Further Development of the ROSI Beamforming Method for the Investigation of Turbomachinery Investigated at an Angle

Zhang, Haitian and Kocsis, Bálint and Horváth, Csaba (2020) Further Development of the ROSI Beamforming Method for the Investigation of Turbomachinery Investigated at an Angle. In: 8th Berlin Beamforming Conference (BeBeC), 2020.03.02-2020.03.03, Berlin-Adlershof.

Preview

Text BeBeC-2020-D13.pdf Download (1MB) | Preview

Abstract

The ROtating Source Identifier (ROSI) beamforming algorithm is a method designed for localizing rotating noise sources in a uniform flow utilizing out-of-flow measurements of the acoustic pressure field. It has been developed for sources moving along a circular path, such as turbomachinery blades. The original ROSI method takes the average of the measured acoustic signal over a long time segment in order to reconstruct the noise sources. By doing so, it does not take into consideration certain features of the noise sources, such as differences between the trailing edge and the leading edge noise sources of the given blades. A further development of the method is presented herein, which separates the pressure signal of one revolution of a given noise source into multiple segments. This method will be referred to as the segmented ROSI method. Using this method, the beamforming maps can provide a better understanding as to the differences between the noise sources as a function of angular position. The goal of this further development is to improve the capability of the method in identifying the position-dependent differences between the noise sources that are rotating around an axis. The new segmented ROSI method is presented through a series of example cases, comparing it to the original ROSI algorithm. Though the publication only presents the theory behind the method under development and a few basic validation cases, the results are promising regarding the applicability of the method to other more complex tasks upon fine tuning the method and the associated code.

Actions (login required)

Edit Item