Diesel Engine Radiation Noise Separation and Identification Based on Dual-channel Source Localization and BSS

The exploration research provides a new view and thinking for the internal combustion engine noise source identification, which proposes to separate engine radiation noise by dual channel localization algorithm that based on time difference and intensity difference cues at the same time. However, due to the inevitable time-frequency aliasing of radiated noise and the problem of multiple excitation sources in the same space location, then the blind source separation method is adopted for further separation. The results show that: dual-channel localization algorithm is equivalent to a kind of "filter", which can eliminate the interference of sources in other location. The joint noise separation scheme successfully separated the combustion excitation component and piston-slap incentives in radiation noise of the engine body side. And the scheme is more superior for separation of complex diesel engine radiated noise compared to the merely BSS method.