Validation of Bayesian design for broadband microslit panel absorbers using causal inference

Area of Science:

• Acoustics and Materials Science
• Computational and Experimental Physics

Background:

• Single-layer microslit panels (MSPs) offer limited sound absorption bandwidth.
• Stacking MSPs enhances sound absorption across wider frequency ranges.
• Bayesian inference is used for designing acoustic absorbers.

Purpose of the Study:

• To experimentally validate multilayer microslit panels (MSPs) designed for simultaneous high sound absorption and wide bandwidth.
• To address challenges in experimental validation arising from deviations from the initial Bayesian design.
• To apply causal analysis for reasoning about and correcting fabrication inaccuracies.

Main Methods:

• Bayesian inference for initial MSP design.
• Experimental validation of multilayer MSPs.
• Causal model-based inference for analyzing fabrication inaccuracies.

Main Results:

• Demonstrated experimental validation of multilayer MSPs.
• Identified and analyzed deviations between designed and experimental performance.
• Utilized causal inference to understand and guide corrections for fabrication inaccuracies.

Conclusions:

• Multilayer MSPs show promise for broadband, high sound absorption.
• Causal analysis is effective in addressing experimental validation challenges caused by fabrication errors.
• Iterative design and validation process, guided by causal reasoning, can optimize MSP performance.