Étude numérique de la robustesse des méthodes TPA et CB-TPA pour l’évaluation du bruit de structure dans les aéronefs induit par les systèmes vibrants

Abstract

Aircrafts are equipped with many vibrating systems. These vibrations propagate from the structure of the devices to the cabin interior panels and thus generating noise. This noise, called Structure Borne Noise (SBN), contribute significantly to the global sound level in the cabin affecting the comfort of passengers and crew members. The SBN results from the assembly of a vibration source on a receiving structure generally developed independently: the receiving structure by the aircraft manufacturer and the vibration source by a supplier. The assembly of both components lead to high level of SBN in the cabin if they are not properly developed, taking into account their assembly. It is therefore essential to use a robust methodology allowing (i) the aircraft manufacturer to deliver relevant specifications to the supplier and give him the means to verify them (ii) characterize the components independently, (iii) predict the SBN generated by the assembly from the characteristics of the decoupled substructures and (iv) modify the components in order to reduce the SBN. Component Based Transfer Path Analysis (CB-TPA) are well suited because they are based on passive properties of the components (mobilities) and intrinsic active properties of the source. Other methods, TPA Inverse Force Synthesis (IFS) allow predicting SBN from the passive properties of the receiving structure (mobilities) and the operational speeds generated by the source on assembly. TPA-IFS are regularly used and although unsuitable to design constraints in aeronautics (because they require measurements on the assembly), they will be used to validate CB-TPA methods. The TPA and CB-TPA methods are suitable for controlling SBN but are widespread in the industry due to experimental errors significantly affecting the accuracy of predicted SBN. Two types of errors associated with the characterization of the mobility of substructures are studied in this work: (i) the model error related to the number of degrees of freedom considered (DoFs) and (ii) the operator error related to “incorrect” hammer impacts. A numerical model has been developed to assess the impact of these two types of errors on the robustness of the TPA and CB-TPA methods for various complexities of assemblies and vibrational excitations. The results show that the translation along the axis normal to the contact allows for perfectly predicting the velocity at a target point on the receiving structure in the normal direction in the case of simple assembly and vibratory behavior (beam assembly and pure normal internal excitation). In the more complex cases (i. e., source which générâtes internal moments and plate and source with a high modal behavior), the predictions are globally in good agreement with the reference but localized errors in frequency can be observed and can lead to large discrepancies in the predictions of SBN for tonal sources (such has hydraulic pumps). For exact predictions, the TPA-IFS methods require the terms of high amplitude of the mobility of the receiving structure and the CB-TPA those of the mobilities of the two substructures (when the mobility ratio is close to 1). Operator errors appear to be less critical than model errors. Mobility at low amplitudes does not improve predictions and increases the effect of operator errors. For obtaining robust predictions, it is necessary to identify and remove the operator errors from calculations.

Date	19 Aug 2020
Original language	French
Awarding Institution	École de technologie supérieure
Supervisor	Olivier Doutres (Supervisor) & Thomas Dupont (Co-supervisor)