250 P. Lubrina et al. Fig. 24.8 Example of amplitude-modulated excitation signal run (ref. [5–7]). The SVDP defines a mathematical construction of a virtual driving point, which would have given rise to vibratory responses strictly similar to those obtained with correlated forces. SVDP relies on the equivalent complex power P .!/ D X shakers Fs .!/ PXs .!/ P .!/ DFV .!/ PXV .!/ whereFs(!) is a excitation force acting on a driving point s, PXs .!/ the velocity at driving point s, Fv(!) the virtual force and PXv .!/ the velocity response of the virtual driving point. Once the SVDP process has been applied, SIMO FRFs are obtained and classical curve-fitting can be directly used on them (Figs. 24.9 and 24.10). 24.7.3 Modal Model Assembly Considering a linear structural behaviour, it would be sufficient to use only very few excitation points to excite all modes of an aircraft. However the practical application shows that several excitation configurations are needed during GVT: vertical and lateral engine excitations, vertical and axial wing excitations, HTP excitation, VTP excitation: : : The general goal is to put as much energy as possible per mode, i.e. to increase the level of generalized force until maximum per mode. These numerous tests are mandatory for optimising the reliability of experimental modal model and taking into account nonlinear structural behaviour. In practice, for each excitation configuration, several runs are performed at different levels of excitations. From all these runs, each structural mode can be identified a significant number of times. During the modes sorting and filtering process, the whole set of modes identified by curve-fitting is carefully analysed by structural engineers and sorted by nature. All identified modes are stored into a database system with multi user access. Each mode is stored not only with its modal properties but also with numerous fields containing meta information. A specially designed software tool called “Correlation Tool” was developed to review the modes in the database. The Correlation Tool can be installed on different computers, even on the customer computer to give online access (read only) to the current modal data. One feature of this database software is that modes which have been identified from different FRF datasets with almost identical properties can be grouped in a mode family based on MAC correlation. For each family, the most representative mode is selected as a member of the final modal model delivered to Airbus. To support the process of correlation of modal datasets and finally the generation of the final modal model different quality indicators and other criteria are applied, for example, level of excitation, generalized force and value of Mode Indicator Function (MIF) are used here. The concept of mode families can also be applied to evaluate scatter on test results or even to analyse the results in terms of non-linear behaviour. If the members of a mode family are considered to be reliable enough (i.e. confidence in the results assessed by quality indicators), they can become affiliated to a “master mode” and their damping ratios and eigenfrequencies can be plotted as a function of force level or other parameter of the database. In this GVT, the work of modal correlation was a specific challenge. Finally, the huge amount of data was condensed down from about 2,600 poles identified from all FRF
RkJQdWJsaXNoZXIy MTMzNzEzMQ==