Eventos Anais de eventos
COBEM 2023
27th International Congress of Mechanical Engineering
INFLUENCE OF SECONDARY PATH MODEL ON NOISE REDUCTION PERFORMANCE OF ACTIVE NOISE CONTROL SYSTEM IN DUCT
Submission Author:
Gabriela Cristina Candido da Silva , DF
Co-Authors:
Gabriela Cristina Candido da Silva, Renato Lopes, Maria Alzira de Araújo Nunes, Bruno Giuliani Gomes, André Murilo
Presenter: Bruno Giuliani Gomes
doi://10.26678/ABCM.COBEM2023.COB2023-0148
Abstract
Due to the increase in acoustic noise pollution, mainly in great cities, noise control has been the subject of much scientific research. In a general way, there are two kinds of noise control: passive and active. Active noise control (ANC) is an effective way to attenuate noise that is difficult and expensive to control using passive means. Developed in 1936, the ANC generates a secondary noise with equal amplitude and the opposite phase of the primary noise to cancel out undesired noise destructively. This technique control is effective in reducing low-frequency noise, and it is extensively adopted in industrial applications. A useful application of the ANC is attenuating the noise generated by rotating machines such as fans and exhausters, which are often periodic and contain multiple tones. Especially since the eighties until nowadays, much research has suggested newer (or modified) control algorithms and techniques improve ANC systems. The well-known and most used control algorithm that attempted to cancel out undesired noise is the Filter-x Least-Mean Square (FXLMS) algorithm. Many scientific works have shown its efficiency and viability. This method requires consistent secondary path modeling to ensure system convergence. It is known that a higher model order increases the model's accuracy. However, it increases computational complexity. This paper aims to investigate the secondary path model's influence on ANC's noise reduction performance using the FXLMS algorithm in an acoustic duct. White noise and pure tone sine waves were used as a noise source. A system identification method performs offline secondary path estimation for different model orders. The experimental results suggest that it is possible to reduce the order of the secondary path model, achieve good modeling accuracy and fast convergence, and maintain noise attenuation performance.
Keywords
ANC experimental bench, duct, State Space
