The applicability of simple adaptive algorithms for the active control of random noise in ventilation ducts.

Abstract

The term Active Noise Control (ANC) describes the suppression of an unwanted sound field by the superposition of an antiphase field. In general, an active control system consists of a sensing mechanism to detect the unwanted noise, a processing element to analyse the sound and produce the antiphase signal and a system of secondary sources to radiate the required antisound. The ideal ANC system would have the ability to modify its own response to accommodate any changes in the environment in which it is placed. Such as system is known as anadaptive one. This thesis is concerned with assessing the suitability of several different recursive algorithms for adapting digital controllers to control random noise within ventilation ducts. Algorithms for adapting both Finite Impulse Response (FIR) and Infinite Impulse Response (IIR) digital filters are studied. Computer simulations are presented which directly compare the performance of the algorithms when used in system identification and when used to control adaptive systems in ducts. Experiments were made in ducts of varying length and ranging from anechoic to highly reverberant in nature. Conclusions drawn from the simulations indicate some very significant savings in terms of economy of filter length when implementing adaptive systems within the duct.