Experimental study of active control of high order acoustic modes in ducts

Tonal noise radiated from an exhaust stack of a spray dryer system used to make powdered milk has been the source of unwanted community noise. The tone is characterised by a frequency above the first mode cut-on frequency of the exhaust duct and it is not steady in either amplitude or frequency. Passive silencing of the exhaust is a very expensive and inconvenient option and for this reason the feasibility of using active noise control was investigated. Zander and Hansen 1 developed a theoretical model for analysing the effect of active noise control on higher order rectangular duct modes, and used it to evaluate the effects of source size, location and strength. It was shown that the total acoustic power reduction was dependent on the relative locations of the control sources. More recently this work was applied to the practical example of a cylindrical spray dryer exhaust stack 1,2 to determine the most suitable location and number of control sources and error sensors. Prior to installation in the exhaust stack a half scale model was constructed to test the theory. Here, the effect of the error sensor locations on the active control on higher order acoustic modes propagating in the half scale circular duct is investigated experimentally using real time control with a multi-channel feedforward controller. A tacho meter signal at the fan blade pass frequency (BPF) was fed into the controller as the reference signal for the feedforward control. Two configurations of error sensors on the duct wall were tested to evaluate how the error sensor locations affect the control performance, which was evaluated in the far field of the duct outlet.