Search

The authors are grateful to Bertrand David (Telecom-ParisTech) for computing the code allowing the STFT filtering procedure used in Section 5.1. The filter has been designed in the framework of the PAFI project (Plateforme d’Aide la facture Instrumentale, www.pafi.fr) which is also thanked.; The vibrations and sounds produced by two notes of a double second steelpan are investigated, the main objective being to quantify the nonlinear energy exchanges occurring between vibration modes that are responsible of the peculiar sound of the instrument. A modal analysis first reveals the particular tuning of the modes and the systematic occurence of degenerate modes, from the second one, this feature being a consequence of the tuning and the mode localization. Forced vibrations experiments are then performed to follow precisely the energy exchange between harmonics of the vibration and thus quantify properly the mode couplings. In particular, it is found that energy exchanges are numerous, resulting in complicated frequency response curves even for very small levels of vibration amplitude. Simple models displaying1:2:2 and 1:2:4 internal resonance are then fitted to the measurements, allowing to identify the values of the nonlinear quadratic coupling coefficients resulting from the geometric nonlinearity. The identified 1:2:4 model is finally used to recover the time domain variations of an impacted note in normal playing condition, resulting in an excellent agreement for the temporal behaviour of the first four harmonics.; International audience; The vibrations and sounds produced by two notes of a double second steelpan are investigated, the main objective being to quantify the nonlinear energy exchanges occurring between vibration modes that are responsible of the peculiar sound of the instrument. A modal analysis first reveals the particular tuning of the modes and the systematic occurence of degenerate modes, from the second one, this feature being a consequence of the tuning and the mode localization. Forced vibrations experiments are then performed to follow precisely the energy exchange between harmonics of the vibration and thus quantify properly the mode couplings. In particular, it is found that energy exchanges are numerous, resulting in complicated frequency response curves even for very small levels of vibration amplitude. Simple models displaying1:2:2 and 1:2:4 internal resonance are then fitted to the measurements, allowing to identify the values of the nonlinear quadratic coupling coefficients resulting from the geometric nonlinearity. The identified 1:2:4 model is finally used to recover the time domain variations of an impacted note in normal playing condition, resulting in an excellent agreement for the temporal behaviour of the first four harmonics.

International audience; This paper presents a modal, time-domain scheme for the nonlinear vibrations of perfect and imperfect plates. The scheme can take into account a large number of degrees-of-freedom and is energy-conserving. The targeted application is the sound synthesis of cymbals and gong-like musical instruments, which are known for displaying a strongly nonlinear vibrating behaviour. This behaviour is typical of a wave turbulence regime, in which the wide-band spectrum of excited modes is observable in the form of an energy cascade. The modal method is selected for its versatility in handling complex damping laws that can be implemented easily by selecting appropriate damping values in each one of the modal equations. In the first part of the paper, the modal method is explained in its generality, and it will be seen that the method is valid for plates with arbitrary geometry and boundary conditions as long as the eigenmodes are known. Secondly, a time-integration, energy-conserving scheme for perfect and imperfect plates is presented, and implementation comments are given in order to treat efficiently the high-dimensionality of the resulting dynamical system. The scheme is run with appropriate parameters in order to produce sound samples. A simple impact law is considered for the exci-tation, whereas the flexibility of the method is highlighted by showing simulations for free-edge circular plates and simply-supported rectangular plates, together with various damping laws.

International audience; A passive control of aeroelastic instabilities on a two-degrees-of-freedom (dofs) system is considered here using shape memory alloys (SMA) springs in their pseudo-elastic regime. SMA present a solid-solid phase change that allow them to face strong deformations (∼10%) ; in the pseudo-elastic regime, an hysteresis loop appears in the stress-strain relationship which in turn gives rise to an important amount of dissipated energy. This property makes the SMA a natural candidate for mitigating undesired vibrations in a passive manner. A 2-dofs system is here used to model the classical flutter instability of a wing section in a uniform flow. The SMA spring is selected to act on the pitch in order to dissipate energy of the predominant motion. A simple phenomenological model for the SMA hysteresis loop is introduced, allowing for a quantitative study of the important parameters to optimize in view of an experimental design. Thanks to a simple phenomenological model for the SMA hysteresis loop, a quantitative numerical study is performed in order to exhibit the best tuning of the material parameters for controlling the flutter instability.

International audience; Non-linear vibrations of free-edge shallow spherical shells are investigated, in order to predict the trend of non-linearity (hardening/softening behaviour) for each mode of the shell, as a function of its geometry. The analog for thin shallow shells of von Kármán's theory for large deflection of plates is used. The main difficulty in predicting the trend of non-linearity relies in the truncation used for the analysis of the partial differential equations (PDEs) of motion. Here, non-linear normal modes through real normal form theory are used. This formalism allows deriving the analytical expression of the coefficient governing the trend of non-linearity. The variation of this coefficient with respect to the geometry of the shell (radius of curvature R, thickness h and outer diameter 2 a) is then numerically computed, for axisymmetric as well as asymmetric modes. Plates (obtained as R → ∞) are known to display a hardening behaviour, whereas shells generally behave in a softening way. The transition between these two types of non-linearity is clearly studied, and the specific role of 2:1 internal resonances in this process is clarified. © 2006 Elsevier Ltd. All rights reserved.

International audience; This paper is devoted to the derivation and the analysis of vibrations of shallow spherical shell subjected to large amplitude transverse displacement. The analog for thin shallow shells of von Kármán's theory for large deflection of plates is used. The validity range of the approximations is assessed by comparing the analytical modal analysis with a numerical solution. The specific case of a free edge is considered. The governing partial differential equations are expanded onto the natural modes of vibration of the shell. The problem is replaced by an infinite set of coupled second-order differential equations with quadratic and cubic non-linear terms. Analytical expressions of the non-linear coefficients are derived and a number of them are found to vanish, as a consequence of the symmetry of revolution of the structure. Then, for all the possible internal resonances, a number of rules are deduced, thus predicting the activation of the energy exchanges between the involved modes. Finally, a specific mode coupling due to a 1:1:2 internal resonance between two companion modes and an axisymmetric mode is studied. © 2004 Elsevier Ltd. All rights reserved.

International audience; This paper summarizes some results obtained in the last few years for the modeling of nonlinear vibrating instruments such as gongs and cymbals. Linear, weakly nonlinear and chaotic regimes are successively examined. A theoretical mechanical model is presented, based on the nonlinear von Kármán equations for thin shallow spherical shells. Modal projection and Nonlinear Normal Mode (NNM) formulation leads to a subset of coupled nonlinear oscillators. Current developments are aimed at using this subset for sound synthesis purpose.

International audience; Non-linear normal modes (NNMs) are used in order to derive accurate reduced-order models for large amplitude vibrations of structural systems displaying geometrical non-linearities. This is achieved through real normal form theory, recovering the definition of a NNM as an invariant manifold in phase space, and allowing definition of new co-ordinates non-linearly related to the initial, modal ones. Two examples are studied: a linear beam resting on a non-linear elastic foundation, and a non-linear clamped-clamped beam. Throughout these examples, the main features of the NNM formulation will be illustrated: prediction of the correct trend of non-linearity for the amplitude-frequency relationship, as well as amplitude-dependent mode shapes. Comparisons between different models - using linear and non-linear modes, different number of degrees of freedom, increasing accuracy in the asymptotic developments - are also provided, in order to quantify the gain in using NNMs instead of linear modes.

International audience; This article is devoted to an experimental validation of a theoretical model presented in an earlier contribution by the same authors. The non-linear forced vibrations of circular plates, with the excitation frequency close to the natural frequency of an asymmetric mode, are investigated. The experimental set-up, which allows one to perform precise measurements of the vibration amplitudes of the two preferential configurations, is presented. Experimental resonance curves showing the amplitude and the phase of each configuration as functions of the driving frequency are compared to the theoretical ones, leading to a quantitative validation of the predictions given by the model. Finally, all the approximations used are systematically discussed, in order to show the scope and relevance of the approach. © 2003 Elsevier Science Ltd. All rights reserved.

International audience; A passive control of aeroelastic instabilities on a two-degrees-of-freedom (dofs) system is considered here using shape memory alloys (SMA) springs in their pseudo-elastic regime. SMA present a solid-solid phase change that allow them to face strong deformations (∼ 10%); in the pseudo-elastic regime, an hysteresis loop appears in the stress-strain relationship which in turn gives rise to an important amount of dissipated energy. This property makes the SMA a natural candidate for damping undesired vibrations in a passive manner. A 2-dofs system is here used to model the classical flutter instability of a wing section in an uniform flow. The SMA spring is selected on the pitch mode in order to dissipate energy of the predominant motion. A simple model for the SMA hysteresis loop is introduced, allowing for a quantitative study of the important parameters to optimize in view of an experimental design.

International audience; We propose a direct method for computing modal coupling coefficients - due to geometrically nonlinear effects - for thin shells vibrating at large amplitude and discretized by a finite element (FE) procedure. These coupling coefficients arise when considering a discrete expansion of the unknown displacement onto the eigenmodes of the linear operator. The evolution problem is thus projected onto the eigenmodes basis and expressed as an assembly of oscillators with quadratic and cubic nonlinearities. The nonlinear coupling coefficients are directly derived from the finite element formulation, with specificities pertaining to the shell elements considered, namely, here elements of the ''Mixed Interpolation of Tensorial Components'' family (MITC). Therefore, the computation of coupling coefficients, combined with an adequate selection of the significant eigenmodes, allows the derivation of effective reduced-order models for computing - with a continuation procedure - the stable and unstable vibratory states of any vibrating shell, up to large amplitudes. The procedure is illustrated on a hyperbolic paraboloid panel. Bifurcation diagrams in free and forced vibrations are obtained. Comparisons with direct time simulations of the full FE model are given. Finally, the computed coefficients are used for a maximal reduction based on asymptotic nonlinear normal modes (NNMs), and we find that the most important part of the dynamics can be predicted with a single oscillator equation.

These lecture notes are related to the CISM course on ”Modal Analysis of nonlinear Mechanical systems”, held at Udine, Italy, from June 25 to 29, 2012. The key concept at the core of all the lessons given during this week is the notion of Nonlinear Normal Mode (NNM), a theoretical tool allowing one to extend, through some well-chosen assumptions and limitations, the linear modes of vibratory systems, to nonlinear regimes. More precisely concerning these notes, they are intended to show the explicit link between Normal Form theory and NNMs, for the specific case of vibratory systems displaying polynomial type nonlinearities. After a brief introduction reviewing the main concepts for deriving the normal form for a given dynamical system, the relationship between normal form theory and nonlinear normal modes (NNMs) will be the core of the developments. Once the main results presented, application of NNMs to vibration problem where geometric nonlinearity is present, will be highlighted. In particular, the developments of reduced-order models based on NNMs expressed asymptotically with the formalism of real normal form, will be deeply presented.

Nonlinear vibrations of thin rectangular plates are considered, using the von kármán equations in order to take into account the effect of geometric nonlinearities. Solutions are derived through an expansion over the linear eigenmodes of the system for both the transverse displacement and the Airy stress function, resulting in a series of coupled oscillators with cubic nonlinearities, where the coupling coefficients are functions of the linear eigenmodes. A general strategy for the calculation of these coefficients is outlined, and the particular case of a simply supported plate with movable edges is thoroughly investigated. To this extent, a numerical method based on a new series expansion is derived to compute the Airy stress function modes, for which an analytical solution is not available. It is shown that this strategy allows the calculation of the nonlinear coupling coefficients with arbitrary precision, and several numerical examples are provided. Symmetry properties are derived to speed up the calculation process and to allow a substantial reduction in memory requirements. Finally, analysis by continuation allows an investigation of the nonlinear dynamics of the first two modes both in the free and forced regimes. Modal interactions through internal resonances are highlighted, and their activation in the forced case is discussed, allowing to compare the nonlinear normal modes (NNMs) of the undamped system with the observable periodic orbits of the forced and damped structure. © 2013 Springer-Verlag Wien.

The effect of damping in the wave turbulence regime for thin vibrating plates is studied. An experimental method, allowing measurements of dissipation in the system at all scales, is first introduced. Practical experimental devices for increasing the dissipation are used. The main observable consequence of increasing the damping is a significant modification in the slope of the power spectral density, so that the observed power laws are not in a pure inertial regime. However, the system still displays a turbulent behavior with a cut-off frequency that is determined by the injected power which does not depend on damping. By using the measured damping power-law in numerical simulations, similar conclusions are drawn out., Comment: 6 pages, 6 figures

International audience; Numerical computation of Nonlinear Normal Modes (NNMs) for conservative vibratory systems is addressed, with the aim of deriving accurate reduced-order models up to large amplitudes. A numerical method is developed, based on the center manifold approach for NNMs, which uses an interpretation of the equations as a transport problem, coupled to a periodicity condition for ensuring manifold's continuity. Systematic comparisons are drawn with other numerical methods, and especially with continuation of periodic orbits, taken as reference solutions. Three di erent mechanical systems, displaying peculiar characteristics allowing for a general view of the performance of the methods for vibratory systems, are selected. Numerical results show that invariant manifolds encounter folding points at large amplitude, generically (but not only) due to internal resonances. These folding points involve an intrinsic limitation to reduced-order models based on the center manifold and on the idea of a functional relationship between slave and master coordinates. Below that amplitude limit, numerical methods are able to produce reduced-order models allowing for a precise prediction of the backbone curve.

International audience; The non-linear dynamic thermomechanical behaviour of superelastic shape memory alloys is investigated. To this end, the Zaki-Moumni model, initially developed for quasi-static loading cases, is extended to simulate the uniaxial forced oscillations of a shape memory alloy device. First, the influence of loading rate is accounted for by considering the thermomechanical coupling in the behaviour of NiTi shape memory alloy. Comparisons between simulations and experimental results show good agreement. Then, the forced response of a shape memory alloy device is investigated at resonance. Both isothermal and non-isothermal conditions are studied, as well as non-symmetric tensile-compressive restoring force. In the case of large values of forcing amplitudes, simulation results show that the dynamic response is prone to jumps, bifurcations and chaotic solutions.

International audience; A thin plate, excited by a harmonic external forcing of increasing amplitude, shows transitions from a periodic response to a chaotic state of wave turbulence. By analogy with the transition to turbulence observed in fluid mechanics as the Reynolds number is increased, a generic transition scenario for thin vibrating plates, first experimentally observed, is here numerically studied. The von Kármán equations for thin plates, which include geometric non-linear effects, are used to model large amplitude vibrations, and an energy-conserving finite difference scheme is employed for discretisation. The transition scenario involves two bifurcations separating three distinct regimes. The first regime is the periodic, weakly non-linear response. The second is a quasiperiodic state where energy is exchanged between internally resonant modes. It is observed only when specific internal resonance relationships are fulfilled between the eigenfrequencies of the structure and the forcing frequency; otherwise a direct transition to the last turbulent state is observed. This third, or turbulent, regime is characterized by a broadband Fourier spectrum and a cascade of energy from large to small wavelengths. For perfect plates including cubic non-linearity, only third-order internal resonances are likely to exist. For imperfect plates displaying quadratic nonlinearity, the energy exchanges and the quasiperiodic states are favored and thus are more easily obtained. Finally, the turbulent regime is characterized in the light of available theoretical results from wave turbulence theory.

International audience; Perturbation methods are routinely used in all fields of applied mathematics where analytical solutions for nonlinear dynamical systems are searched. Among them, normal form theory provides a reliable method for systematically simplifying dynamical systems via nonlinear change of coordinates, and is also used in a mechanical context to define Nonlinear Normal Modes (NNMs). The main recognized drawback of perturbation methods is the absence of a criterion establishing their range of validity in terms of amplitude. In this paper, we propose a method to obtain upper bounds for amplitudes of changes of variables in normal form transformations. The criterion is tested on simple mechanical systems with one and two degrees-of-freedom, and for complex as well as real normal form. Its behavior with increasing order in the normal transform is established, and comparisons are drawn between exact solutions and normal form computations for increasing levels of amplitudes. The results clearly establish that the criterion gives an upper bound for validity limit of normal transforms.

International audience; This paper investigates experimentally the quasi-static and dynamic torsional behaviour of shape memory alloys wires under cyclic loading. A specifically designed torsional pendulum made of a Ni–Ti wire is described. Results on the quasi-static behaviour of the wire obtained using this setup are presented, giving an overall view of the damping capacity of the material as function of the amplitude of the loading (imposed torsional angle), the frequency and the temperature. The dynamical behaviour is then presented through measured frequency response function between forcing angle at the top of the pendulum and the difference between top and bottom rotation angles in the vicinity of the first eigenfrequency of the wire, i.e. in the range [0.3 Hz, 1 Hz]. The softening-type non-linearity and its subsequent jump phenomenon, predicted theorically by the decrease of the effective stiffness when martensite transformation starts is clearly evidenced and analysed.

International audience; The transition from periodic to chaotic vibrations in free-edge, perfect and imperfect circular plates, is numerically studied. A pointwise harmonic forcing with constant frequency and increasing amplitude is applied to observe the bifurcation scenario. The von Kármán equations for thin plates, including geometric non-linearity, are used to model the large-amplitude vibrations. A Galerkin approach based on the eigenmodes of the perfect plate allows discretizing the model. The resulting ordinary-differential equations are numerically integrated. Bifurcation diagrams of Poincaré maps, Lyapunov exponents and Fourier spectra analysis reveal the transitions and the energy exchange between modes. The transition to chaotic vibration is studied in the frequency range of the first eigenfrequencies. The complete bifurcation diagram and the critical forces needed to attain the chaotic regime are especially addressed. For perfect plates, it is found that a direct transition from periodic to chaotic vibrations is at hand. For imperfect plates displaying specific internal resonance relationships, the energy is first exchanged between resonant modes before the chaotic regime. Finally, the nature of the chaotic regime, where a high-dimensional chaos is numerically found, is questioned within the framework of wave turbulence. These numerical findings confirm a number of experimental observations made on shells, where the generic route to chaos displays a quasiperiodic regime before the chaotic state, where the modes, sharing internal resonance relationship with the excitation frequency, appear in the response.

International audience; A numerical study of the transition from periodic to chaotic motions in forced vibrations of circular plates, is proposed. A pointwise harmonic forcing of constant excitation frequency Ω and increasing values of the amplitude is considered. Perfect and imperfect circular plates with a free edge are studied within the von Kármán assumptions for large displacements (geometric non-linearity). The transition scenario is observed for different excitation frequencies in the range of the first eigenfrequencies of the plate. For perfect plate with no specific internal resonance relationships , a direct transition to chaos is at hand. For imperfect plate tuned so as to fulfill specific internal resonance relations, a coupling between internally resonant modes is first observed. The chaotic regime shows an attractor of large dimension, and thus is studied within the framework of wave turbulence.

International audience; In this paper, we present a validation on a practical example of a harmonic-based numerical method to determine the local stability of periodic solutions of dynamical systems. Based on Floquet theory and Fourier series expansion (Hill method), we propose a simple strategy to sort the relevant physical eigenval-ues among the expanded numerical spectrum of the linear periodic system governing the perturbed solution. By mixing the Harmonic Balance Method and Asymptotic Numerical Method continuation technique with the developed Hill method, we obtain a purely-frequency based continuation tool able to compute the stability of the continued periodic solutions in a reduced computation time. This procedure is validated by considering an externally forced string and computing the complete bifurcation diagram with the stability of the periodic solutions. The particular coupled regimes are exhibited and found in excellent agreement with results of the literature, allowing a method validation.

International audience; A vibrating plate is set into a chaotic state of wave turbulence by a forcing having periodic and random components. Both components are weighted in order to explore continuously intermediate forcing from the periodic to the random one, but keeping constant its rms value. The transverse velocity of the plate is measured at the application point of the force. It is found that whatever the detail of the forcing is, the velocity spectra exhibit a universal cascade for frequencies larger than the forcing frequency range. In contrast, the velocity spectra strongly depend on the nature of the forcing within the range of forcing frequencies. The coherence function is used to extract the contribution of the velocity fluctuations that display a linear relationship with the forcing. The nonlinear contribution to the velocity fluctuations is found to be almost constant, about 55% of the total velocity fluctuations whatever the nature of the forcing from random to periodic. On the other hand, the nonlinear contribution to the fluctuations of the injected power depends on the nature of the forcing; it is significantly larger for the periodic forcing (60%) and decreases continuously as the randomness is increased, reaching a value of 40% for the pure random forcing. For all the cases of intermediate forcing from random to periodic, a simple model of the velocity response recovers in a fairly good agreement the probability density function of the injected power. The consequence of the existence of a linear-response component is discussed in the context of the fluctuation-dissipation theorem validation in experiments of out-of-equilibrium systems. © 2010 The American Physical Society.

Mots cles : Plaque stratifiee / non-linearites geometriques / flambage / vibrations / piezoelectrique Abstract - Buckling and non-linear vibrations of a piezoelectric stratified plate. Application to a MEMS mass sensor. This study proposes a model of a circular, non-symetrically laminated, piezoelectric plate, subjected to non-linear large amplitude vibrations. This model is built to simulate the behavior of a MEMS bio-sensor, designed to detect, automatically and autonomously, the presence of a given molecule in an aqueous solution. Because of both the laminated structure and the fabrication process, prestresses are observed, with different signs and intensities from one layer to another. This is responsible of a non-planar deformed geometry of the plate at rest. Moreover, experimental results show that a geometrically non-linear response is observed, with curved freqeuncy responses. A continuous non- linear model of the von-Karman type is used and discretized by an expansion of the solution onto the mode shape basis of the plate without prestresses. It's response is numerically computed by a continuation method, in order to predict the system's non-linear behavior.

The nonlinear interaction of waves in a driven medium may lead to wave turbulence, a state such that energy is transferred from large to small length scales. Here, wave turbulence is observed in experiments on a vibrating plate. The frequency power spectra of the normal velocity of the plate may be rescaled on a single curve, with power-law behaviors that are incompatible with the weak turbulence theory of Düring et al. [Phys. Rev. Lett. 97, 025503 (2006)10.1103/PhysRevLett.97.025503]. Alternative scenarios are suggested to account for this discrepancy -- in particular the occurrence of wave breaking at high frequencies. Finally, the statistics of velocity increments do not display an intermittent behavior.

International audience; Non-linear normal modes (NNMs) are used in order to derive reduced-order models for large amplitude, geometrically non-linear vibrations of thin shells. The main objective of the paper is to compare the accuracy of different truncations, using linear and non-linear modes, in order to predict the response of shells structures subjected to harmonic excitation. For an exhaustive comparison, three different shell problems have been selected: (i) a doubly curved shallow shell, simply supported on a rectangular base; (ii) a circular cylindrical panel with simply supported, in-plane free edges; and (iii) a simply supported, closed circular cylindrical shell. In each case, the models are derived by using refined shell theories for expressing the strain-displacement relationship. As a consequence, in-plane inertia is retained in the formulation. Reduction to one or two NNMs shows perfect results for vibration amplitude lower or equal to the thickness of the shell in the three cases, and this limitation is extended to two times the thickness for two of the selected models. © 2008 Elsevier B.V. All rights reserved.

19 pages; International audience; The effect of geometric imperfections and viscous damping on the type of nonlinearity (i.e., the hardening or softening behaviour) of circular plates and shallow spherical shells with free edge is here investigated. The Von Kármán large-deflection theory is used to derive the continuous models. Then, nonlinear normal modes (NNMs) are used for predicting with accuracy the coefficient, the sign of which determines the hardening or softening behaviour of the structure. The effect of geometric imperfections, unavoidable in real systems, is studied by adding a static initial component in the deflection of a circular plate. Axisymmetric as well as asymmetric imperfections are investigated, and their effect on the type of nonlinearity of the modes of an imperfect plate is documented. Transitions from hardening to softening behaviour are predicted quantitatively for imperfections having the shapes of eigenmodes of a perfect plate. The role of 2:1 internal resonance in this process is underlined. When damping is included in the calculation, it is found that the softening behaviour is generally favoured, but its effect remains limited.

International audience; A vibrating plate is set into a chaotic state of wave turbulence by either a periodic or a random local forcing. Correlations between the forcing and the local velocity response of the plate at the forcing point are studied. Statistical models with fairly good agreement with the experiments are proposed for each forcing. Both distributions of injected power have a logarithmic cusp for zero power, while the tails are Gaussian for the periodic driving and exponential for the random one. The distributions of injected work over long time intervals are investigated in the framework of the fluctuation theorem, also known as the Gallavotti-Cohen theorem. It appears that the conclusions of the theorem are verified only for the periodic, deterministic forcing. Using independent estimates of the phase space contraction, this result is discussed in the light of available theoretical framework.

International audience; The aim of the present paper is to compare two different methods available for reducing the complicated dynamics exhibited by large amplitude, geometrically nonlinear vibrations of a thin shell. The two methods are: the proper orthogonal decomposition (POD), and an asymptotic approximation of the nonlinear normal modes (NNMs) of the system. The structure used to perform comparisons is a water-filled, simply supported circular cylindrical shell subjected to harmonic excitation in the spectral neighbourhood of the fundamental natural frequency. A reference solution is obtained by discretizing the partial differential equations (PDEs) of motion with a Galerkin expansion containing 16 eigenmodes. The POD model is built by using responses computed with the Galerkin model; the NNM model is built by using the discretized equations of motion obtained with the Galerkin method, and taking into account also the transformation of damping terms. Both the POD and NNMs allow to reduce significantly the dimension of the original Galerkin model. The computed nonlinear responses are compared in order to verify the accuracy and the limits of these two methods. For vibration amplitudes equal to 1.5 times the shell thickness, the two methods give very close results to the original Galerkin model. By increasing the excitation and vibration amplitude, significant differences are observed and discussed. The response is investigated also for a fixed excitation frequency by using the excitation amplitude as bifurcation parameter for a wide range of variation. Bifurcation diagrams of Poincaré maps obtained from direct time integration and calculation of the maximum Lyapunov exponent have been used to characterize the system. © 2007 Elsevier Ltd. All rights reserved.

International audience; This study is devoted to the experimental validation of a theoretical model of large amplitude vibrations of thin spherical shells described in a previous study by the same authors. A modal analysis of the structure is first detailed. Then, a specific mode coupling due to a 1:1:2 internal resonance between an axisymmetric mode and two companion asymmetric modes is especially addressed. The structure is forced with a simple-harmonic signal of frequency close to the natural frequency of the axisymmetric mode. The experimental setup, which allows precise measurements of the vibration amplitudes of the three involved modes, is presented. Experimental frequency response curves showing the amplitude of the modes as functions of the driving frequency are compared to the theoretical ones. A good qualitative agreement is obtained with the predictions given by in the model. Some quantitative discrepancies are observed and discussed, and improvements of the model are proposed.

International audience; In order to build efficient reduced-order models (ROMs) for geometrically nonlinear vibrations of thin structures, a normal form procedure is computed for a general class of nonlinear oscillators with quadratic and cubic nonlinearities. The linear perturbation brought by considering a modal viscous damping term is especially addressed in the formulation. A special attention is focused on how all the linear modal damping terms are gathered together in order to define a precise decay of energy onto the invariant manifolds, also defined as nonlinear normal modes (NNMs). Then, this time-independent formulation is used to reduce the dynamics governing the oscillations of a structure excited by an external harmonic force. The validity of the proposed ROMs is systematically discussed and compared with other available methods. In particular, it is shown that large values of the modal damping of the slave modes may change the type of nonlinearity (hardening/softening behaviour) of the directly excited (master) mode. Two examples are used to illustrate the main features of the method. A two-degrees-of-freedom (dof) system allows presentation of the main results through a simple example. Then a water-filled circular cylindrical shell with external resonant forcing is considered, in order to show the ability of the method to substantially reduce the dynamics of a continuous structure. © 2006 Elsevier Ltd. All rights reserved.

International audience; The aim of the present paper is to compare two different methods available to reduce the complicated dynamics exhibited by large amplitude, geometrically nonlinear vibrations of a thin shell. The two methods are: the proper orthogonal decomposition (POD) and an asymptotic approximation of the Nonlinear Normal Modes (NNMs) of the system. The structure used to perform comparisons is a water-filled, simply supported circular cylindrical shell subjected to harmonic excitation in the spectral neighbourhood of the fundamental natural frequency. A reference solution is obtained by discretizing the Partial Differential Equations (PDEs) of motion with a Galerkin expansion containing 16 eigenmodes. The POD model is built by using responses computed with the Galerkin model; the NNM model is built by using the discretized equations of motion obtained with the Galerkin method, and taking into account also the transformation of damping terms. Both the POD and NNMs allow to reduce significantly the dimension of the original Galerkin model. The computed nonlinear responses are compared in order to verify the accuracy and the limits of these two methods. For vibration amplitudes equal to 1.5 times the shell thickness, the two methods give very close results to the original Galerkin model. By increasing the excitation and vibration amplitude, significant differences are observed and discussed.

International audience; The definition of a non-linear normal mode (NNM) as an invariant manifold in phase space is used. In conservative cases, it is shown that normal form theory allows one to compute all NNMs, as well as the attendant dynamics onto the manifolds, in a single operation. Then, a single-mode motion is studied. The aim of the present work is to show that too severe truncature using a single linear mode can lead to erroneous results. Using single-non-linear mode motion predicts the correct behaviour. Hence, the nonlinear change of co-ordinates allowing one to pass from the linear modal variables to the normal ones, linked to the NNMs, defines a framework to properly truncate non-linear vibration PDEs. Two examples are studied: a discrete system (a mass connected to two springs) and a continuous one (a linear Euler-Bernoulli beam resting on a non-linear elastic foundation). For the latter, a comparison is given between the developed method and previously published results.

International audience; Non-linear normal modes (NNMs), defined as invariant manifolds, are introduced through Normal Form theory. In a conservative framework, it is shown that all NNMs, as well as the attendant dynamics onto the manifolds, are computed in a single operation. The general third-order approximation of the dynamics onto a single NNM is derived. It is underlined that single linear mode truncation can lead to erroneous results which are corrected when considering NNMs. These results are illustrated by studying the vibrations of a linear beam resting on a non-linear elastic foundation.

International audience; In this article, a detailed study of the forced asymmetric non-linear vibrations of circular plates with a free edge is presented. The dynamic analogue of the von Kàrmàn equations is used to establish the governing equations. The plate displacement at a given point is expanded on the linear natural modes. The forcing is harmonic, with a frequency close to the natural frequency $\omega_n$ of one asymmetric mode of the plate. Thus, the vibration is governed by the two degenerated modes corresponding to $\omega_n$, which are one-to-one internally resonant. An approximate analytical solution, using the method of multiple scales, is presented. Attention is focused on the case where one configuration which is not directly excited by the load gets energy through non-linear coupling with the other configuration. Slight imperfections of the plate are taken into account. Experimental validations are presented in the second part of this paper.

Geometrically nonlinear vibrations of free edge circular plates subjected to a harmonic excitation are discussed. Particularly, transition from periodic to chaotic motion is observed when increasing the amplitude of the forcing. The present work is devoted to reproduce numerically these highly nonlinear behaviours. The temporal integration of such dynamics, including instabilities and chaotic regimes, is not straight forward because a stiff problem with a very large number of dofs is at hand. Consequently, numerical instabilities are observed when typical Runge‐Kutta schemes are applied. To settle the matter, two methods have been tested and compared. They both rely on a modal approach applied to the von Karman's model for large amplitude vibrations of plate. For the first one, the energies of the plates are expressed at the continuous level. The Hamiltonian of the system is then derived and discretized using the eigenmodes. The Hamiltonian formulation ensures the conservation of energy. An implicit time ...

A new step in the analysis of cymbal vibrations is presented. The goal of this analysis is to gain a better understanding of the underlying mechanisms that govern the sound of these instruments. In a previous work, geometrical invariants (correlation dimension and percentage of false nearest neighbors) were extracted from the reconstruction of the cymbal vibration in phase space [Touze et al., Proceedings of ISMA’98, Leavenworth (1998)]. Here, the Lyapunov exponents of the system are computed in order to assess the presence of chaos and to quantify it. The method used consists of approximating the tangent flow of the trajectory in the reconstructed phase space. An algorithm which uses a higher‐order polynomial approximation has been developed. This algorithm is validated on a number of well‐known theoretical time series (Henon map, Duffing equation, etc.). The Lyapunov spectrum of the cymbal shows, in particular, the convergence of the greatest exponent, which is a strong indication for the presence of chaos. A closer analysis reveals a kind of Ruelle–Takens scenario for the transition from linear to chaotic motion: Power spectra and Lyapunov exponents show that mode‐locking and quasi‐periodicity are present in the vibrations just before the chaotic regime.

National audience; L'effet Trou Noir Acoustique (TN) est une technique passive d'amortissement de vibrations sans ajout de masse fondée sur les propriétés des ondes de flexion dans des structures minces d'épaisseur variable. La mise en œuvre habituelle consiste en une plaque avec une extrémité profilée selon une loi exponentielle, recouverte d'un film viscoélastique. L'inhomogénéité de la structure conduit a une baisse de la célérité et une augmentation de l'amplitude des ondes de flexion, ce qui a pour conséquence une dissipation d'énergie efficace quand un film amortissant est placé dans la zone de faible épaisseur. L'amplitude des ondes a l'extrémité peut facilement atteindre l'ordre de grandeur de l'épaisseur de la plaque, ce qui est une source de non-linéarités géométriques. Ces non-linéarités peuvent avoir pour conséquences des couplages entre les modes de vibration de la structure, et induire un transfert d'énergie des basses fréquences vers les hautes fréquences. Ce phénomène de transfert d'énergie peut être exploité pour augmenter l'efficacité du traitement dans le domaine des basses fréquences pour lequel le trou noir est peu efficace. Une expérience montre que la terminaison TN se comporte de manière non linéaire et permet un couplage entre modes. Un régime fortement non-linéaire peut également être observé, qui est associé au phénomène de turbulence d'ondes. Une modélisation d'une poutre TN comme une plaque de von Kármán d'épaisseur variable et une résolution du problème par une méthode modale permet de confirmer les effets observés dans l' expérience et d'analyser plus finement ces résultats .

When vibrating with large amplitudes, thin structures experience geometric nonlinearity due to the nonlinear relationship between strains and displacements. Because full-order nonlinear analysis on geometrically nonlinear models are computationally very expensive, the derivation of efficient reduced-order models (ROMs) has always been a topic of interest.In this thesis, nonlinear reduction methods for building ROMs with geometric nonlinearity in the framework of the Finite Element (FE) procedure, are investigated. Three non-intrusive nonlinear reduction methods are specifically investigated and systematically compared. They are: implicit condensation and expansion (ICE), modal derivatives (MD), and the reduction to invariant manifold. Theoretical analysis shows that the first two methods can give reliable results only if a slow/fast assumption between slave and master coordinates holds. On the other hand, reduction to invariant manifolds allows proposing a simulation-free reduction method that can be applied without restricting assumptions on the frequencies of the slave modes.Numerical comparisons and numerous applications to continuous structures discretized with the FE procedure, are given subsequently. For application of the invariant manifold-based method, the computation is based on a direct application of the normal form to the physical space and hence to the nodes of the FE mesh, a method recently developed. The examples show the advantages and drawbacks of each reduction method when deriving ROM, and the results of the theoretical comparison are validated.Finally, the analysis of the dynamics of a system with 1:2 internal resonance and cubic nonlinearity is given in the last part of the thesis. The real normal form of the problem is first derived. Then the solution branches of the problem are investigated and compared to simpler solutions with the dynamics truncated at order two. The divergent behaviour of the hardening/softening characteristics for single-mode reduction is investigated with this more complete model., Lorsqu'elles vibrent avec de grandes amplitudes, les structures minces montrent un comportement non linéaire géométrique, provenant de la relation non linéaire entre les déformations et les déplacements. Les analyses des systèmes complets font appel à des calculs extrêmement couteux de telle sorte que l'établissement de modèles d'ordre réduit efficaces est un sujet d'intérêt majeur pour le calcul prédictif de vibrations de structures minces.Dans cette thèse, des méthodes non linéaires de réduction de modèle pour les structures discrétisées par la méthode des éléments finis et comportant une non-linéarité géométrique, sont étudiées. Trois méthodes non intrusives sont plus particulièrement examinées et systématiquement comparées: la méthode de condensation implicite, la méthode des dérivées modales, et la réduction sur variétés invariantes du système. Les analyses théoriques montrent que les deux premières méthodes ne peuvent donner de résultats fiables que sous hypothèse d'une séparation spectrale entre les fréquences propres des modes maitres et celles des modes esclaves. La méthode de réduction sur variétés invariantes permet quant à elle d'avoir une méthode directe, ne nécessitant pas de pré-calculs, ni d'hypothèses préalables sur les fréquences propres des modes esclaves, afin de fournir des résultats corrects.De nombreuses applications et de comparaisons numériques sont montrées sur diverses structures discrétisées avec la méthode des éléments finis. Pour appliquer la méthode des variétés invariantes, une méthode récemment développée, permet de proposer un calcul direct de la forme normale du problème, à partir de la base physique et donc des degrés de liberté du maillage éléments finis. Les exemples montrent clairement les avantage et inconvénients de chaque méthode, validant aussi les résultats théoriques montrés précédemment.Dans la dernière partie de la thèse, la dynamique non linéaire d'un système présentant une relation de résonance interne 1:2 est analysée, en tenant compte des termes cubiques. La forme normale réelle du problème est d'abord établie. Ensuite les branches de solution du problème sont analysées et comparées avec celles du modèle plus simple négligeant la non-linéarité cubique. Le comportement divergent observé lorsqu'on réduit le problème à un seul mode et que l'on cherche à prédire le comportement raidissant ou assouplissant, est ensuite étudié avec ce modèle plus complet.

The objective of this thesis is to improve the understanding of the acoustics of the piano in the context of physically-based sound synthesis. The manuscript is decomposed in three parts, the first two being devoted to the undertsanding of the origin of the double polarisation in piano string, while the third one is dedicated to the identification of sound sources of a complete piano.In the first part, the geometric (large-amplitude) nonlinearity is studied in order to understand if the nonlinear coupling can transfer energy to an initially non excited polarisation, thus leading to the double polarisation phenomenon. A multiple-scaleanalysis is conducted on a Kirchhoff-Carrier string model with fixed boundary conditions at both ends. Each polarisation is restrained to its fundamental mode, leading to two oscillors having nearly equal eigenfrequencies, and thus presenting a 1:1 internal resonance. The existence condition and stability criteria for double polarisation to occur are obtained and validated numerically based on the complete Kirchhoff-Carrier equations, as well as a more enriched third-order string model. Experiments are carried out on a monochord setup where the natural polarisation angles of the string, detuning between the two polarisations and its nonlinear behaviour are observed and identified.The second part is devoted to the string/bridge coupling. The degrees of freedom of the string are coupled to the bridge whose translational and rotational motions are respresented by a set of oscillators. The eigenfrequencies of various coupled systems are analysed. Numerical schemes are proposed and implemented where the string is solved via high-order finite-element method while the lumped bridge is solved analytically and coupled to the string by Lagrange multipliers. Experimentally, the string is strung over a bridge in a zig-zag configuration and excited vertically and horizontally. In both cases, double polarisation and double decay are observed and similar results are also obtained qualitatively in numerical models.The last part is devoted to a quantitative description of the vibroacoustic sources of a Bösendorfer 280VC-9 piano via operational transfer path analysis. The contribution of the soundboard, inner and outer rim, iron frame and lid are investigated in the frequency domain. It is found out that the soundboard is the primary contributor but the iron frame and the lid also play a significant role, especially at high frequencies., L’objectif de cette thèse est d’améliorer la compréhension de l’acoustique du piano dans le contexte de la synthèse sonore par modèles physiques. Le manuscrit est décomposé en trois parties principales, dont les deux premières ont pour but la compréhension de l'origine de la double polarisation de la corde de piano, tandis que la dernière se focalise sur l’identification de sources d’un piano complet.Dans la première partie, la non linéarité géométrique, intervenant lorsque les amplitudesde vibration sont grandes, est étudiée afin de comprendre si le couplage non linéaire peut transmettre de l'énergie à une polarisation non initialement excitée et mener ainsi au phénomène de double polarisation. Un développement en échelles multiples est mené sur un modèle de corde de Kirchhoff-Carrier avec les deux extrémités fixes, restreint au mode fondamental de chacune des polarisations. Les deux oscillateurs ont alors des fréquences très proches, on parle de résonance 1:1. La condition d’existence et le critère de stabilitépour l’apparition de double polarisation sont obtenus et validés numériquement sur la base des équations de Kirchhoff-Carrier, ainsi qu’avec un modèle de corde enrichi.Des expériences sont menées sur un dispositif monocorde où les angles de polarisation naturelle de la corde, le désaccord entre les deux polarisations et le comportement non linéaire son observés et identifiés.La seconde partie se concentre sur le couplage entre la corde et le chevalet. Les degrés de liberté de la corde sont couplés au chevalet dont les mouvements (translation/rotation) sont représentés par un ensemble d'oscillateurs. Les fréquences propres des différents systèmes couplés sont analysés. Des schémas numériques sont proposés et mis en {oe}uvre pour une résolution directe. Ces schémas résolvent les équations de corde par une méthode d’éléments finis d’ordre élevé et les équations du chevalet analytiquement. Les conditions de couplage entre corde et chevalet sont assurées par des multiplicateurs de Lagrange. Expérimentalement, la corde est tendue sur le chevalet dans une configuration de type zig-zag et excitée verticalement ou horizontalement. Dans les deux cas, les phénomènes de double polarisation et de double décroissance sont observés et des résultats qualitativement similaires sont obtenus avec les modèles numériques.La dernière partie s'attache à décrire quantitativement les différentes sources vibro-acoustiques d'un piano complet. Une étude est menée en utilisant une analyse des chemins de transfert (transfer path analysis en anglais) sur un piano Bösendorfer 280VC-9. Les contributions de la table d’harmonie, des parties interne et externe de la ceinture, du cadre en fonte et du couvercle sont étudiées dans le domaine fréquentiel. L’analyse montre que la table d’harmonie est le principal contributeur mais que le cadre en fonte et le couvercle jouent également un rôle significatif, en particulier à hautes fréquences.

Collisions between a vibrating string and a unilateral obstacle arise in a number of instruments such as tanpuras, sitars and electric basses. Contacts introduce a strong nonlinearity in the string motion, resulting in energy transfers. In this document, a numerical and experimental study of string/obstacle contacts is investigated. Three conservative numerical schemes are developed. They rely on a modal description of the string, so that a fine tuning of eigenfrequencies and damping parameters is possible for each mode. Moreover, contact and friction models are numerically treated either with a penalty approach or with a nonsmooth contact dynamics method. In the case of a point obstacle, comparisons to an analytical solution constitute a first validation of models. Experiments are then carried out in the case of an isolated string vibrating against a point obstacle, as well as in the case of a string vibrating against the neck of a fretted or fretless electric bass on which the string is installed. A good agreement is observed between numerical and experimental signals over long durations, showing the ability of models to take into account the essential physical features experimentally observed. Finally, a parametric study is led on a fretted electric bass, showing the influence of parameters such as the plucking position and the fretboard profile on the resulting sound. These adjustements are directly related to common and central issues met by guitar makers and players., Les contacts entre une corde vibrante et un obstacle unilatéral interviennent dans de nombreux instruments de musique tels que la tampoura, le sitar ou encore la basse électrique. Ils introduisent une forte non-linéarité dans le mouvement de la corde, qui se traduit notamment par d'importants transferts d'énergie. Dans ce manuscrit, une étude à la fois numérique et expérimentale du contact corde/obstacle est proposée. Trois types de schémas conservatifs sont développés. Ils s'appuient sur une description modale de la corde, ce qui permet un ajustement des fréquences propres et amortissements pour chaque mode. De plus, deux modèles de contact et frottement sont considérés, l'un régularisant et l'autre non. Des comparaisons à la solution analytique dans le cas d'un obstacle ponctuel sont menées et permettent une première validation des modèles. Des protocoles expérimentaux sont ensuite mis en place pour l'étude d'une corde isolée en présence d'un obstacle ponctuel, puis dans le cas d'une corde installée sur une basse électrique, munie ou non de frettes. Une forte concordance des signaux expérimentaux et numériques est observée sur des temps longs, montrant la capacité des modèles à rendre compte des éléments principaux observés expérimentalement. Enfin, une étude paramétrique est conduite dans le cas d'une basse munie de frettes, montrant l'influence de paramètres tels que la position de pincement de la corde et le réglage des frettes sur le son produit. Ces ajustements sont directement liés à des problématiques récurrentes et centrales rencontrées par les musiciens et les luthiers.

The aim of this thesis is to study the effect of passive nonlinear absorbers on the two degrees of freedom airfoil flutter. When an airfoil is subject to flutter instability, it oscillates increasingly until stabilizing on a limit cycle, the amplitude of which can be possibly substantial and thus damage the airfoil structure. The control has two main objectives : delay the instability and decrease the limit cycle amplitude. The flutter instability, and the post-flutter regime in particular, were studied first. A flutter experiment on a flat plate airfoil was conducted and the airfoil behavior was modeled, taking into account dynamic stall. Regarding the passive control, the first absorber studied was a hysteretic damper, realized using shape memory alloys springs. The characteristic of such dampers is their hysteretic restoring force, allowing them to dissipate a large amount of energy. Their main goal was thus to decrease the limit cycle amplitude caused by the flutter instability. This expected effect was observed and quantified both experimentally and numerically, using heuristic model. The second absorber studied was a nonlinear tuned vibration absorber. This absorber consists of a light mass attached to the airfoil through a spring having both a linear and a cubic stiffness. The role of the linear part of such absorber was to repel the instability threshold, while the aim of the nonlinear part was to decrease the limit cycle amplitude. It was found, analytically and numerically, that the instability threshold is substantially shifted by this absorber, whereas the limit cycle amplitude decrease is relatively modest., Cette thèse est consacrée à l'étude d'amortisseurs passifs non linéaires innovants pour le contrôle de l'instabilité de flottement sur un profil d'aile à deux degrés de libertés. Lorsqu'un profil d'aile entre en flottement, il oscille de façon croissante jusqu'à se stabiliser sur un cycle limite dont l'amplitude peut être significative et détériorer sa structure. Le contrôle a ainsi deux objectifs principaux : retarder l'apparition de l'instabilité et réduire l'amplitude des cycles limites. Avant d'étudier l'influence des amortisseurs passifs, l'instabilité de flottement, et notamment le régime post-flottement, a été étudié. Une expérience de flottement sur une plaque plane a été menée et sa modélisation, prenant en compte le phénomène de décrochage dynamique, a été réalisée. Concernant le contrôle passif, le premier type d'amortisseur étudié est un amortisseur hystérétique réalisé à l'aide de ressorts en alliage à mémoire de forme. La caractéristique principale de tels amortisseurs est que leur force de rappel étant hystérétique, elle permet de dissiper une grande quantité d'énergie. L'objectif principal est ainsi de réduire l'amplitude des cycles limites provoqués par l'instabilité de flottement. Cet effet escompté a été observé et quantifié expérimentalement et numériquement à l'aide de modèles semi-empiriques. Le second type d'amortisseur utilisé est un amortisseur non linéaire de vibration accordé. Il est composé d'une petite masse connectée au profil d'aile à l'aide d'un ressort possédant une raideur linéaire et une raideur cubique. La partie linéaire de ce type d'amortisseur permet de retarder l'apparition de l'instabilité tandis que la partie non linéaire permet de réduire l'amplitude des cycles limites. L'influence de l'amortisseur non linéaire de vibration accordé a été étudiée analytiquement et numériquement. Il a été trouvé que l'apparition de l'instabilité est significativement retardée à l'aide de cet amortisseur, l'effet sur l'amplitude des cycles limites étant plus modeste.

This study focuses on the characterization of plastic fatigue damage process of an austenitic stainless steel. The aim is to clarify the nature and proportions of the different stages of damage in the life of the material. Interrupted fatigue tests were established to monitor the evolution of the damage and to propose a description. At each stop, surface observations were performed for different amplitudes total strain. These observations have been completed by monitoring acoustic emission. Fatigue tests under constants amplitudes loading highlight three stages of damage. A model was proposed and compared to results obtained in fatigue under variables amplitudes loading., Cette étude porte sur la caractérisation des processus d’endommagement en fatigue plastique d’un acier inoxydable austénitique. L’objectif est de préciser la nature et la part respective des différents stades d’endommagement dans la durée de vie du matériau. Des essais de fatigue oligocyclique, interrompus, ont été mis en place afin de suivre l’évolution de l’endommagement et d’en proposer une description. A chaque arrêt, des observations en surface, pour les différentes amplitudes de déformations totales imposées ont été réalisées. Ces observations ont été complétées par un suivi par émission acoustique. Les essais de fatigue sous chargement à amplitude constante mettent en évidence trois stades d’endommagement. Une modélisation a été proposée puis confrontée aux résultats obtenus en fatigue sous chargement à amplitude variable.

Wave turbulence theory aims at describing the long time behavior of weakly non-linear, out-of-equilibrium systems. For thin vibrating plates, this framework allows predicting a Kolmogorov-Zakharov Spectrum (KZ) with an energy flux transfered from the injection to the dissipative scales along a transparency window. Previous experimental studies have pointed out some discrepancies between mesured and theoretical spectra. The fact that, in solid, damping acts at all scales, is here studied in order to explain this disagreement. By an experimental control of the dissipation, it is observed that dissipation determines the shape of spectra. Experimental measurement of the dissipation shows that damping can here be described, as a function of the frequency, by a power law. This behavior allows us to introduce directly damping in a numerical simulation of the Föppl-von Kàrmàn equations. It leads to pass from the theoretical solution KZ obtained without dissipation to spectra which are very closed to the experimental ones. These observations do not mean that wave turbulence theory should not be applied to thin plates excited by a strong forcing but encourage to extend our theoretical tools when there is no transparency window. By doing this in a phenomenological way, a new stationary solution, different from KZ and valid for any dissipation law, has been derived., Turbulence d’ondes dans les plaques minces en vibration : étude expérimentale et numérique de l’effet de l’amortissement La théorie de turbulence d'ondes a pour but de décrire le comportement à long terme de systèmes faiblement non linéaires hors équilibre. Pour les plaques minces en vibration, ce formalisme permet de prédire un spectre de Kolmogorov-Zakharov (KZ) avec un flux d'énergie transféré des échelles d'injection à celles de dissipation le long d’une fenêtre de transparence. Les études expérimentales antérieures à cette thèse ont mis en lumière des différences notables entre les spectres mesurés et les specres théoriques. La présence, dans un solide, de l’amortissement à toutes les échelles, est ici étudié afin d’expliquer ce désaccord. En contrôlant expérimentalement l’amortissement, il a été montré que la dissipation déterminait la forme des spectres. Par la caractérisation de l’amortissement, il a été trouvé que ce dernier peut être décrit en fonction de la fréquence par une loi de puissance. La dissipation expérimentale a alors pu être introduite directement dans une simulation numérique des équations de Föppl-von Kàrmàn. Cette démarche conduit à passer de la solution théorique KZ obtenue en l’absence de dissipation à des spectres très proches des spectres expérimentaux. Ces observations ne remettent pas en cause l’emploi de la turbulence d’ondes pour décrire la vibration d’une plaque mince sollicitée par un forçage d’une grande amplitude mais encouragent à étendre les outils théoriques à notre disposition en l’absence d’une fenêtre de transparence. En réalisant cette généralisation de façon phénoménologique, une nouvelle solution stationnaire, différente de KZ et valable pour toute loi d’amortissement, a pu être calculée numériquement.

Thin plate vibrations display a rich and complex dynamics that ranges from linear to strongly nonlinear regimes when increasing the vibration amplitude with respect to the thickness. This thesis is concerned with the development of a numerical code able to simulate without restrictions this large spectrum of dynamical features, described by the von Kármán equations, in the case of flat, homogeneous plates presenting a rectangular geometry. The main application of such a code is to produce gong-like sounds, in the context of sound synthesis by physical modelling. For that, a modal approach is used, in order to reduce the original Partial Differential Equations to a set of couped Ordinary Differential Equations. An energy-conserving, second-order accurate time integration scheme is developed in order to yield a stability condition. The most appealing features of the modal scheme are its accuracy and the possibility of implementing a rich loss mechanism by selecting an appropriate damping factor for each one of the modes. The sound produced by the numerical code is systematically compared to another numerical technique based on Finite Difference techniques. Fundamental aspects of the physics of nonlinear vibrations are also considered in the course of this work. When a plate vibrates in a weakly nonlinear regime, modal couplings produce amplitude-dependent vibrations, internal resonances, instabilities, jumps and bifurcations. The modal scheme is used to construct and analyse the nonlinear response of the plate in the vicinity of its first eigenfrequencies, both in free and forced-damped vibrations, showing as a result the effect of damping and forcing on the nonlinear normal modes of the underlying Hamiltonian system. When plates vibrate in a strongly nonlinear regime, the most appropriate description of the dynamics is given in terms of the statistical properties of the system, because of the vast number of interacting degrees-of-freedom. Theoretically, this framework is offered by the Wave Turbulence theory. Given the large amount of modes activated in such vibrations, a Finite Difference, energy-conserving code is preferred over the modal scheme. Such a scheme allows to produce a cascade of energy including thousands of modes when the plate is forced sinusoidally around one of its lowest eigenfrequencies. A statistical interpretation of the outcome of the simulation is offered, along with a comparison with experimental data and other numerical results found in the literature. In particular, the effect of the pointwise forcing as well as geometrical imperfections of the plates are analysed.; Les vibrations de plaques minces présentent des dynamiques très riches et variées, allant de comportements linéaire à fortement non linéaire en fonction de leur amplitude par rapport à l'épaisseur. Cette thèse présente le développement d'un code numérique capable de simuler sans restriction cette dynamique complexe, décrite par les équations de Von Karman, dans le cas de plaques homogènes et présentant une géométrie rectangulaire. L'application principale de ce code est de produire des sons de cymbales et de gongs: cette partie du travail s'inscrit donc dans le contexte de la synthèse sonore par modélisation physique. Pour cela, une approche modale est utilisée, afin de réduire les équations aux dérivées partielles à un ensemble d'équations différentielles ordinaires couplées. Un schéma d'intégration temporelle est proposé, pour lequel la conservation de l'énergie produit une condition de stabilité. Les caractéristiques les plus intéressantes du schéma modal sont sa précision et la possibilité de mettre en place des lois d'amortissements complexes à moindre coût. Le son produit par le code numérique est systématiquement comparée à celui calculé par une autre méthode fondée sur les techniques de différences finies . Les aspects fondamentaux de la physique des vibrations non linéaires sont également pris en compte au cours de ce travail. Lorsqu'une plaque vibre dans un régime faiblement non linéaire, les couplages modaux produisent des vibrations qui dépendent de l'amplitude, des résonances internes, des instabilités, des sauts et des bifurcations. Le schéma modal est utilisée pour construire et analyser la réponse non linéaire de la plaque au voisinage de ses premières fréquences propres, dans un cas conservatif, puis en prenant en compte les effets de l'amortissement et du forçage, montrant ainsi leurs effets sur les modes normaux non linéaires du système hamiltonien. Lorsque les plaques vibrent dans un régime fortement non linéaire, la description la plus appropriée de la dynamique est donnée en termes des propriétés statistiques du système, en raison de la multitude de degrés de liberté activés. Théoriquement, ce cadre est offert par la théorie de turbulence d'ondes. À cause de la grande quantité de modes activés dans de telles vibrations, le code en différence finie conservatif est préféré au schéma modal. Lorsque la plaque est excitée avec un forçage sinusoïdal autour d'un de ses modes propres de basse fréquence, une cascade d'énergie se produit, activant des milliers de modes à plus hautes fréquences. Une interprétation statistique des résultats des simulations est proposée, ainsi qu'une comparaison avec des données expérimentales et des autres résultats numériques trouvés dans la littérature. En particulier, les effets du forçage ponctuel ainsi que des imperfections géométriques des plaques sont analysées.

This study is devoted to the steelpan, a harmonically tuned percussion from Trinidad and Tobago. It consists in a steel barrel subjected to several steps of irreversible metal forming to obtain a main spherical bowl within wich convex domes are formed. This work is mainly focused on two topics. Firstly, the study of the making processes allows us to gain a better understanding of the work of the tuner in order to propose a model of the first step of steelpan making. Secondly, a vibration analysis of the steelpan, at the end of this process, is performed to understand its dynamics which is responsible of its particular tone. In the first part of the manuscript we study the evolution of the dynamical parameters during manufacturing. In particular, mode localization after note forming, strong variations of modal dampings, natural frequencies after burning and harmonic relationships after fine tuning are observed. The second part of the manuscript focusses on the first step of manufacturing during which the top of the barrel is hammered to obtain the spherical shell. The shell is analytically modeled as a nonlinear von Karman plate and the plastic strains are introduced as initial inelastic strains. The effect of the geometry and of the residual stress fields are separately quantified on the dynamical parameters of the structure in its final equilibrium state. The last part is concerned with the complex nonlinear dynamics of the steelpan at the end of the forming process. During normal playing, energy exchanges between modes are clearly identifiable: they result in a rich sound due to numerous internal resonances. Modal coupling are measured even for small amplitudes of vibrations. Original models of internal resonances are proposed.; Cette étude porte sur le steeldrum, aussi appelé steelpan, percussion mélodique de Trinidad et Tobago, fabriquée à partir de bidons métalliques dont une des faces subit un ensemble de déformations irréversibles afin d'obtenir une cuve principale à l'intérieur de laquelle sont façonnées les différentes notes de l'instrument. Cette thèse s'organise autour de deux axes. D'une part les processus de fabrication sont étudiés afin de mieux comprendre le travail du facteur et de proposer une modélisation de la première étape de fabrication. D'autre part, des études vibratoires sont menées sur l'instrument fini, pour comprendre la dynamique complexe responsable de son timbre. La première partie rend compte de l'évolution des caractéristiques vibratoires au cours de la fabrication. On remarque une forte localisation vibratoire au moment de la création des notes, des évolutions marquées des amortissements et des fréquences propres lors du chauffage, et des relations harmoniques après l'accordage final. La seconde partie se concentre sur la première étape de fabrication au cours de laquelle le haut du bidon est martelé pour obtenir une calotte sphérique concave. Ce processus de plasticité est modélisé analytiquement sous les hypothèses de von Karman par une structure mince soumise à des contraintes initiales. L'influence du changement géométrique et celle de l'état de contraintes résiduelles sont alors quantifiées sur les paramètres dynamiques de la structure dans son nouvel état d'équilibre. Enfin la dernière partie s'intéresse aux vibrations de l'instrument terminé qui présente un comportement dynamique caractéristique des systèmes non linéaires géométriques. En mode de jeu usuel, des échanges d'énergie entre les modes sont clairement audibles. Cette richesse de timbre provient des résonances internes dues à l'accordage, si bien que des couplages complexes sont mesurés pour des amplitudes vibratoires très faibles. Des modèles originaux de résonances internes sont proposés.

This study is devoted to the steelpan, a harmonically tuned percussion from Trinidad and Tobago. It consists in a steel barrel subjected to several steps of irreversible metal forming to obtain a main spherical bowl within wich convex domes are formed. This work is mainly focused on two topics. Firstly, the study of the making processes allows us to gain a better understanding of the work of the tuner in order to propose a model of the first step of steelpan making. Secondly, a vibration analysis of the steelpan, at the end of this process, is performed to understand its dynamics which is responsible of its particular tone. In the first part of the manuscript we study the evolution of the dynamical parameters during manufacturing. In particular, mode localization after note forming, strong variations of modal dampings, natural frequencies after burning and harmonic relationships after fine tuning are observed. The second part of the manuscript focusses on the first step of manufacturing during which the top of the barrel is hammered to obtain the spherical shell. The shell is analytically modeled as a nonlinear von Karman plate and the plastic strains are introduced as initial inelastic strains. The effect of the geometry and of the residual stress fields are separately quantified on the dynamical parameters of the structure in its final equilibrium state. The last part is concerned with the complex nonlinear dynamics of the steelpan at the end of the forming process. During normal playing, energy exchanges between modes are clearly identifiable: they result in a rich sound due to numerous internal resonances. Modal coupling are measured even for small amplitudes of vibrations. Original models of internal resonances are proposed., Cette étude porte sur le steeldrum, aussi appelé steelpan, percussion mélodique de Trinidad et Tobago, fabriquée à partir de bidons métalliques dont une des faces subit un ensemble de déformations irréversibles afin d'obtenir une cuve principale à l'intérieur de laquelle sont façonnées les différentes notes de l'instrument. Cette thèse s'organise autour de deux axes. D'une part les processus de fabrication sont étudiés afin de mieux comprendre le travail du facteur et de proposer une modélisation de la première étape de fabrication. D'autre part, des études vibratoires sont menées sur l'instrument fini, pour comprendre la dynamique complexe responsable de son timbre. La première partie rend compte de l'évolution des caractéristiques vibratoires au cours de la fabrication. On remarque une forte localisation vibratoire au moment de la création des notes, des évolutions marquées des amortissements et des fréquences propres lors du chauffage, et des relations harmoniques après l'accordage final. La seconde partie se concentre sur la première étape de fabrication au cours de laquelle le haut du bidon est martelé pour obtenir une calotte sphérique concave. Ce processus de plasticité est modélisé analytiquement sous les hypothèses de von Karman par une structure mince soumise à des contraintes initiales. L'influence du changement géométrique et celle de l'état de contraintes résiduelles sont alors quantifiées sur les paramètres dynamiques de la structure dans son nouvel état d'équilibre. Enfin la dernière partie s'intéresse aux vibrations de l'instrument terminé qui présente un comportement dynamique caractéristique des systèmes non linéaires géométriques. En mode de jeu usuel, des échanges d'énergie entre les modes sont clairement audibles. Cette richesse de timbre provient des résonances internes dues à l'accordage, si bien que des couplages complexes sont mesurés pour des amplitudes vibratoires très faibles. Des modèles originaux de résonances internes sont proposés.