Vehicle positioning is a vital issue for new vehicle design and engineering. It is critical to effectively and efficiently represent the idle noise so that proper development activities can be directed. The A-weighted sound level is widely used in characterising engine idle noise. However, the use of A-weighted sound levels to describe sounds or noises can be quite misleading and may lead to confusion, since it can be shown that two sounds or noises of totally different spectra can have the same value of sound level. Multivariate statistical methods such as principal component analysis (PCA) and correspondence analysis (CA) can be used in order to represent the multidimensional response of idle noise in a simpler and meaningful way. Also, another important benefit is that PCA and results do not depend on different weighting schemes, such as A, B and C weights, since PCA is based on a covariance matrix of the sound pressure level (SPL) and adding constant values does not change the covariance structure. PCA is applied to idle noise in order to understand the difference among vehicles. The idle noise in order to understand the difference among vehicles. The idle noise can be attributed to 4 components representing the SPL on the entire frequency bands. The first component can be interpreted as the overall SPL, overall dB(A), on the entire frequency band and the second component as the SPL at lower frequencies between 25 Hz and 50 Hz. The first component has a very high correlation with traditional measurements, overall dB(A). The PCA results show that there are differences among vehicles with a different number of cylinders, especially at low frequencies. This analysis provides a simplified comparison of vehicles regarding idle noise. The position of vehicles projected on PC space using the PCA can be used in order to evaluate the relative position of new vehicles to be introduced in a market. Also, CA is applied in order to identify vehicles with different profiles of SPL, since the shapes of SPL are also important factor to consider. CA shows that the shape difference occurs mostly at low frequencies and that there is not much difference beyond 60 Hz. These two methods have advantages over traditional measures, such as dB(A) for idle noise, since the SPL on the entire spectrum and the shape of the SPL can be analysed and represented in a simpler way. In conclusion, PCA is useful to represent vehicles with a large variation in SPL, and CA is helpful to identify vehicles with different profiles of SPL. By applying both PCA and CA, we have a better understanding of idle noise.
Multivariate statistical analyses of idle noise and vehicle positioning
Mehrdimensionale statistische Analyse von Leerlaufgeräusch und Fahrzeugrangliste
International Journal of Vehicle Noise and Vibration ; 2 , 2 ; 156-175
2006
20 Seiten, 23 Bilder, 3 Tabellen, 11 Quellen
Article (Journal)
English
Anwendung im Fahrzeugbau , Anwendung im Motorenbau , Leerlauf , Schalldruck , Eigenfunktion , statistische Analyse , Motorgeräusch , Kraftfahrzeugmotor , Branchenuntersuchung , Engineering , Ton (Akustik) , Lärm , Lautstärkemessung , Bewertungsverfahren , Schwingungsfrequenz , akustische Eigenschaft , Schall , Schallwelle , Rangordnung , technische Akustik , akustische Messung
Multivariate statistical analyses of idle noise and vehicle positioning
Automotive engineering | 2006
|Multivariate statistical analyses of sediment geochemistry
Tema Archive | 1990
|Method for maintaining Idle Noise Improvability and Vehicle thereof
European Patent Office | 2021