Noise source separation of diesel engine by combining binaural sound localization method and blind source separation method. (November 2017)

• Record Type: Journal Article
• Authors: Yao, Jiachi; Xiang, Yang; Qian, Sichong; Li, Shengyang; Wu, Shaowei
• Is Part Of: Mechanical systems and signal processing. Volume 96(2017)
• Page Start: 303
• Page End: 320
• Publication Date: 2017-11
• Subjects: Acoustics -- Diesel engine -- Binaural sound localization -- Blind source separation -- Combustion noise -- Piston slap noise; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2017.04.027
• Languages: English
• ISSNs: 0888-3270

Abstract

Highlights: A combined binaural sound localization method and blind source separation method is proposed to separate and identify the combustion noise and the piston slap noise of a diesel engine. The correlation coefficient and variance are utilized to determine the best dual channel measurement positions. Thus, the combustion characteristics and the mechanical characteristics of the diesel engine can be fully obtained in the measured dual-channel noise signals. In the test, a lead covering method was carried out on a diesel engine to isolate other interference noise from the No. 1–5 cylinders. Only the No. 6 cylinder parts were left bare. Abstract: In order to separate and identify the combustion noise and the piston slap noise of a diesel engine, a noise source separation and identification method that combines a binaural sound localization method and blind source separation method is proposed. During a diesel engine noise and vibration test, because a diesel engine has many complex noise sources, a lead covering method was carried out on a diesel engine to isolate other interference noise from the No. 1–5 cylinders. Only the No. 6 cylinder parts were left bare. Two microphones that simulated the human ears were utilized to measure the radiated noise signals 1 m away from the diesel engine. First, a binaural sound localization method was adopted to separate the noise sources that are in different places. Then, for noise sources that are in the same place, a blind source separation method is utilized to further separate and identify the noise sources. Finally, a coherence function method, continuous wavelet time-frequency analysis method, and prior knowledge of the diesel engine are combined to further identify the separation results. The results show that the proposed method can effectively separate and identify the combustion noise and the piston slap noise of a diesel engine. The frequency of the combustion noise and the piston slap noise are respectively concentrated at 4350 Hz and 1988 Hz. Compared with the blind source separation method, the proposed method has superior separation and identification effects, and the separation results have fewer interference components from other noise.