Abstract:

A kind of environment-friendly sound-absorbing material was fabricated by means of concrete molding , with expended cement , light ceramisite , fly ash and polypropylene fiber as the raw materials. The effects of the fiber length and dosage on the sound-absorbing property were experimentally investigated , finding that the soundabsorbing property , especially that at a high frequency , improves significantly with the doping of polypropylene
fiber. Moreover , a prediction system based on back-propagation (BP) neural network was proposed by means of Matlab , which was further trained with the data of fiber length , fiber dosage , test frequency and practical absorption coefficient. Thus , an input-output nonlinear mapping relationship was presented for the fabricated material. The predicted results of the test data indicate that the proposed system is of high accuracy.

Key words: sound-absorbing property, polypropylene fiber, mechanical strength, neural network, simulation