The positioning of the error sensors for a multichannel control system in a circular duct is currently an important subject of research(Baumann 1998). Indeed, the sensors positioning plays a decisive role on the noise reduction obtained. However, the control sources positioning is still a problem, and it is almost absent of the literature. Even so, the use of several speakers for a multichannel control system presents an important problem regarding the necessary acoustic power needed for each speaker. In fact, the necessary power for the control sources can occasionally become greater than the one generated by the primary source, depending on the transfer function combination between different speakers and error microphones. The required power for the control sources is, sometimes, an important limitation for industrial applications, mostly when the primary acoustic field is very high. Thus it seems that the optimization of the control sources positioning is an important factor, in order to minimize the power used.
One of the objectives of this paper is to quantify the importance of the limited power for various parameters: frequency, ducts length, control sources positioning and error sensors plane. The first section will present an experimental analysis of the distribution pressure field in a semi-open duct. The effect of the source position and of the duct length on the interference pattern will be analyzed from an active control point of view. The second section will study an analysis of the acoustic power used by actuators in a multichannel system case. The effects of closed-end reflections, actuators position and error sensors position on the power used for the control sources will be analyzed. With this in mind, a positioning method has been developed and evaluated by experimental tests to minimize the power used by control actuators.
2. Longitudinal distribution of acoustic pressure field
In the case of the inside duct sound propagation, it is common knowledge that the stationary waves are established to form an interference pattern, following the reflection coefficients and radiation impedance at the duct end(Bouchard 1998). Understanding this phenomenon of propagation is relatively simple for low frequencies when the plane mode (0,0) is the only one to be propagated in the duct. However, when several propagation modes occur, the inside duct pressure field distribution becomes more complex because of the modes combination. In those cases, the pressure field is not uniform, not only following the duct axis, but also in each duct section. Before analyzing the power actuators problem for the multichannel active control, it appears important to review the inside duct pressure field longitudinal distribution for the higher order modes.
2.1. Experimental assembly
To study the inside duct pressure field longitudinal distribution, an experimental study has been realized with the assembly shown in Fig. 1. The pressure field is generated in a PVC duct of 30 cm diameter and 2.05 m in length, with a closed and open end. A fixed speaker on the duct wall is used to generate a pressure field in the ...