10 Source Characterization for Automotive Applications Using Innovative Techniques 121 Fig. 10.3 Sensor consistency of (left) raw FRF data and (right) rigidified FRF data The FRFs were then transformed to a single 6-DoF virtual point and projected back to the original degrees of freedom, as described above. This set of rigidized FRFs was merged with the original FRFs with a cutoff frequency centered at 25 Hz. The sensor consistency of the new set of FRFs is presented on the right side of Fig. 10.3. As expected, the “filtered” FRFs behave very rigidly and consistently at the low frequencies, while leaving the sensor consistency at higher frequencies unaffected. An operational test with broadband content was then performed to do in-situ source characterization and TPA. Of the 32 sensors used in the test, 30 were used as indicators (u4) and 2 were used as validation sensors (u3). A total of 40 interface DoF were considered. The blocked forces were calculated using both the raw FRFs and the rigidified FRFs. These blocked forces were applied to the system using Eq. 10.3, and TPA results are compared to the measured response at one of the validation sensors in Fig. 10.4. As seen, both predictions match the measured response well at higher frequencies, but there is a discrepancy between the measured response and the raw FRF TPA at very low frequencies (below 10 Hz). By rigidifying the FRFs, the TPA results are greatly improved at those low frequencies. 10.2.2 Reciprocal FRFs for Mid-Frequency TPA Predictions A separate test was recently done on a full vehicle for source characterization of the tire noise. Sensors were placed around each of the wheel hubs during FRF and operational tests, again for in-situ blocked force TPA. Impacts were made around each wheel hub and transformed to virtual points, and these virtual point FRFs were used to calculate blocked forces during a constant speed test with Eq. 10.2. For this test series, the response of interest was the sound at the driver’s ear. To calculate this response due to the blocked forces, two sets of FRFs were measured. First, the FRFs YAB 32 between a microphone at the driver’s ear and the impacts at the wheel hubs were measured, as is usually done. Additionally, a low frequency volume source (LFS) was placed at the driver’s ear position, and the accelerometers around the wheel hubs were used to measure YAB 23 . A sample noise transfer function (NTF) from the volume source measurement is shown in Fig. 10.5 where the coherence looks reasonable up to approximately 500 Hz. A few comparisons between the impact test and the volume source NTFs are shown in Fig. 10.6 and Fig. 10.7. Most of the NTFs compare quite well, with similar quality as in Fig. 10.6. Some of the comparisons are not quite as good, as in Fig. 10.7; the worse comparisons are generally for rotational DoF. There are a couple possible reasons for the discrepancies. There may be inaccuracies in the impact measurements that are more pronounced in the rotational DoF. There could also be inaccuracies in the accelerometer responses during the volume source measurement, as the discrepant region is near the region of low coherence in Fig. 10.5. The primary cause of the discrepancies was not further investigated.
RkJQdWJsaXNoZXIy MTMzNzEzMQ==