SEA substructuring using cluster analysis: The MIR index
Résumé
Statistical Energy Analysis (SEA) is a powerful tool to describe energy sharing in a complex structure. To apply this method, the structure must be divided into subsystems. One of the main drawback of this method is the difficulty in finding subsystems which verify SEA assumptions. Moreover, the weak coupling between subsystems is not well defined despite it being one of the most important assumptions of SEA. This paper deals with the derivation of a tool to aid decision-making in SEA substructuring. It is based on a cluster analysis to classify energy transfer functions and on principal component analysis (PCA) to reduce data size. The method is applied to three test cases: an L-shaped plate, an assembly of three plates and a complex thin shell structure.