Reduced basis method with parameterized boundary conditions for room acoustics

Room acoustic simulations can be performed by means of numerical methods, which typically solve the wave equation in an enclosure through discretization techniques. These methods provide high-fidelity solvers that include all the wave phenomena but are computationally costly. This paper investigates the potential of a reduced basis method for simulating room acoustics when the boundary properties are changed iteratively, e.g., for the investigation of effects of design changes in rooms where different boundary properties are simulated. Such studies increase the computational cost when using traditional high-fidelity solvers. The presented framework allows reducing the computational burden by applying reduced basis methods and solving the problem in a low-dimensional subspace where the absorption properties of complex boundary conditions are parameterized, e.g., the thickness of a porous material. The potential of the proposed framework is analyzed in terms of computational efficiency, accuracy and storage requirements. A computational speedup of two orders of magnitude is demonstated for a 2D case with an upper frequency of 2 kHz, and impressive three orders of magnitude for a 3D case with an upper frequency of 1 kHz showing a potential of having an increase in terms of speedup when increasing the size of the domain. The storage requirement for a 2D case with an upper frequency of 8 kHz is 10.8GB while for a 3D case with an upper frequency of 1 kHz is 8 GB.