Robust estimation of coupling loss factors from finite element analysis

There are well-established techniques by which the coupling loss factors (CLFs) of statistical energy analysis (SEA) can be estimated from finite element analysis (FEA). These are typically based on a single, selected system. A slightly different choice of system would give different estimates. There is a need for robust methods that give good estimates of the SEA average CLFs, independent of the details of the chosen system. Estimates of variance, confidence limits, are also of interest. Two approaches to this problem are discussed. These involve attempts to randomise the properties of the system and averaging the resulting estimates, but without repeating the full FEA. The first involves perturbation of component modal properties which can be related to perturbations in the modes of the assembled structure and hence to the energies and CLFs. In the second approach, it is assumed that the statistics of the modal properties of the system analysed are a fair representation, when taken over a wide enough frequency range, of the statistics of the modes of the SEA ensemble. The modes of the system are then randomly sampled to provide robust estimates. Numerical examples are presented. The methods are computationally very cheap.