| Popis: |
The present work is an exploration of the feasibility and the accuracy of a hybrid FDTD/GA simulation model when applied to broadband acoustic simulations of large-scale environments. The model selected, which has been developed by the Acoustics and Audio Group research team of the University of Edinburgh (UK), combines a wave-based method, a finite-difference time-domain (FDTD) code, with a traditional GA approach, the ray-tracing method. During the investigation, the attention is focussed on the possibilities recently achieved in the wave-based part of the model thanks to the increasing computational power and to new efficient algorithms adopted. The present thesis concerns the method employed to test and assess these new potentials through various application contexts. Moreover, a thorough analysis of the most significant factors affecting the simulation process is carried out with specific assessments of the outcomes. A sample of four large environments is employed in this study, each one of them for specific peculiar reasons: a traditional opera house, a modern theatre, a multipurpose elliptical hall, and a historical university lecture hall. All the simulation process steps and the method adopted to validate the models are reported and described in detail: the 3D modeling phase, the input data, the setting of proper calculation parameters, the time cost, and the acoustic metrics considered time by time in the halls assessed. The thesis allows defining an objective benchmark between the state-of-art of acoustic simulation in medium and large-scale environments and the new advances allowed by recent research in wave-based methods. Finally, the work outlines theoretical and practical remarks on the effectiveness of the hybrid simulation method adopted, enhancing the existing scientific literature with further data and case studies for an increasingly thorough benchmark of computational acoustics techniques. |
