Prediction of sound absorption property of metal rubber using general regression neural network

Metal rubber (MR) is an excellent sound absorption material which can be utilized in extremely harsh environments. Traditional experimental studies are incomplete for MR with random structural parameters. In this article, the general regression neural network (GRNN) method is developed to comprehensively predict the sound absorption behaviors of MR with random structural parameters. Sixty samples are utilized and divided into training and test set. Training set contains 50 samples to establish the GRNN model. Input training parameters include the porosity, wire diameter and thickness, while the target dates consist of sound absorption coefficients at six central frequencies as well as their average values. The remaining 10 samples constitute the testing set; sound absorption coefficients can be obtained by inputting their structure parameters. Results indicate that the proposed approach is reliable to design and predict the sound absorption properties of MR in engineering field.