Large-distance outdoor sound propagation with high-frequency noise sources and multiple obstacles/geometry with varying acoustic impedance is common in real-life applications. To resolve the acoustic governing equations directly is often computationally costly, especially in the three-dimensional space. Methods based on geometric acoustics can be more efficient. However, efforts are still being made to improve the efficiency, robustness, and capability for complex configurations of such methods. In this thesis, an efficient method of the rectilinear Gaussian beam tracing is developed, which combines the rectilinear ray tracing with a new efficiency-matched dynamic ray tracing algorithm. A continuous medium stratification treatment is employed to improve the robustness of Gaussian beam t...[ Read more ]

Large-distance outdoor sound propagation with high-frequency noise sources and multiple obstacles/geometry with varying acoustic impedance is common in real-life applications. To resolve the acoustic governing equations directly is often computationally costly, especially in the three-dimensional space. Methods based on geometric acoustics can be more efficient. However, efforts are still being made to improve the efficiency, robustness, and capability for complex configurations of such methods. In this thesis, an efficient method of the rectilinear Gaussian beam tracing is developed, which combines the rectilinear ray tracing with a new efficiency-matched dynamic ray tracing algorithm. A continuous medium stratification treatment is employed to improve the robustness of Gaussian beam tracing computations. Moreover, a ray compression algorithm is proposed to save computation time. The efficiency and capability of the solver are demonstrated by studying several benchmark problems with varying degrees of complexity.

An omnidirectional source model was commonly implemented in Gaussian beam tracing, which neglected the influence of generic directivity patterns in practical acoustic problems. There were efforts to synthesize or reproduce the target directivity pattern over an observation surface using multiple distributed point sources. However, the efficiency and applicability of general applications still call for improvement. To this end, a complex-valued radiation function model is developed in this thesis study to realize the generic source directivity for Gaussian beam tracing computation. An advantage of the method is that only one source is required such that computation cost can be reduced. The verification cases show that this method can give good agreement with benchmarking analytical or wave-based numerical solutions. It can also model a complex source of the spinning sound field to mimic the propeller noise.

The developed Gaussian beam tracing method can efficiently compute sound waves with broadband or multiple frequency contents, which is common in practice. The key measure is to decouple the influences of target source frequencies and ray properties in the proposed formulation. As the high-frequency asymptotic solutions of the wave equation are sought in this method, the ray tracing results are frequency-independent. Therefore, only one computation is needed to determine the ray tracing and ray-centered coordinates, while different frequency contents of the sound wave can be introduced simultaneously and emitted to the far-field. Consequently, the computational cost is significantly reduced for multi-frequency or broadband cases as there is no need for repeated computations, as shown in various validation cases.

This thesis research develops an efficient Gaussian beam tracing method to solve large-distance outdoor sound propagation problems for generic sources in complex environments. The new contributions include the improved numerical implementation of dynamic ray tracing, a numerical acceleration with a ray compression method, and an algorithm for general tonal/broadband source modeling in Gaussian beam tracing. The outcome of this research can facilitate noise simulation of flying vehicles in realistic urban environments, especially for low-altitude flight cases where detailed urban geometries should be considered.