Modeling acoustic metamaterials based on reused buttons using data fitting with neural network.

Area of Science:

• Acoustics and Materials Science
• Focuses on the development and acoustic characterization of novel metamaterials for sound absorption applications.

Background:

• Metamaterials, engineered with periodic structures, offer unique wave interaction properties.
• Their application has expanded from electromagnetic to acoustic waves, showing promise for building soundproofing.

Purpose of the Study:

• To develop and evaluate a novel polyvinyl chloride (PVC) membrane-based metamaterial for acoustic sound absorption.
• To utilize artificial neural networks (ANNs) for accurate prediction and sensitivity analysis of the metamaterial's acoustic performance.

Main Methods:

• Fabrication of a new metamaterial using a PVC membrane with radially arranged buttons of varying weights.
• Acoustic absorption coefficient measurements using an impedance tube across three different cavity configurations.
• Development and training of an ANN model using measured data for prediction and sensitivity analysis.

Main Results:

• The ANN model demonstrated excellent generalization capabilities, providing highly accurate estimates of the acoustic absorption coefficient.
• Sensitivity analysis revealed the contribution of input variables to the metamaterial's acoustic performance.

Conclusions:

• The developed PVC membrane metamaterial shows potential for effective sound absorption in building applications.
• The trained ANN model serves as a reliable tool for predicting acoustic performance and guiding future metamaterial design.