Sound quality of diesel engines

Aaron Lee Hastings, Purdue University

Abstract

Variations in timing and amplitude of combustion events in diesel engines and the presence of tonal components affect the perceived attributes of the sound generated by these engines. People use the levels of these attributes when making overall judgments of the quality of the sound and these attributes influence how annoying the sound is perceived to be. The objective of the research was to be able to relate annoyance, combustion variations, and sound characteristics. Sounds from many diesel engines were measured and a sound simulation program was developed whereby timing and amplitude variation, combustion pressure profiles, transfer paths, the noise floor, and the presence of additional tonal components could be controlled. Aures Tonality was examined and refined to improve its predictions of stationary narrow-band noise. Aures Tonality was combined with statistical attributes of the Zwicker and Fastl sound quality metrics: Loudness, Sharpness, Fluctuation Strength, and Roughness, and used as inputs to an annoyance model. The influences of combustion variations and tonal presence on sound attributes and annoyance were also examined. It was found that the least annoying timing variations were at a level of 16 percent of the engine firing period and the least annoying amplitude variations were at a level of 20 percent of the nominal combustion amplitude. As the variation level increased or decreased from these levels, annoyance increased. For periodic variations, as the frequency of variation increased from 2 to 8 Hz, the annoyance decreased. A linear model which combined the effects of several sound quality metrics was developed from which it could be seen that annoyance increased as the frequency of tones, Sharpness, Loudness, Tonality, and Fluctuation Strength increased. The psychoacoustic model described by Fastl was also examined and a new term based on the modified Tonality metric was included to model the observed increase in annoyance with increasing Tonality. Recommended future work includes modifying the Tonality model to take into account the asymmetry of auditory filters and validating the modified Tonality and annoyance models to demonstrate their broader applicability beyond explaining the responses of the subjects taking part in the tests described in this manuscript.

Degree

Ph.D.

Advisors

Bolton, Purdue University.

Subject Area

Mechanical engineering|Acoustics|Automotive materials

COinS