Nonlinear active noise control system based on correlated EMD and Chebyshev filter

Abstract

• A nonlinear ANC system using signal decomposition and Chebyshev filter is proposed. • The non-stationary primary noise is decomposed by the improved correlated EMD. • The Chebyshev nonlinear filter is investigated to expand the decomposed IMFs. • Simulation results indicate that proposed method outperforms conventional methods. Investigations into active noise control (ANC) have been widely conducted with the aim of effective control of low-frequency noise. This paper proposes a novel nonlinear ANC system to control non-stationary noise produced by rotating machinery under nonlinear primary path. A real-time correlated empirical mode decomposition (CEMD) is first presented to decompose non-stationary noise into intrinsic mode functions (IMFs), some of which are chosen by the correlation analysis between IMFs and primary noise. Subsequently, the second-order Chebyshev nonlinear filter is applied to expand selected IMFs that are controlled individually by the filtered-x LMS. The convergence of proposed nonlinear ANC system is also investigated. Simulation results demonstrate that proposed method outperforms the filtered-x LMS, filtered-s LMS and Volterra filtered-x LMS algorithms with respect to noise reduction and convergence rate.