192 E. Mola et al. • accelerometer piezo-electric PCB393B31 (channels 5, 7, 11, 13, 20, 22, 24, 25, 26, 27) • accelerometer piezo PCB 393A03 (channels 9, 21) • accelerometer piezo-electric PCB 393B12 (other channels) All signals were recorded using a 24 bit A/D converter with anti-aliasing filters built inside, with a sampling frequency equal to 2048 Hz. 23.4.2 Test Results and Identification Issues The technique used for the extraction of modal parameters from the output data only is the Polyreference Least Squares Complex Frequency Domain method. One of the specific advantages of the technique lies in the very stable identification of the system poles and participation factors as a function of the specified system order, leading to easy-to-interpret stabilization diagrams, [4, 5]. Hence it keeps a rather good accurate performance to identify the parameters for small damping and dense modal system by using stabilization diagram. Figure 23.6 shows the stabilization diagram obtained. Since the first three/four vibration modes are enough for the model validation, a frequency band from 1 to 3 Hz with a rather high mode density was chosen (Fig. 23.6). From the stabilization diagram, the poles marked by arrows are very stable, so these modes can be identified in a reliable and robust way [6, 7]. The averaged Power Spectral Densities (PSD) of Accelerometer 2, 5 and 6 have been computed and showed in Fig. 23.7. It is clear that auto-spectra in Fig. 23.7 have the first significant peaks from 1 to 3 Hz, which reveals that the structure has very closely spaced modes, in some cases just with a 0.05 Hz gap among them. Because some modes, for instance the two modes marked by arrows in Fig. 23.7, are highly damped and close to the next peaks (1.27 and 1.36 Hz respectively), it is rather challenging to identify these modes. Table 23.2 shows the first identified natural frequencies and corresponding damping ratios. It must be noticed that this table only shows modes for which the identified eigenmodes were reliable. Figure 23.8 shows the mode shapes correctly identified and their corresponding simulations. Figure 23.8 shows the experimental identified mode shapes and numerical identified mode shapes have a high degree of correspondence and the identified frequencies have a good match with the numerical ones. Fig. 23.6 MI-TO way: stabilization diagram
RkJQdWJsaXNoZXIy MTMzNzEzMQ==