Identification et optimisation des propriétés dynamiques des matériaux viscoélastiques
In the automotive, railway and aerospace industries, interior noise is an important consideration in design and operation. Among the available technologies to reduce the structure-borne vibration and noise, the use of Metal-Polymer Sandwich (MPS) panels is attracting more interest from Original Equipment Manufacturers (OEMs). As for constrained-layer damping (CLD) treatments, besides developing more accurate models (theoretical and finite element) to simulate the vibroacoustic performance, it is very important to 'accurately identify the properties of the constituent materials of an MPS. Since the core materials in MPS exhibit viscoelastic properties which vary significantly with temperature and frequency, it is necessary to develop experimental and/or optimization methods to characterize the dynamic properties so that they may be well matched to specific noise and vibration control applications. This is the objective of this thesis.In this thesis, a simple free-free beam based setup, together with an identification algorithm has been developed to identify the dynamic properties of core materials from the measured frequency response functions. The setup involved circumventing some drawbacks of the traditional clamped-free setup.In particular, a new optimization method is brought' forward wherein a four-parameter fractional derivative model plus a three-parameter Williams-Landel-Ferry (WLF) equation are used to describe the temperature and frequency dependent behaviour of core materials. Therefore, few parameters are optimized for the temperature and frequency dependent properties. The objective function in the optimization is based on the so-called amplitude correlation coefficient which can be calculated directly by the frequency response functions. The normal mode superposition method taking the added mass into account, as well as Ross-Kerwin-Ungar (RKU) equations, as a solver, is used to calculate the predicted frequency response functions. The Pattern Search algorithm is used to find the best values of design parameters. This algorithm is a global optimization algorithm and is less sensitive to the initial values of design parameters. Numerical examples and tests on several MPS panels were used to validate the free-free setup and optimization method by systematic comparison with the ASTM E756-04 Standard and with DMA when the latter is possible. However, with some MPS panels, the proposed method failed to provide satisfactory results. It was further postulated that the manufacturing process of these MPS panels may somehow have modified the properties or the constitutive law of the polymer itself.
- Génie – Thèses