Comparison of analysis and experiment for gearbox noise

Low-contact-ratio spur gears were tested in the NASA gear-noise rig to study the noise radiated from the top of the gearbox. Experimental results were compared with a NASA acoustics code to validate the code for predicting transmission noise. The analytical code is based on the boundary element method (BEM) which models the gearbox top as a plate in an infinite baffle. Narrow-band vibration spectra measured at 63 nodes on the gearbox top were used to produce input data for the BEM model. The BEM code predicted the total sound power based on this measured vibration. The measured sound power was obtained from an acoustic intensity scan taken near the surface of the gearbox at the same 63 nodes used for vibration measurements. Analytical and experimental results were compared at four different speeds for sound power at each of the narrow-band frequencies over the range of 400 to 3200 Hz. Results are also compared for the sound power level at meshing frequency plus three sideband pairs and at selected gearbox resonant frequencies. The difference between predicted and measured sound power is typically less than 3 dB with the predicted value generally less than the measured value.