Abstract
In this paper, a full methodology to deal with microvibration predictions onboard satellites is described. Two important aspects are tackled: 1) the characterization of the sources with a pragmatic procedure that allows integrating into the algorithm the full effect of the sources, including their dynamic coupling with the satellite structure; 2) the modeling of the transfer function source receivers with a technique named in this paper as the Craig-Bampton stochastic method, which allows prediction of a nominal response and variations due to structural uncertainties as accurate as full Monte Carlo simulations but at a fraction of the computational effort. The paper then includes a statistical study of the data from the structural dynamic testing of the five identical craft of the Rapid-Eye constellation to set the magnitude of the uncertainties that should be applied in the analysis. Finally, the computational procedure is applied to the new high-resolution satellite SSTL-300-S1 and the predictions compared with the experimental results retrieved during the physical microvibration testing of the satellite, which was carried out at the Surrey Satellite Technology Limited facilities in the United Kingdom.