Robustness of the extended delayed-X harmonics synthesizer algorithm applied to engine exhaust noise control

Abstract

Exhaust noise of a combustion engine contains several harmonic components of the fundamental frequency depending on the engine speed. Active noise cancellation (ANC) in the exhaust line of a diesel-electric drive faces serious challenges such as high temperature, high sound pressure levels, standing waves phenomena, a time-varying environment and slightly varying disturbance frequencies. In the case of a non-acoustic reference sensor also an offset or a mismatch in the estimation of the disturbance frequencies affects the performance. The extended delayed-X harmonics synthesizer (DXHS) is known from several research projects to be appropriate for control of periodic noise with multiple harmonic components. The algorithm is robust against observation noise and fluctuation of the plant under control. In this paper the extended DXHS is implemented in an ANC-system for a diesel-electric drive and experimentally investigated in laboratory experiments. A modified update rule, a convergence factor regulation and the pre-filtering of the error signal are implemented to improve the DXHS. Performance and robustness are tested in a mockup of an exhaust line in which realistic fluctuations of temperature and gas flow as well as varying frequencies and amplitudes of the exhaust noise are simulated.