To set up condition indicators for machinery health monitoring it is preferable to run multiple machines with seeded defects and with no defects and measure the resulting vibration signatures; however, this can be a prohibitively expensive undertaking, especially when the machinery has many monitored components. Instead, measurement of the defect-to-monitoring-sensor transfer path dynamics provides a picture of the frequency bands which will amplify the desired vibration signal and those bands which will attenuate it. Since measurement of the true transfer path (from gear tooth or bearing race to the vibration sensor) can also be a costly task, this work seeks to demonstrate that, in cases where the monitored bearing is near the machine surface, a transfer path measurement from the surface near the bearing load zone reasonably approximates the true transfer path, sufficiently for selecting frequency bands of vibration signal amplification and avoidance of bands of signal attenuation.
To demonstrate this in the lab, piezoelectric bearing excitation devices were constructed and placed in a US Army Apache helicopter’s intermediate gearbox and the true-transfer-path measurements were compared to surface-mounted excitation FRFs. As these compared favorably, the surface measurement method was used on all the gearboxes for each of the following rotorcraft: Apache, Kiowa, Chinook, and Blackhawk. Since the Chinook and the Blackhawk are built in a special-ops version which is equipped with a rotor brake, those components were measured on both standard and special-ops versions. In the interest of space, only the complete results from the Blackhawk are shown here; however, the US Army has made all of the data and reports from the project publicly available.