A new feedforward hybrid active noise control system using an adaptive notch filter bank and its application to fan noise

Fan noise in duct systems has three major components, namely, the blade passing frequency (BPF) narrowband component, turbulence noise, and resonance noise. Such fan noise can be suppressed comprehensively by using feedforward hybrid active noise control (FFHANC). In this paper, a new FFHANC system is proposed and applied to reduce each of the three fan-noise components effectively. Two microphones are placed close to the fan, one on the exhaust side and the other on the intake side. An adaptive notch filter bank (ANFB) is applied to the intake microphone measurement to detect and extract the BPF narrowband component, which is then fed to two-weight finite impulse response (FIR) magnitude and phase adjusters (MPAs) to cancel the BPF narrowband component that is contained in the exhaust microphone measurement. The two decoupled noise components, namely, the BPF narrowband component (the ANFB bandpass output) and the broadband noise (the MPA error), are adopted as the input to the FIR-type feedforward broadband ANC sub-controller and the MPA-based narrowband ANC sub-controller, respectively. The proposed system has the following two major advantages. 1) The ANFB swiftly detects and tracks the BPF component, rendering our system more stable and more capable. 2) In addition to the BPF component, the resonance noise whose frequencies are close to those of the BPF component can also be reduced effectively. Extensive simulation and experimental results are provided to demonstrate the effectiveness and improved performance of the proposed system as compared to its counterparts.