An artificial neural network approach to modelling absorbent asphalts acoustic properties

Sound-absorbing asphalts are particularly useful for reducing noise emissions from vehicular traffic. This solution is perfectly suited for urban areas, in fact the use of sound-absorbing asphalt represents a noise control measure with a negligible environmental impact. In the present work, the results of an experimental investigation on sound-absorbing asphalts were reported. First, the characteristics of the sound-absorbing asphalts used were experimentally found. Then, the measurements of the sound absorption coefficient of the asphalt specimens were investigated. In the final part, numerical simulation model with artificial neural networks of the acoustic coefficient were compared with the data obtained from the measurements. The neural network model showed good Pearson correlation coefficient values (0.894) which can be used with good accuracy to predict the sound absorption coefficient.