Abstract
The Finite Element Method (FEM) has become the most utilized tool to carry out structural analysis. It is implemented in various software packages which are commonly used in Industry. FEA gives accurate predictions up until the first few structural modes of vibration, where the behaviour of real structures is quite deterministic. In the high frequency range statistical approaches are more suitable, and here Statistical Energy Analysis (SEA) has been applied quite successfully. In the mid-frequency range FEM predictions start to become unreliable, and SEA is not applicable as some of its basic assumptions are not verified. This paper has been developed in the context of a project concerning analyses of transmission of micro-vibrations in satellite structures. In addition to the ones related to the mid-frequency range, micro-vibrations introduce other issues: being very small entities, the related uncertainties are more substantial. Because of the large bandwidth of the frequency range related to micro-vibrations, their modelling and analysis pose a challenge, in particular in the mid-frequency range, where many of the micro-vibration sources on board a spacecraft tend to act. In this context, this paper will deal with two different aspects: on one side the development of a method aimed at reducing the computational effort nowadays involved to overtake the mid-frequency issue (we propose to merge the efficiency of the Craig-Bampton reduction with the simplicity and reliability of the Monte Carlo Simulation for the various subsystems to produce an overall analysis algorithm); on the other side the validation of finite element models of satellite structures (the effectiveness of Modal Assurance Criteria and Normalised Cross Orthogonality on the response prediction of spacecraft models is here carried and a new criterion, Base Force Assurance Criterion, is defined using the experimentally determined dynamic force at the base and the finite element predicted force). The method (Craig-Bampton Stochastic Method, CBSM) will be described in this article; with a benchmark example shown. A proof of the validity of the method in a real industrial application will also be presented, which will be performed by comparing the results obtained in applying the CBSM and the MCS to results obtained during an experimental campaign. This campaign has been carried out on the spacecraft SSTL 300 (made available by the company Surrey Satellite Technologies Limited in Guildford, UK).