Reduced order modelling using parameterized non-uniform boundary conditions in room acoustic simulations.

Area of Science:

• Acoustics
• Computational physics
• Building science

Background:

• Optimizing building acoustics requires rapid simulations for evaluating design variables and boundary conditions.
• Realistic room acoustic simulations often involve complex, inhomogeneous surface properties.

Purpose of the Study:

• To investigate the application of the reduced basis method (RBM) for efficient room acoustic simulations with inhomogeneous boundary conditions.
• To assess the RBM's performance across various room geometries, source positions, and surface acoustic parameterizations.

Main Methods:

• Implementing the reduced basis method (RBM) for acoustic simulations.
• Parameterizing inhomogeneous surface properties, including material thickness and cavity depth.
• Comparing RBM simulation speeds against a high-fidelity numerical solver.

Main Results:

• Achieved speedups exceeding two orders of magnitude for 3D room acoustic simulations.
• Demonstrated RBM's effectiveness with diverse geometries and frequency-dependent boundary conditions.
• Verified simulation accuracy where reverberation time changes minimally.

Conclusions:

• The reduced basis method (RBM) significantly accelerates complex room acoustic simulations.
• RBM enables efficient evaluation of inhomogeneous surface properties for building acoustic optimization.
• This approach is valuable for iterative design processes in architectural acoustics.