The separation and identification of distributed noise sources with local coherence by a subspace equivalent-source method. (1st April 2022)

• Record Type: Journal Article
• Authors: Pan, Xingjian; Jiang, Weikang
• Is Part Of: Mechanical systems and signal processing. Volume 168(2022)
• Publication Date: 2022-04-01
• Subjects: Source identification -- Equivalent-source method -- Distributed sources -- Airfoil noise; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2021.108601
• Languages: English
• ISSNs: 0888-3270

Abstract

Highlights: A subspace equivalent-source method is proposed for the separation and identification of distributed noise sources with local coherence. Sound pressure radiated from each component is represented by coherent structures of the overall sound field, which are calculated by using the Proper Orthogonal Decomposition. The influence of background noise on the inverse problem is reduced by source decomposition and the selection of principle components. Intrinsic coherence among equivalent sources accounts for better performance of the proposed method than beamforming algorithms. Abstract: Identification of distributed and locally coherent aerodynamic noise sources, such as airfoil noise and blade noise, attracts considerable attention these years. When using beamforming techniques, errors in source positions may be caused by the assumption of incoherent sources. A subspace equivalent-source method is proposed to decompose and identify different components of sources. Sound pressure radiated from each component is represented by coherent structures of the overall sound field, which are calculated by using the Proper Orthogonal Decomposition. The equivalent-source model is modified for the influence of the flow field and the shear layer by using a geometry method. Truncated Singular Value Decomposition is applied to avoid singularities generated by the inverse problem and reduce errors of source identification. The proposed algorithm is verified by numerical simulations and a benchmark experiment of airfoil noise in the wind tunnel. Leading edge noise, trailing edge noise and tip noise are successfully separated and located. The distribution of each source is estimated more accurately than Conventional Beamforming and deconvolution methods. Source decomposition and intrinsic coherence among equivalent sources both account for better performance.