'utilisation du logiciel LMGC90 basé sur la méthode de la Dynamique des Contacts Non Réguliers. Les premières applications de la méthodologie sont exposées : la création d'une roche numérique pour simuler un essai de compression triaxiale, et la simulation d'une excavation multi phases d'un tunnel au rocher. La méthodologie proposée a été appliquée sur le marbre blanc de Saint Béat (Haute Garonne, France) et les résultats préliminaires sont donnés dans le chapitre trois. Les réponses mécaniques de la roche numérique sont analysées et son comportement est caractérisé. / Aimed at studying the mechanical behaviour of rock mass and considering the presence of the discontinuity network in the intact rock, this research concentrates on how the rock can be represented in suitable geometrical models, on the basis of site measurements, and then appropriately analysed using computer tools developed for the study of granular media. The first chapter deals with a bibliographical study on fractured rock and tunnel engineering. Different computational methods of rock mechanics are introduced. Simultaneously, three principal approaches for tunnel structural design are recalled. These studies lead to the proposition of a methodology from the in situ investigation to in-door modelling and mechanical analysis, presented in the second chapters. The rock mass is first geometrically represented through the distribution of discontinuities in the rock mass and the use of the RESOBLOK code based on the Discrete Fracture Network method. Mechanical models of rock mass are then presented from the data of historical studies on the rock mass and from laboratory and in situ measurements. The 3D computational models are analysed using the LMGC90 based on the Non Smooth Contact Dynamics method. The first two applications of the methodology are introduced: the generation of the numerical rock for the simulation of the triaxial compression test, and the simulation of multi-phase excavation of rock tunnel. The proposed methodology has been applied on the white marble of Saint Béat (Haute Garonne, France) and the initial results are given in the third chapter. The mechanical responses of the numerical rock mass are analysed and the bulk behaviour of the rock is evaluated.
|
42 |
Análise estrutural e de superfície de madeiras submetidas a tratamento térmico e degradadas pelo fungo Pycnoporus sanguineus / Structural analysis and surface of wood heat-treated and degraded by the fungus Pycnoporus sanguineusSantos, Sabrina Nicoleti Carvalho dos 17 July 2015 (has links)
Neste trabalho, analisou-se como o tratamento térmico e a degradação pelo fungo de podridão branca influenciaram as propriedades dos componentes das amostras de madeira das espécies Araucaria sp., Pinus sp., Erisma sp. e Hymenaea sp.. Essas amostras foram cortadas em diferentes planos de corte e submetidas inicialmente a um teor de umidade de equilíbrio de 12%. As caracterizações das amostras foram realizadas empregando a técnica de ângulo de contato, medidas de perda de massa (somente para amostras degradadas pelo fungo), a análise dinâmico-mecânica (DMA) e difração de raios X. Para as amostras degradadas pelo fungo, o valor de perda de massa foi maior para a espécie Pinus sp. e menor para Hymenaea sp.. Os difratogramas de raios X apontaram que o corte transversal apresentou um caráter mais amorfo em relação ao corte longitudinal. As medidas de ângulo de contato mostraram que com o aumento da temperatura de tratamento térmico a superfície se tornou mais hidrofóbica. O mesmo comportamento foi verificado para amostras submetidas à degradação pelo fungo. De forma que, tanto o tratamento térmico quanto a degradação pelo fungo, ocasionaram um aumento de grupos não-polares presentes na superfície. Já nas medidas de DMA foi observado que o módulo de armazenamento (E\') diminuiu com o aumento da temperatura. E o grau de mobilidade das cadeias poliméricas foi analisado por meio das temperaturas de transições vítreas das hemiceluloses e lignina (Tg), bem como, as relaxações secundárias através do fator de perda (tan δ). Esses resultados possibilitam uma maior elucidação sobre como o tratamento térmico e a degradação fúngica afetam diferentes tipos de amostras de várias espécies de madeira, essas informações são relevantes para a utilização e preservação dessas madeiras. / The present dissertation addresses the analysis of the influence of heat treatment and degradation by white rot fungus on the properties of the components of wood species Araucaria sp., Pinus sp., Erisma sp. and Hymenaea sp. Samples were cut at different cutting plans and initially subjected to a 12% equilibrium moisture content. They were then characterized by a contact angle technique, weight loss measurements (for samples degraded by the fungus), dynamic mechanical analysis (DMA) and X-ray diffraction. The mass loss values of samples degraded by the fungus were higher for Pinus sp. and lower for Hymenaea sp. The X-ray diffraction showed the transversal cut had an amorphous character in relation to the longitudinal cut. According to the contact angle measurements, showed that with increasing heat treatment temperature over the surface became hydrophobic. The same behavior was observed for samples submitted to degradation by fungus. In order, that both heat treatment and degradation by the fungus, caused an increase of non-polar groups on the surface. The DMA measurements showed the storage modulus (E\') decreases with increasing temperature. The degree of mobility of the polymer chains was analyzed through glass transition temperatures of hemicelluloses and lignin (Tg) and the secondary relaxations were assessed through the loss factor (tan δ). The results enabled greater elucidation on how the heat treatment and fungal degradation affect different types of samples of various wood species. Such information is relevant for the use and preservation of woods.
|
43 |
Thermo-mechanical Finite Element Analysis And Design Of Tail Section For A Ballistic MissleGuler, Togan Kemal 01 October 2012 (has links) (PDF)
During the flight of missiles, depending on the flight conditions, rotation of missiles around its centerline can cause instabilities. To override this issue, missile generally is designed in 2 sections. In the missile, the rear tail section and the front section are to rotate freely by means of bearings. Tail section on which bearings are mounted is designed according to thermal loads due to flow of hot gasses through the nozzle and mechanical loads due to inertial load, interference fit and thread preload which appear during flight of missile.
The purpose of this thesis is to determine the most suitable structural parameters according to the flight conditions of missile.
The geometrical and load parameters which have effect on the results were determined. Finite element model is formed by using FEA software. After that, transient nonlinear thermo-mechanical analyses are performed and the most effective parameter on VM (Von-Mises) stress and force is determined.
DOE (Design of Experiments) method was used to determine the most suitable values for the structural parameters. Totally 27 different configurations are studied to achieve to the most suitable values for variable set.
It is observed that VM stress and force results for all configurations are within the ± / %5 ranges. So this means parameters don&rsquo / t affect the systems response very much. By taking manufacturing processes into consideration, configuration with the highest bearing inner/outer ring interference is taken. From the comparison of the results, the most suitable configuration is obtained after checking forces and VM stress on the bearings.
|
44 |
Numerical Modeling of Hydraulic Fracture Propagation Using Thermo-hydro-mechanical Analysis with Brittle Damage Model by Finite Element MethodMin, Kyoung 16 December 2013 (has links)
Better understanding and control of crack growth direction during hydraulic fracturing are essential for enhancing productivity of geothermal and petroleum reservoirs. Structural analysis of fracture propagation and impact on fluid flow is a challenging issue because of the complexity of rock properties and physical aspects of rock failure and fracture growth. Realistic interpretation of the complex interactions between rock deformation, fluid flow, heat transfer, and fracture propagation induced by fluid injection is important for fracture network design. In this work, numerical models are developed to simulate rock failure and hydraulic fracture propagation. The influences of rock deformation, fluid flow, and heat transfer on fracturing processes are studied using a coupled thermo-hydro-mechanical (THM) analysis.
The models are used to simulate microscopic and macroscopic fracture behaviors of laboratory-scale uniaxial and triaxial experiments on rock using an elastic/brittle damage model considering a stochastic heterogeneity distribution. The constitutive modeling by the energy release rate-based damage evolution allows characterizing brittle rock failure and strength degradation. This approach is then used to simulate the sequential process of heterogeneous rock failures from the initiation of microcracks to the growth of macrocracks. The hydraulic fracturing path, especially for fractures emanating from inclined wellbores and closed natural fractures, often involves mixed mode fracture propagation. Especially, when the fracture is inclined in a 3D stress field, the propagation cannot be modeled using 2D fracture models. Hence, 2D/3D mixed-modes fracture growth from an initially embedded circular crack is studied using the damage mechanics approach implemented in a finite element method.
As a practical problem, hydraulic fracturing stimulation often involves fluid pressure change caused by injected fracturing fluid, fluid leakoff, and fracture propagation with brittle rock behavior and stress heterogeneities. In this dissertation, hydraulic fracture propagation is simulated using a coupled fluid flow/diffusion and rock deformation analysis. Later THM analysis is also carried out. The hydraulic forces in extended fractures are solved using a lubrication equation. Using a new moving-boundary element partition methodology (EPM), fracture propagation through heterogeneous media is predicted simply and efficiently. The method allows coupling fluid flow and rock deformation, and fracture propagation using the lubrication equation to solve for the fluid pressure through newly propagating crack paths.
Using the proposed model, the 2D/3D hydraulic fracturing simulations are performed to investigate the role of material and rock heterogeneity. Furthermore, in geothermal and petroleum reservoir design, engineers can take advantage of thermal fracturing that occurs when heat transfers between injected flow and the rock matrix to create reservoir permeability. These thermal stresses are calculated using coupled THM analysis and their influence on crack propagation during reservoir stimulation are investigated using damage mechanics and thermal loading algorithms for newly fractured surfaces.
|
45 |
Análise estrutural e de superfície de madeiras submetidas a tratamento térmico e degradadas pelo fungo Pycnoporus sanguineus / Structural analysis and surface of wood heat-treated and degraded by the fungus Pycnoporus sanguineusSabrina Nicoleti Carvalho dos Santos 17 July 2015 (has links)
Neste trabalho, analisou-se como o tratamento térmico e a degradação pelo fungo de podridão branca influenciaram as propriedades dos componentes das amostras de madeira das espécies Araucaria sp., Pinus sp., Erisma sp. e Hymenaea sp.. Essas amostras foram cortadas em diferentes planos de corte e submetidas inicialmente a um teor de umidade de equilíbrio de 12%. As caracterizações das amostras foram realizadas empregando a técnica de ângulo de contato, medidas de perda de massa (somente para amostras degradadas pelo fungo), a análise dinâmico-mecânica (DMA) e difração de raios X. Para as amostras degradadas pelo fungo, o valor de perda de massa foi maior para a espécie Pinus sp. e menor para Hymenaea sp.. Os difratogramas de raios X apontaram que o corte transversal apresentou um caráter mais amorfo em relação ao corte longitudinal. As medidas de ângulo de contato mostraram que com o aumento da temperatura de tratamento térmico a superfície se tornou mais hidrofóbica. O mesmo comportamento foi verificado para amostras submetidas à degradação pelo fungo. De forma que, tanto o tratamento térmico quanto a degradação pelo fungo, ocasionaram um aumento de grupos não-polares presentes na superfície. Já nas medidas de DMA foi observado que o módulo de armazenamento (E\') diminuiu com o aumento da temperatura. E o grau de mobilidade das cadeias poliméricas foi analisado por meio das temperaturas de transições vítreas das hemiceluloses e lignina (Tg), bem como, as relaxações secundárias através do fator de perda (tan δ). Esses resultados possibilitam uma maior elucidação sobre como o tratamento térmico e a degradação fúngica afetam diferentes tipos de amostras de várias espécies de madeira, essas informações são relevantes para a utilização e preservação dessas madeiras. / The present dissertation addresses the analysis of the influence of heat treatment and degradation by white rot fungus on the properties of the components of wood species Araucaria sp., Pinus sp., Erisma sp. and Hymenaea sp. Samples were cut at different cutting plans and initially subjected to a 12% equilibrium moisture content. They were then characterized by a contact angle technique, weight loss measurements (for samples degraded by the fungus), dynamic mechanical analysis (DMA) and X-ray diffraction. The mass loss values of samples degraded by the fungus were higher for Pinus sp. and lower for Hymenaea sp. The X-ray diffraction showed the transversal cut had an amorphous character in relation to the longitudinal cut. According to the contact angle measurements, showed that with increasing heat treatment temperature over the surface became hydrophobic. The same behavior was observed for samples submitted to degradation by fungus. In order, that both heat treatment and degradation by the fungus, caused an increase of non-polar groups on the surface. The DMA measurements showed the storage modulus (E\') decreases with increasing temperature. The degree of mobility of the polymer chains was analyzed through glass transition temperatures of hemicelluloses and lignin (Tg) and the secondary relaxations were assessed through the loss factor (tan δ). The results enabled greater elucidation on how the heat treatment and fungal degradation affect different types of samples of various wood species. Such information is relevant for the use and preservation of woods.
|
46 |
Modelling Concrete Behaviour At Early-Age : Multiscale Analysis And Simulation Of A Massive Disposal Structure / Modélisation du comportement du bêton au jeune âge : analyse multiéchelle et simulation d'une structure massive de stockageHonorio de Faria, Tulio 28 September 2015 (has links)
La prédiction précise du comportement à long et court terme des structures en béton dans le domaine nucléaire est essentielle pour assurer des performances optimales (intégrité, capacité de confinement) pendant leur durée de vie. Dans le cas particulier des structures massives en béton, la chaleur produite au jeune âge par les processus d'hydratation ne peut pas s’évacuer rapidement, si bien que des températures élevées peuvent être atteintes et les gradients de température qui en résultent peuvent conduire à la fissuration, en fonction des conditions aux limites et contraintes internes auxquelles ces structures sont soumises. Les objectifs de cette étude sont (1) d'effectuer des simulations numériques afin de décrire et prédire le comportement thermo-chimio-mécanique au jeune âge d'une structure massive en béton dédiée au stockage de déchets en surface, et (2) de développer et appliquer des outils de changement d'échelle pour estimer rigoureusement, à partir de la composition du matériau, les propriétés physiques du béton nécessaires à une analyse au jeune âge. Une étude chimio-thermique visant à déterminer l'influence de la convection, du rayonnement solaire, du re-rayonnement et de la chaleur d'hydratation sur la réponse thermique de la structure est tout d’abord menée. Des recommandations pratiques concernant les températures de bétonnage sont fournies afin de limiter la température maximale atteinte au sein de la structure. Ensuite, au moyen d'une analyse mécanique, des stratégies de modélisation simplifiées et plus complexes (prenant en compte l’endommagement couplé au fluage) sont mises en œuvre pour évaluer des scénarios intégrant différentes conditions aux limites issues de l'analyse chimio-thermique précédente. Dans un second temps, une étude prenant en compte le caractère multi-échelle du béton est réalisée. Un modèle simplifié de cinétique d'hydratation du ciment est proposé. Les évolutions des fractions volumiques des différentes phases au niveau de la pâte de ciment peuvent être alors estimées. Par la suite des outils d’homogénéisation analytiques et numériques développés dans un cadre vieillissant sont présentés et appliqués pour estimer les propriétés mécaniques et thermiques des matériaux cimentaires. Les données d’entrée utilisées dans l'analyse structurelle sont finalement comparées avec les estimations obtenues dans l'analyse multiéchelle. Pour conclure, la stratégie proposée dans cette thèse vise à prédire le comportement des structures massives en béton à partir de la composition du béton au moyen d'une approche séquentielle: le comportement du béton est estimé via les outils de changement d’échelle, fournissant ainsi les données d'entrée pour l'analyse phénoménologique à l’échelle de la structure. / The accurate prediction of the long and short-term behaviour of concrete structures in the nuclear domain is essential to ensure optimal performances (integrity, containment roperties) during their service life. In the particular case of massive concrete structures, at early age the heat produced by hydration reactions cannot be evacuated fast enough so that high temperatures may be reached and the resulting gradients of temperature might lead to cracking according to the external and internal restraints to which the structures are subjected. The goals of this study are (1) to perform numerical simulations in order to describe and predict the thermo-chemo-mechanical behaviour at early-age of a massive concrete structure devoted to nuclear waste disposal on surface, and (2) to develop and apply upscaling tools to estimate rigorously the key properties of concrete needed in an early-age analysis from the composition of the material. Firstly, a chemo-thermal analysis aims at determining the influence of convection, solar radiation, reradiation and hydration heat on the thermal response of the structure. Practical recommendations regarding concreting temperatures are provided in order to limit the maximum temperature reached within the structure. Then, by means of a mechanical analysis, simplified and more complex (i.e. accounting for coupled creep and damage) modelling strategies are used to assess scenarios involving different boundary conditions defined from the previous chemo-thermal analysis. Secondly, a study accounting for the multiscale character of concrete is performed. A simplified model of cement hydration kinetics is proposed. The evolution of the different phases at the cement paste level can be estimated. Then, analytical and numerical tools to upscale the ageing properties are presented and applied to estimate the mechanical and thermal properties of cementbased materials. Finally, the input data used in the structural analysis are compared with the estimations obtained in the multiscale analysis. To conclude, the entire strategy proposed in this thesis aims at predicting the behaviour of massive concrete structures from the composition of the concrete by means of a sequenced approach: concrete behaviour is estimated using the upscaling tools, providing then the input data to the phenomenological analysis at the structure level.
|
47 |
DIRECT TESTING OF TIRE TREAD COMPOUNDS AT HIGH FREQUENCIES USING A NEWLY DEVELOPED DYNAMIC MECHANICAL ANALYSIS (DMA) SYSTEMEsmaeeli, Roja 25 August 2020 (has links)
No description available.
|
48 |
Wood/Polymeric Isocyanate Resin Interactions: Species dependenceDas, Sudipto 28 September 2005 (has links)
The performance of polymeric diphenylmethane diisocyanate (PMDI) resin is known to be highly dependent on the wood species. This species dependence may be due to differences in: cure chemistry, interphase morphology, or both of these factors. This study addresses aspects of the cure chemistry and interphase morphology of wood/PMDI bondlines; specifically these effects are compared using two woods: yellow-poplar and southern pine.
In this study, the cure chemistry of wood-PMDI system was analyzed with solid state NMR (SSNMR) using wood samples cured with doubly labeled (15N,13C) PMDI resin. The kinetics of PMDI cure in the presence of wood was analyzed with differential scanning calorimetry. Thermogravimetric analysis was used to analyze the effect of resin impregnation on the degradation patterns of wood. The wood-PMDI bond morphology was probed with dynamic and static (creep) mechanical analyses in both dry and plasticized conditions. The effect of resin on wood polymer relaxations was quantitatively analyzed by both the time-temperature superposition principle and the Kohlrausch-Williams-Watts equation.
The presence of a small but statistically significant species effect was observed on both the cure chemistry and bond morphology of wood-PMDI system at low cure temperatures. The cure of PMDI resin was found to be significantly faster in pine relative to corresponding poplar samples. Resin impregnation showed a significant species dependent effect on the wood mechanical properties; the resinated pine samples showed increase in compliance while the corresponding poplar samples became stiffer.
The in situ lignin relaxation was studied with both dynamic and static modes, using plasticized wood samples. Results showed that the lignin relaxation was slightly affected by resin impregnation in both woods, but the effect was relatively larger in pine. Static experiments of dry wood samples showed a significant reduction in the interchain interactions of wood polymers in pine samples, exclusively. Investigation of plasticized pine samples, which focuses on the in situ lignin relaxations, showed only minor changes with resin impregnation. This led us to hypothesize that the large changes observed in dry samples, were due to the in situ amorphous polysaccharides. The wood-PMDI interactions were significantly reduced upon acetylation of wood. This study also discusses three new and highly sensitive methods for the analysis of wood-resin interactions. / Ph. D.
|
49 |
High Temperature SiC Embedded Chip Module (ECM) with Double-sided Metallization Structureyin, jian 03 January 2006 (has links)
The work reported in this dissertation is intended to propose, analyze and demonstrate a technology for a high temperature integrated power electronics module, for high temperature (e.g those over 200oC) applications involving high density and low stress.
To achieve this goal, this study has examined some existing packaging approaches, such as wire-bond interconnects and solder die-attach, flip-chip and pressure contacts. Based on the survey, a high temperature, multilayer 3-D packaging technology in the form of an Embedded Chip Module (ECM) is proposed to realize a lower stress distribution in a mechanically balanced structure with double-sided metallization layers and material CTE match in the structure.
Thermal and thermo-mechanical analysis on an ECM is then used to demonstrate the benefits on the cooling system, and to study the material and structure for reducing the thermally induced mechanical stress. In the thermal analysis, the high temperature ECM shows the ability to handle a power density up to 284 W/in3 with a heat spreader only 2.1x2.1x0.2cm under forced convection. The study proves that the cooling system can be reduced by 76% by using a high temperature module in a room temperature environment.
Furthermore, six proposed structures are compared using thermo-mechanical analysis, in order to obtain an optimal structure with a uniform low stress distribution. Since pure Mo cannot be electroplated, the low CTE metal Cr is proposed as the stress buffering material to be used in the flat metallization layers for a fully symmetrical ECM structure. Therefore, a chip area stress as low as 126MPa is attained.
In the fabrication process, the high temperature material glass and a ceramic adhesive are applied as the insulating and sealing layers. Particularly, the Cr stress buffering layer is successfully electroplated in the high temperature ECM by means of the hard chrome plating process. The flat metallization layer is accomplished by using a combined structure with Cr and Cu metallization layers.
The experimental evaluations, including the electrical and thermal characteristics of the ECM, have been part of in the study. The forward and reverse characteristics of the ECM are presented up to 250oC, indicating proper device functionality. The study on the reverse characteristics of the ECM indicates that the large leakage current at high temperature is not due to the package surrounding the chip, but chiefly caused by the Schottky junction and the chip passivation layer. Finally, steady-state and transient measurements are conducted in terms of the thermal measurements. The steady-state thermal measurement is used to demonstrate the cooling system reduction. To obtain the thermal parameters of the different layers in the high temperature ECM, the transient thermal measurement is applied to a single chip ECM based on the temperature cooling-down curve measurement. / Ph. D.
|
50 |
Identification expérimentale de sources vibratoires par résolution du problème inverse modélisé par un opérateur éléments finis local / Experimental identification of vibration sources by solving the inverse problem modeled by local finite element operatorRenzi, Cédric 16 December 2011 (has links)
L'objet de cette thèse est l'extension aux structures complexes de la méthode de Résolution Inverse Fenêtrée Filtrée (RIFF). L'idée principale se base sur le modèle Eléments Finis local et libre d'une partie de la structure étudiée. Tout d'abord, la méthode a été développée dans le cas des poutres. Les mesures de vibrations sont alors injectées dans le modèle Eléments Finis de la partie de poutre analysée. Les rotations sont estimées à l'aide de mesures de déplacements supplémentaires et des fonctions de forme sur le support élémentaire. La méthode étant sensible vis-à-vis des incertitudes de mesures, une régularisation a dû être développée. Celle-ci repose sur une double inversion de l'opérateur où une régularisation de type Tikhonov est appliquée dans la seconde inversion. L'optimisation de cette régularisation est réalisée par le principe de la courbe en L. A cause des effets de lissage induits par la régularisation, les moments ne peuvent être reconstruits mais ils apparaissent comme des ''doublets'' de forces. Ceci nous a conduit à résoudre le problème en supposant que seules des forces agissent sur la poutre. Enfin, une étude des effets de la troncature du domaine a été menée dans le but de s'affranchir des efforts de couplage apparaissant aux limites de la zone étudiée. Le cas des plaques a été considéré ensuite afin d'augmenter progressivement la complexité des modèles utilisés. L'approche Eléments Finis a permis d'intégrer à la méthode des techniques de condensation dynamique et de réduction par la méthode de Craig-Bampton. Le nombre de degrés de liberté est trop élevé pour permettre une estimation des rotations par mesures de déplacements supplémentaires, la condensation dynamique est employée afin de les supprimer dans le modèle théorique. Par ailleurs, la régularisation induisant une perte de résolution spatiale à cause de son effet de lissage, une procédure de déconvolution spatiale basée sur l'algorithme de Richardson-Lucy a été ajoutée en post traitement. Enfin, une application de la méthode à la détection de défauts a été envisagée de même que l'application de la méthode à l'identification des efforts appliqués par une pompe à huile sur un banc d'essais industriel. Le travail s'est donc appuyé sur des développements numériques et la méthode a été validée expérimentalement en laboratoire et en contexte industriel. Les résultats de la thèse fournissent un outil prédictif des efforts injectés par des sources de vibrations raccordées à une structure en s'appuyant sur un modèle Eléments Finis local et des mesures de déplacements vibratoires, le tout en régime harmonique. / The object of this thesis is the extension to complex structures of the RIFF method (Résolution Inverse Fenêtrée Filtrée). Considering a subpart of a structure, the main idea is to build a local Finite Element model using free boundary conditions. First, the general method was developed on beams. Vibration measurements are injected in the Finite Element model of the analysed part of the beam. Rotations are estimated using extra-displacement measurements and elementary shape functions. The method is highly sensitive towards errors present in measurements, so a regularisation had to be used. This one consists in a double inversion of the operator where a Tikhonov regularisation is applied when performing the second inversion. The regularisation parameter is tuned by the L-curve principle. Because of the smoothing effect of the Tikhonov procedure, moments cannot be reconstructed anymore at this stage, but they do still appear as sets of opposite forces. This setback led us to solve the problem by restricting it to forces only equations. At last, the study of the truncature of the domain was conducted in the aim to suppress coupling forces appearing at the limits of the studied area. Then, the case of plates was considered in order to increase progressively the models’ complexities. The Finite Element approach permitted us to implement dynamical condensation as well as Craig-Bampton reduction techniques. This allowed us to reduce the total number of degrees of freedom to be taken into account both from a numerical and an experimental standpoint. For example, dynamical condensation allows to eliminate rotations in the model. Besides, regularisation induces a lack of spatial resolution because of its smoothing effect. A spatial deconvolution technique was therefore developed; it is based on the Richardson-Lucy algorithm which is applied at a post-processing stage. At last, it was successfully proposed to extend the method to the application of detecting defaults present in the structure. The method was also validated on an industrial test bench in order to identify the forces applied by an oil pump taken from a truck’s engine. This phD thesis relied on numerical developments and the method was validated experimentally both in laboratory and industrial context. Main results provide a predictive tool to evaluate injected forces by vibration sources linked to a structure. It necessitates to inject vibratory displacements measurements into a Finite Element model.
|
Page generated in 0.0392 seconds