Prediction of structure-borne sound in buildings using the substructure synthesis method

Accurate predictions of structure-borne sound in buildings using numerical simulation methods, such as the finite element method, require detailed modeling, which consumes much memory and computational time. The substructure synthesis method has been progressively developed since 1970 to obtain dynamic characteristics of large-scale structures, such as machines and vehicles. It can obtain the dynamic characteristics of a structure as a result of the synthesis of vibration modes in each substructure, which can be independently computed. However, it is still time consuming when applied to analyze a building. This study presents two ideas to reduce the computational time of the substructure synthesis method. The first is to reuse the calculation result. In buildings, once a wall, column, or beam has been analyzed, the result can be applied to the other walls, columns, and beams having the same material and size. The second idea is to multiplex the calculation. A group of substructures can be analyzed and synthesized into a new substructure. This procedure is repeated until the substructure becomes the entire structure. These two ideas have been applied to compute a frame structure and have shown to be effective in reducing the computing time without decrease in calculation accuracy.