FEM/SEA Hybrid Method for Predicting Mid and High Frequency Structure-Borne Transmission

• Published in: The open acoustics journal Vol. 2; no. 1; pp. 45 - 60
• Main Authors: Troclet, B., Hiverniau, B., Ichchou, M. N., Jezequel, L., Kayvantash, K., Bekkour, T., Mouillet, J. B., Gallet, A.
• Format: Journal Article
• Language: English
• Published: 01.01.2009
• Subjects: Finite element method; High frequencies; Influence coefficient; Joining; Mathematical models; Methodology; Statistical energy analysis; Transmissions (automotive)
• DOI: 10.2174/1874837600902010045
• ISSN: 1874-8376

Launch Vehicles are subject, at lift-off and during flight ascent, to acoustic and aeroacoustic loads, which are random in nature. Because electronic units are very sensitive to mid and high frequency loads, it is important to numerically predict and specify the vibration levels to be applied to units for qualification test. The general objective of the activity presented in this paper is to develop a methodology to predict mid and high frequency structure-borne transmissions in launch vehicles. As the loads of interest are random, it has been chosen to investigate energy-based modeling approaches, combined with the Finite Element Method. For energy-based modeling, the structure is divided into subsystems. For high frequency predictions, the purely numerical Power Injection Method, derived from Statistical Energy Analysis, is used to estimate the Coupling Loss Factor between structural subsystems. For the mid frequency predictions, an approach close to Statistical Energy Analysis, called Statistical Energy Analysis-Like (SEA-like), is investigated. In this approach, a relation between total energies of subsystems and input powers is established, by identifying a matrix composed of Energy Influence Coefficients. The objective of the study is to establish the methodology to compute with accuracy, using the Finite Element Method, Coupling Loss Factors and Energy Influence Coefficient. It is shown that the excitation of subsystems by 'Rain on the Roof' loads defined by the 'Influence Circle' and the Optimal Latin Hypercube methods provide accurate coupling data. A validation of the methodology on academic and industrial cases is presented.