Identification of statistical energy analysis parameters based on ERA

Statistical energy analysis (SEA) is the most effective tool for the solution to high-frequency dynamical problems of complex structure. However, a reliable solution can be obtained with SEA under the premise that SEA parameters in SEA equations are estimated exactly. A method for identifying the parameters from experimental data was proposed here. The methodology was based on identifying the minimum order power flow model utilizing the eigensystem realization algorithm(ERA), then the eigenpairs of the identified model were used to update an initial computation SEA model having a discrepancy with the true model. The SEA parameters were extracted from the updated SEA model ultimately. Simulations were performed for a three-subsystem model at different noise-levels, and the results showed that the methodology presented here works quite satisfactorily. Another analogous simulation was performed for a micro-satellite and the results indicated the methodology can be extended to identify the SEA parameters of a complicated model.