Global Active Noise Control of a One-Dimensional Acoustic Duct Using a Feedback Controller

Abstract

Active noise control of acoustic enclosures is a classical engineering problem. The active noise control of a one-dimensional hard-walled duct with a partially dissipative boundary condition is addressed in this paper. Previous techniques have attacked this problem by developing adaptive filters designed to cancel acoustic noise at a single measurement location. The work presented here applies modern, state space, control theory to globally reduce noise levels in a one-dimensional acoustic enclosure rather than at a single location. This global control requires only the addition of a single response measurement microphone and control speaker to the open-loop system. Pressure measurements are taken at the microphone location and passed to an observer, which generates estimates of the system states. Using these state estimates, a pole placement feedback control algorithm is used to lower the noise level. Pole placement reduces noise levels globally by increasing the decay rates of all the controlled acoustic states. Experimental results presented here demonstrate reduction in the noise level at all points in the duct by 58 percent when the system is excited by random amplitude pressure input. System ModelAn accurate discretized system model is required for the control system's state observer. The linear second order wave 488 / Vol. 115, SEPTEMBER 1993Transactions of the ASME Copyright © 1993 by ASME Downloaded From: http://dynamicsystems.asmedigitalcollection.asme.org/ on 05/24/2015 Terms of Use: http://asme.org/terms

Theoretical modeling issue in active noise control for a one-dimensional acoustic duct system

“…The developed model serves a solid and sufficient basis for ANC design using advanced control techniques, such as the pole-placement design [8], the lead-lag compensators [18], the mode predictive control [19]. The global attenuation [8], the selection of best actuatoring/measureing position [15] can also be efficiently analyzed by simply define the performance point x p or x s and x m as (sliding) variables in the model [16].…”

“…The global attenuation [8], the selection of best actuatoring/measureing position [15] can also be efficiently analyzed by simply define the performance point x p or x s and x m as (sliding) variables in the model [16]. The developed model is also suitable for applying advanced filtering techniques, such as Kalman filters, for ANC design.…”

“…However, some simplification or approximation of this parameter could make the developed model (18) standard. For instance,in case that K is assumed constant, then system (18) can be simplified to a two-input one-output LTI system [8], [15], [17].…”

“…• The model developed in this way consists of the understandings of the physical system to the maximal extent without losing its flexibility, so that it can be used for a class of systems instead of a specific setup; • The state space model has a good orientation for applying advanced filtering techniques (e.g., Kalman filtering techniques, virtual sensor concepts [1], and advanced control techniques as well [12], [8], [19]; Furthermore, • The models used by most of existing work are obtained mainly using the black-box method.…”

“…This model is a two-input one-output LTI state space description. Comparing with the existing models [7], [8], [14], this model consider the coupling acoustic dynamics coming from the front duct as well as the rear enclosure. The usage of rear enclosure is quite important to avoid the "acoustic shortcircuit" from the practical point of view [17].…”