Ce mémoire est consacré à l'étude d'un absorbeur dynamique non linéaire hybride passif-actif (ADNLH) pour la réduction du bruit en basses-fréquences. La partie passive de l'ADNLH est une membrane en latex à déformée non linéaire dont la face avant est couplée au champ acoustique que l'on souhaite réduire. Cette membrane se comporte comme un oscillateur non linéaire et fait partie de la famille des absorbeurs non linéaires connus sous le nom de Nonlinear Energy Sink (NES). La face arrière de la membrane est encoffrée et un dispositif de contrôle actif est inclus dans le volume d'encoffrement. Ce dispositif est conçu pour modifier la raideur linéaire et l'amortissement de la membrane. Des précédents travaux ont été réalisés uniquement sur la partie passive (la membrane) et ont permis de valider le principe de pompage énergétique dans le domaine acoustique. Cependant la membrane seule possède des limitations (notamment le seuil de déclenchement du pompage) qui restreignent les applications possibles. L'objectif de l'ADNLH est d'améliorer les performances du pompage énergétique acoustique en modifiant les propriétés linéaires de la membrane grâce à la boucle d'asservissement. Dans un premier temps une étude théorique et expérimentale est réalisée sur l'ADNLH. L'ADNLH est ensuite couplé à un tube résonant avec une excitation sinusoïdale et en bruit blanc. Il permet bien d'écrêter le premier pic de résonance du tube avec de meilleures performances que la version passive. Enfin l'ADNLH est installé dans une salle peu amortie. Il permet d'atténuer la première résonance acoustique de la salle dans le cas d'une excitation sinusoïdale. / This thesis is devoted to the study of a hybrid passive-active nonlinear dynamic absorber for the reduction of noise in low frequencies. The passive part of the ADNLH is a membrane in latex with a nonlinear deformation and its front face coupled to the acoustic field to be reduced. This membrane is acting as a nonlinear oscillator and is part of the family of absorbers known as Nonlinear Sink Energy (NES). The rear face is enclosed and a active device is included inside this enclosure. This device is designed in order to modify the linear stiffness and the damping of the membrane. Previous work has been done only on the passive part (the membrane) and has validated the principle of energy pumping for Acoustics. However the membrane has some limitations (like the threshold of energy pumping) that restrain the practical applications. The goal of the ADNLH is to improve the performance of the energy pumping by modifying the linear properties of the membrane with the help of the active device. In a first time an experimental and theoretical study of the ADNLH is done. Then the ADNLH is coupled to a tube of air thanks to an academic assembly under a sinusoidal excitation or broadband. It allows to cut the top off the first acoustic resonance of the tube with better performances than the membrane alone. At last the ADNLH is set inside a weakly damped room. The ADNLH allows to attenuate the first resonance of the room in the case of a sinusoidal excitation. One also shows that the control of the damping of the membrane is the key parameter for the performance of the ADNLH.

Auxiliary absorbers provide an effective means to attenuate the vibrations of a structural or mechanical system (the "primary structure"). The simplest auxiliary absorber, a tuned mass damper (TMD), provides reliable narrow-band attenuation but is not robust to the effects of detuning. Strongly nonlinear tuned mass dampers (NTMDs) are capable of wide-band, irreversible energy transfer known as "energy pumping" but can also exhibit high amplitude solutions which significantly amplify the response of the primary structure. Semi-active tuned mass dampers (STMDs) incorporate an actuating element in order to achieve real-time tuning adjustment capability. This thesis presents a global dynamic analysis of the response of a primary structure with an NTMD and then explores the performance of a novel absorber configuration consisting of an NTMD and STMD attached to the primary structure in series. The global dynamic analysis is conducted using a new cell mapping method developed by the author and introduced within the thesis: the parallelized multi-degrees-of-freedom cell mapping (PMDCM) method. The benefits of the additional STMD component are explored for two distinct applications: (1) restoring the performance of a linear TMD which develops a weak nonlinearity due to operation outside of the intended range or other means, and (2) acting as a safety device to eliminate or minimize convergence to the detached high-amplitude response. In the weakly nonlinear case, the STMD is shown to reduce the effects of the nonlinearity and improve attenuation capability by constraining the motion of the NTMD. In the strongly nonlinear case, the STMD effectively eliminates the complex response behavior and high amplitude solutions which were present in the original system, resulting in a single low amplitude response. Experimental tests using an adjustable-length pendulum STMD verify the numerical results.