Using Numerical Models within an SEA Framework

Statistical Energy Analysis (SEA) is widely used for modeling the vibro-acoustic response of large and complex structures. SEA makes simulations practical thanks to its intrinsic statistical approach and the lower computational cost compared to FE-based techniques. However, SEA still requires underlying models for subsystems and junctions to compute the SEA coefficients which appear in the power balance equations of the coupled system. Classically, such models are based on simplified descriptions of the structures to allow analytical or semi-analytical developments. To overcome this limitation, the authors have proposed a general approach to SEA which only requires the knowledge of impedances of the structures to compute SEA coefficients. Such impedances can always be computed from an accurate FE model of each component of a coupled system. This yields to a numerical SEA framework able to account for the dynamic complexity of the FE model without any a priori assumption on the type of waves traveling in the subsystem. In turns, this allows to (i) extend the range of applicability of SEA towards lower frequencies thanks to the higher fidelity models used to compute the power coefficients and (ii) improve accuracy in the mid- and high-frequency domains for complex structures. Moreover, the computational cost is kept low thanks to the substructuring approach which is inherent in SEA. Representative examples are considered to show the capabilities of the proposed numerical SEA framework.