Validating impeller geometry optimization for sound quality based on psychoacoustics metrics

Abstract

The noise generated by vacuum cleaners is one of the dominant noise sources among household appliances with its suction unit as the key element of noise generation. The dominant noise source in the suction unit is its impeller, and that is why there were numerous efforts made to reduce its noise. The years of its optimization have consequently led to a stage, where the possibilities for further noise level reduction are limited. Because human perception of noise is not only influenced by sound pressure levels, it is possible to alter the sound of the impeller to be less annoying to the user. A new centrifugal impeller design was optimized for sound quality, resulting in triangular-shaped flow channels. The optimization process of impeller geometry was based on the sound quality calculated from measured psychoacoustic metrics using two noise annoyance models, Guoqing’s and Lipar’s. The validation of the result was performed by listening tests. Forty individuals participated in the listening tests of the suction units tested and rated the sounds using two sets of semantic differentials. The first set was used to evaluate the sound quality of the impeller, while the second set was used for the validation of these listening tests by evaluating psychoacoustic metrics. Validation results confirm that Lipar’s model of noise annoyance is suitable for evaluating centrifugal blower impeller noise. The subjective evaluation of psychoacoustic metrics, via listening tests of impeller noise, suggests that fluctuation strength and tonality are important features for the sound quality of impeller noise.