An active noise control casing using the multi-channel feedforward control system and the relative path based virtual sensing method

Abstract

Active noise control (ANC) is a noise reduction technique based on acoustic wave superposition. The sound pressure level is reduced in a zone of quiet (ZoQ) by an anti-noise wave transmitted from the control source. The anti-noise wave has the same amplitude and opposed phase of the noise wave. An error microphone is conventionally placed at the target ZoQ to monitor the sound pressure level, forming a closed-loop control in an ANC system. However, due to application constraints or physical limitations, the error microphone sometimes cannot be placed at the target ZoQ. Virtual sensing (VS) methods are developed for such situations. There are two most commonly used VS methods. They are the auxiliary filter based VS (AF-VS) method and the remote microphone based VS (RM-VS) method. The AF-VS method preserves the information regarding the optimal control filter that can achieve the maximum noise reduction at the target ZoQ. The RM-VS method estimates the disturbance signal at the target ZoQ based on remote measurements. In this paper, we propose a new VS method, the relative path based VS (RP-VS) method, which estimates both the disturbance signal and the anti-noise signal at the target ZoQ. A theoretical analysis is provided to demonstrate that under different assumptions of varying acoustic paths, the RP-VS method can behave in the same way as the AF-VS method or the RM-VS method. Simulation results validate this theoretical analysis and demonstrate that improved noise reduction can be achieved by the RP-VS method when the noise frequency varies. Lastly, an ANC casing is built up with the RP-VS method to reduce a varying broadband fan noise. The RP-VS method is validated to be as effective as the AF-VS method and the RM-VS method with the implementation of the ANC casing.