Search results

(1 - 1 of 1)

Title

ANALYSIS OF MECHANICAL NOISE GENERATION IN WIND TURBINE DRIVE TRAIN

Creator

Patel, Hirenkumar J.

Date

2012-05-01, 2012-05

Description

The research work presented here is a part of a project, funded by U.S. Department of Energy to study mechanical noise generated by a wind...
Show moreThe research work presented here is a part of a project, funded by U.S. Department of Energy to study mechanical noise generated by a wind turbine drive train. In our study a Viryd 8 kW wind turbine drive train test bed located at the Illinois Institute of Technology was used. Various wind speeds and turbulence levels could be simulated using a computer program that is used to control the test bed. Acoustic measurements were carried out using a single microphone and a microphone array. The microphone array was used to localize noise sources on the drive train. Various beamforming algorithms such as FDBF, DAMAS2, CLSC, DAS and TIDY were used to study the noise sources. Quali cation experiments using synthetic sources showed that \Clean based on spatial coherence" beamforming algorithm localizes noise sources very accurately for narrowband frequency analysis and TIDY was found to work best for broadband analysis. The resolution of the beamform maps improved for higher frequencies of interest (>700 Hz). The continuous variable planetary (CVP) gearbox, which is a proprietary gearbox by Viryd was used in the drive train to optimize the generator rotational speed. An interesting trend was observed in active power generated for the wind speeds greater than 10 m/s, where the power does not increase signi cantly as it is regulated at 6000 Watts. CVP speed ratio, ratio of input rotational speed to output rotational speed of CVP, was also found to be having similar e ect after wind speed reaches a value of 10 m/s. Vibrations of the drive train test bed were studied using accelerometers. It was observed that the test bed was vibrating at a fundamental frequency of 120 Hz, with harmonics of decreasing amplitude at 240 and 360 Hz. Vibrations in all degrees of motion were found to be occurring at similar frequencies. Acoustic beamforming using a microphone array showed that the test bed was a dominant noise source at the same frequencies. Initially the entire test bed was covered by a Plexi-glass casing for safety reasons. It was found that the glass casing a ected the microphone array measurements as the noise produced by the components had no direct path to the array. Almost all the measured noise was refracted through the gaps between the glass casing and the stretcher holding it, that led to spurious microphone array results. As a result of this, the experiments were conducted without the glass casing. It was discovered after the experiments that the glass casing not only a ects the path of sound but the amplitude is also a ected. The components of the drive train namely gearbox, brake, CVP and generator, were found to be emitting sound at various discrete frequencies ranging from 165 to 3885 Hz. They were also found to be emitting broadband noise, where gearbox and generator were found to be most dominant noise sources. We were able to separate each noise source on a complex wind turbine drive train that contributed to the mechanical noise generation from a wind turbine.
M.S. in Mechanical and Aerospace Engineering, May 2012
Show less