82 S. Marwitz and V. Zabel a1 b1 c1 d1 a2 b2 c2 d2 a3 b3 c3 d3 a4 b4 c4 d4 a5 b5 c5 • d5 366 366 192 200 200 200 188 SLDVTop Accx Accy x y z a x y 366 60 60 366 60 60 5 × 5 M8 z b SLDVRight SLDVLeft Fig. 7.4 3D view of the test structure with dimensions in mm. (a) Overview, labeling of points. (b) Detailed view of one platform of the tower c5 b3 b4 a3 c2 b1 a5 c4 a1 a2 c1 b5 a4 c3 b2 c5 a2 c1 b5 a4 c3 b2 a1 b3 b4 a3 c2 b1 a5 c4 c5 a2 c1 b5 a4 c3 b2 a1 b4 a3 c2 b1 a5 c4 b3 c5 c5 c5 Full Set SLDV1 SLDV2 SLDV3 Set 1 Set 2 Set 3 Ref. Meas. Fig. 7.5 All setups used in the experimental study, refer to Fig. 7.4 for labels of the measuring points The final experimental setup for the application of the proposed approach was created based on the results of the preliminary measurement. The reference locations were chosen to be at the same point, namely point c5, for all scan heads in order to reduce influencing factors on the final results. A sufficient representation of all modes of interest in each single setup was the objective for combining moving points into setups. The resulting 22 setups are shown in Fig. 7.5. An independent positioning of the laser beams was achieved by utilizing the macro programming capabilities of the software provided by the manufacturer. Another aim of the experimental application was the comparison and evaluation of the performance of the three multisetup merging strategies presented in Sect. 7.2.3. Therefore ambient excitation (i.e. non-stationary power spectrum) was required and achieved through the use of two moving ventilators, exciting the structure aerodynamically. The vibration velocities were recorded at a sampling frequency of 512 Hz for a duration of 256 s. However prior to the analysis the data was decimated by a factor of 2. The analysis was performed with an own implementation of the SSI cov/ref algorithm, with parameters i D200 (number of block rows) and nmax D200 (maximum model order). After the manual selection of stable poles for each mode of interest, the setups were merged following the PoSER strategy. In the case of 22 setups this is a time consuming process. To evaluate the performance and accuracy of the other two merging strategies, those were applied as well. In that case all setups belonging to one set were merged following the PoGER and PreGER approach, respectively, and subsequently the three sets were merged using the PoSER algorithm. Finally, the coordinates of the mode shapes were transformed to Cartesian coordinates. That allowed not only for the visualization, but also for the comparison with the mode shapes obtained from the preliminary measurement.
RkJQdWJsaXNoZXIy MTMzNzEzMQ==