Active Noise Control of Axial-Fan Using Singular Spectrum Analysis

Active noise control (ANC) of sound radiated by axial-fan is a challenging work, since microphones may be affected by wind. The induced wind noise will decrease the coherence between primary noise and reference input and result in degradation of noise reduction performance of feedfoward active noise control system. Singular spectrum analysis (SSA) is utilized to extract axial-fan sound from microphone signal contaminated with wind noise in this paper. The characteristic of axial-fan sound is exploited to estimate the phase and frequency of each harmonic component to reconstruct the artificial reference input. Simulation result shows that the artificial reference signal is effective in reducing axial-fan sound.