334 T. Marinone et al. Fig. 31.5 Drive point FRFs from impact runs, showing mass lines only the six rigid body modes. The synthesized CMIF shows a high level of correlation with the test data. As such, there is confidence that the extracted rigid body modes will accurately describe the inertia properties of the hub as well. Figure 31.5 shows the drive point measurements for all of the DOFs where the six rigid body modes are evident in the peaks seen in the responses. There are clear mass lines from 20 to 50 Hz, showing good separation between the rigid body and flexible modes. As seen in the zoomed-in portion, however, there is a small but noticeable oscillation in some of the mass lines, which is associated with experimental error caused by double impacts at some of the less-repeatable drive points. As a result, there is expected to be some error involved in the estimation of the terms using this method. The inertia results from modal inversion for the nacelle are shown in Table 31.1. The masses of the nacelles were also measured independently by a crane company during transportation, and these are included for completeness. The mass and CG properties of nacelle 3 were also obtained, but modal testing was not performed on that nacelle. The two nacelles have fairly consistent properties, which were expected due to the assumed similarity in manufacture of the nacelles. The CG of the nacelles in the X direction (fore-aft) is aft of the origin; however, the CG measured in the modal inversion method is closer to the origin than that measured with load cells. The aft CG is expected, as the nacelle should be balanced when the rotor is assembled: without the rotor, the nacelle will be significantly back-heavy. The CG about the Y direction (side-side) is relatively close to the origin, although computed on the opposite side for nacelle 2 using the modal inversion method. The offset is expected to be positive due to the generator being primarily located on the positive-Y side of the nacelle. Substantial rigging was used when the nacelles were being lifted by the crane, which accounts for some of the discrepancy in the masses obtained between the crane, load cell, and modal inversion methods. Additionally, oil and equipment were installed after the initial inertia testing, changing the mass and CG of the nacelle. The results obtained from the crane and the load cell are assumed to be more accurate than the modal inversion method, as those methods obtained the loads directly from the calibrated load cells, in contrast to estimating them from the modal mass of the rigid body shapes. The error seen using modal inversion may be due to error in measurement, curve-fitting, or accelerometer location measurement. However, the mass and CG are within 20 % of the load cell results, which indicates that the moment of inertia terms are reasonably accurate. Finally, the moment of inertia about the Y-axis is the largest, which is expected since the mass is most spread out about the Y-axis.
RkJQdWJsaXNoZXIy MTMzNzEzMQ==