Acoustic beamforming in angular domain applied for rotating machines

Standard acoustic beamforming methods are common for noise source identification in far field. Steady noise sources are identified quite accurately within the available methods precisions. Additionally, methods for transient sources have been published. However, the time dependence does not provide a clear picture in all cases, especially for rotating machines. The published methods use either de-rotating or order-tracking based algorithms. Our algorithm employs a measurement with rotary encoders placed on selected rotating machine parts. The angular position of mechanisms is thus known more precisely and the noise sources are identified more accurately in the angular domain. Furthermore, we can obtain the results in a plane that is not necessarily parallel with the microphone array plane (which is usual in the commercial systems with planar arrays). In this contribution, we compare the results of a standard beamforming in a plane parallel with the array with the results in a tilted plane.