Active Control of Noise in Ventilation Systems : Analysis and Experiments

Abstract: In many kinds of buildings, the ventilation system constitute a well known source of broadband noise. Traditionally, duct born noise is attenuated using passive resistive silencers which produce a high level of attenuation over a broad frequency range. However, they tend to become large and bulky if designed for low frequency attenuation. The active noise control (ANC) technique is known for its ability to produce high levels of attenuation in the low frequency range even with a relatively moderate sized ANC system. Accordingly, a combination of active- and passive techniques, i.e. the construction of a hybrid active/passive silencer, provides a duct silencer solution of manageable size which also covers the low frequency range. However, adequate levels of attenuation are not likely to be obtained if the installation of the ANC system is not designed to account for the physical factors that may degrade its performance. This thesis focuses on applying ANC in ventilation systems, with particular emphasis on analyzing the limiting effect of some of these physical factors on the controller, and installation design for the purpose of reducing the influence of them. The degrading factors of particular interest include: flow induced noise in the microphone signals, acoustic feedback between the control loudspeaker and reference microphone, and standing waves and higher order acoustic modes inside the ducts. With respect to installation design, focus is also placed upon industry requirements for the ANC system. This has led to a module based approach, in which the microphones and the loudspeaker are installed in separate modules based on standard duct parts. This thesis is comprised of six parts. The first and third parts analyze the influence of flow-induced noise on the adaptive digital controller theoretically, through simulations and experiments. The second part describes investigations of several microphone installations intended to reduce flow induced noise. Further, results of measurements conducted in an acoustic laboratory according to an ISO-standard are presented. The attenuation produced by the ANC system was approximately 15-25 dB between 50-315 Hz, even for airflow speeds up to 20 m/s. The fourth part focuses on the possibility of using a passive silencer in combination with ANC, to reduce acoustic feedback and standing waves, while the fifth part investigates the possibility of using a passive silencer to reduce standing waves in the duct when the ANC system is not installed near the duct outlet. In ducts of larger dimensions, higher order acoustic modes may be in the frequency range adequate for ANC. The final part presents initial investigations concerning the feasibility of dividing a duct of large dimension into two more narrow ducts to remove higher order acoustic modes in the ANC frequency range, and the feasibility of applying single-channel ANC in each duct.