Artificial neural network-based models for predicting the sound absorption coefficient of electrospun poly(vinyl pyrrolidone)/silica composite. (1st December 2020)

• Record Type: Journal Article
• Authors: Ciaburro, Giuseppe; Iannace, Gino; Passaro, Jessica; Bifulco, Aurelio; Marano, Aniello Daniele; Guida, Michele; Marulo, Francesco; Branda, Francesco
• Is Part Of: Applied acoustics. Volume 169(2020)
• Publication Date: 2020-12-01
• Subjects: Sound absorption coefficient -- Deep neural network -- Acoustic measurements -- Electrospinning -- PVP/silica composite; Acoustical engineering -- Periodicals; Periodicals; 620.2
• DOI: 10.1016/j.apacoust.2020.107472
• Languages: English
• ISSNs: 0003-682X

Abstract

Highlights: Polymeric sound absorbers can be produced through electrospinning, materials with a fiber diameter from few nanometers to several micrometers. A numerical simulation model of the acoustic behavior of poly vinyl pyrrolidone/silica composites were developed. The results of the measurements of the sound absorption coefficient were analyzed. The results of the numerical modeling of the acoustic coefficient were reported. The neural network-based model showed high Pearson correlation coefficient values (0.942), indicating many correct predictions. Abstract: Polymeric sound absorbers can be produced through electrospinning, a process which allows to fabricate high specific surface materials with a fiber diameter from few nanometers to several micrometers. In this study, a numerical simulation model of the acoustic behavior of poly vinyl pyrrolidone/silica composites were developed. First, the characteristics of the poly vinyl pyrrolidone/silica composites were examined, and the manufacturing of the material were described. Subsequently, the results of the measurements of the sound absorption coefficient were analyzed. Finally, the results of the numerical modeling of the acoustic coefficient were reported. The neural network-based model showed high Pearson correlation coefficient values (0.942), indicating many correct predictions. Taking into account the bell shaped acoustic response of the studied blankets as a function of frequency, the possibility to foresee the needed mass with the neural network-based model will be of great value for the applications where high acoustic absorption is required in specific limited frequency ranges.