Figure 6: 20th Mode of the Cylinder + transmission simulator. To see whether these modes were truly insignificant for the substructure model, the full system described in [2] was assembled and its frequency response functions were calculated. Figures 7 and 8 show the drive point FRF of the coupled system at point 301z, which is shown in Fig. 4. Three cases are shown. The solid blue line labeled “analytical” is the FRF of a finite element truth model. The green dashed line is the FRF predicted by substructuring using all of the modes of the cylinder (the C system), and is the same result that was reported in [2]. One can see that the substructuring prediction agrees very well with the truth model. The red dash-dot line is the substructuring prediction obtained after deleting modes 19-22 and mode 26 of the cylinder, as just described. The truncated substructuring prediction also agrees very well with the truth model, except near a few peaks at 1300 and 2300 Hz. The region where the discrepancies occur is magnified in Fig. 8. The FRFs of the system in the axial direction were also shown in [2], but they are not repeated here since there was no discernible difference between the predictions when modes 19-22 and 26 were deleted. These results seem to confirm that these modes are not very important to the subcomponent model. Even then, one would prefer to avoid decreasing the accuracy of the subcomponent model, but for this system none of these problematic modes can be discarded without degrading the substructuring predictions to some extent at least. 129
RkJQdWJsaXNoZXIy MTMzNzEzMQ==