21 Preliminary Results of Vibration Measurements on a Wind Turbine Test Bench 219 0 5 10 15 20 25 30 10-4 FRF [(m/s2)/N] 0 5 10 15 20 25 30 0 0.5 1 Coherence 0 5 10 15 20 25 30 Frequency [Hz] 1000 2000 3000 4000 5000 Force spectrum [Ns] (a) 0 5 10 15 20 25 30 Frequency [Hz] 10-3 10-2 10-1 100 101 102 Acceleration PSD [(m/s2)2/Hz] (b) Fig. 21.9 Frequency response function, coherence and force spectrum from impact test and PSD from operational measurements for a geophone in DOF 10 which is on the concrete wall more thorough excitation of the modes which is also shown in the plots where the peaks are more defined. The frequency values for operational measurements are for that reason thought to be more precise. For impact tests when the DUT is not attached, the peak at 13.7 Hz is shown for the concrete walls but not for the rest of the sensors. When the DUT is attached both for operational measurements and impact tests, the peak at 13.7 Hz are visual for all sensors. The vertical sensor at DOF 5 showed the same peaks at 3.9, 9.8 and 13.7 Hz for the operational measurements as were seen the horizontal sensors. This could indicate a rigid body mode for the concrete which had both horizontal and vertical movement. Another aspect that supports assumption that the low frequencies are rigid body modes, were that the eigenfrequencies both with and without a DUT remain the same for the experimental vibration tests. This supports the assumption that the frequencies could be rigid body modes of the entire concrete structure. If it is not a rigid body mode it means that the DUT does not change these low frequency modes for the HALT tester. This has to be examined further, and more measurements have to be performed on the concrete in order to determine if it is a rigid body mode. 21.4 Conclusion The HALT tester is a large and complex structure for which satisfactory measurements are difficult to obtain. Especially impact tests were difficult to conduct due to the size and complexity of the machine. Different adjustments were made to improve the tests. Two types of sensors were used and for impact tests two sizes of impact hammers were tried. A comparison between the hammers was made and it was concluded that the sledgehammer performed best with better coherence and a FRF with better details. Accelerometers and geophones were compared and showed that the geophones produced the best results. The geophones showed a better coherence overall. It especially outperformed the accelerometer when the excitation level was low, due to the low noise floor of the geophone. The FRF for the geophones had better defined peaks which were easier to interpret. For operational measurement the accelerometers on the concrete wall showed clear peaks where the rest performed poorly. This indicates that the concrete foundation was better excited than the steel parts during operation. Frequencies for the first 3–4 peaks are shown in Table 21.1. Operational measurements showed an additional peak compared to the impact tests at a lower frequency at 1.5 Hz. The peak at 3.9 Hz can be compared across the two test types and is consistent at 3.9 Hz, both with the DUT attached and without it. This could indicate that this frequency could be a rigid body mode for the concrete foundation, which has to be examined further. The peak after the 3.9 Hz varies depending on test and which sensors for the impact tests. The peak at 13.7 Hz is consistent in frequency but is not visual on the Hexapod when the DUT is not attached.
RkJQdWJsaXNoZXIy MTMzNzEzMQ==