Measurement techniques for the acoustic analysis of synchronousbelts

Abstract

At present, in the automotive field, the noise generated by beltdrives is evaluated by using microphones in the proximity of the belt,crankshafts, idlers and so on. Such a method can be misleading, since it mayeasily include the contributions of other noise sources present during themeasurement. Moreover, a large amount of data is needed in order to testvarious layouts and various running conditions.We present a method for the analysis and prediction of the noise generated bybelt drives which consists of two distinct phases in this paper.For simplicity, a two-pulley belt drive has been considered and the resultshave been validated at the meshing frequency, at which, as has been shown inthe existing literature, the phenomenon of noise generation is mainlyconcentrated.In the first stage of the work, the acoustic power generation of the beltdrive being tested was measured by means of acoustic intensity techniques.Subsequently, an acoustic prediction was performed by using vibration dataobtained with a scanning laser Doppler vibrometer (SLDV) as inputs for aboundary element code. The SLDV was used because of its capability ofmeasuring in-operation data on the running belt, which would not have beenpossible using traditional contact sensors (accelerometers and so on).The results obtained in the two phases were finally compared in order toevaluate the relation between the vibratory behaviour and the total acousticradiation determined experimentally. The experimental and numerical data agreefairly well, adding precious information on the noise generation mechanismsand showing the feasibility of modelling the vibro-acoustic behaviour of beltdrives and the possibility of a totally numerical procedure. In particular theimplementation of an entirely numerical procedure using, for example, datagenerated through the use of codes for the dynamic characterization ofmechanical systems (multi-body mechanical models and so on) seems foreseeable.In the final section of the present work, the uncertainty arising from themeasurement processes of the investigation method presented is also discussed.