REVIEW OF DATA REDUCTION TECHNIQUES FOR THE INTERPRETATION OF DYNAMIC QUALIFICATION WITH SWEPT-SINE EXCITATION

摘要

The dynamic qualification procedure for spacecraft structures usually includes broadband base excitation using vibration tables on an acceleration level defined by the launcher. In most cases, the frequency range of interest is covered by a logarithmic sweep. In order to prevent the structure from being overtested, low and intermediate level runs are performed before the actual qualification run. The latter is tailored to fulfil the qualification requirements based on preceding lower-level runs. The load level of the structure is generally measured by means of accelerometers sometimes using additional information from other sensor types like strain gauges. In order to define the high-level swept-sine runs it is necessary to compute the applied loads on the structure as a function of frequency. It is, thus, necessary to transform the recorded time domain measurements in a reliable way to frequency domain to obtain a realistic picture of the excitation level of critical parts of the structure. This transformation can be carried out by several means performing Hilbert transform of the time domain data or calculating RMS values and replacing the time axis by an equivalent frequency axis or using standard procedures of spectral analysis based on the Fourier transform. The paper critically reviews the benefits and problems of these procedures when using them to analyse data from swept-sine base excitation. The procedures will be thoroughly checked by using simulation data for a numerical 1 Degree of Freedom (DoF) system and a weak non-linear beam. Furthermore, they will be applied to data from the qualification test campaign of the Automated Transfer Vehicle (ATV) in 2003.The results for the several data reduction techniques from the numerical simulations, a simple beam structure and the test on the ATV are compared and evaluated. The dependency on noise, sweep-velocity and non-linearities is discussed. By means of these investigations suggestions for future measurement campaigns are made.