36 Producing Simulated Time Data for Operational Modal Analysis 345 Fig. 36.6 Stabilization diagram from the experimental test of the Plexiglass plate. ‘C’ mean stable pole and mode shape, ‘ı’ unstable and ‘˘’ only stable frequency. The solid line in the background is the power spectrum from DOF 1 (Color figure online) 0 200 400 600 800 1000 0 20 40 60 80 Frequency [Hz] Number of Poles 0 500 1000 1500 10−10 100 1010 Frequency [Hz] Principal Components 0 500 1000 1500 10−10 100 1010 Frequency [Hz] Principal Components 0 500 1000 1500 10−10 100 1010 Frequency [Hz] Principal Components 0 500 1000 1500 10−10 100 1010 Frequency [Hz] Principal Components a b c d Fig. 36.7 Principal components of the responses (a) for (Sim. 1) the single DOF excitation case. The first principal component is approx. 60 dB higher than the next, indicating a single independent source as expected. (b) For (Sim. 2) the case of independent excitation of all DOFs. Ten principal components of similar level can be observed (approx. 100 dB higher than the rest), indicating ten independent sources. The reason why 35 independent sources are not observed is that model has only ten modes. (c) For (Sim. 3) the case of excitation of all DOFs randomly mixed of 35 independent sources. Ten principal components of equivalent level can be observed (approx. 90 dB higher than the rest), indicating ten independent sources. (d) Principal components for (Sim. 4) the case of excitation of all DOFs randomly mixed of four independent sources. Four principal components are observed to be at an equal level, with the next one approx. 30 dB below
RkJQdWJsaXNoZXIy MTMzNzEzMQ==