SEA-FEM hybrid analysis for predicting Inter-floor impact noise

Abstract

A hybrid method for predicting inter-floor impact noise in an apartment house is developed. Statistical Energy Analysis (SEA) is useful for predicting mid- to high-frequency noise by utilizing simple equivalent models of structures that have complex geometry and material properties. For low frequencies with concomitant low modal density, however, the error can be significant. Finite Element Method (FEM) is useful for predicting low frequency noise but entails complex modelling process and can be computation intensive. The inter-floor impact noise contains a significant low frequency component arising from floor vibration engendered by a person walking on the floor. The force of impact at 30–300Hz frequency range excites the first and second bending modes of the floor and causes sound to radiate, and for the component of the sound less than 100Hz, the SEA prediction error can be quite significant. In the present study, a modular hybrid method in which the floor structure is modelled using FEM and the vibration energy transmission through walls is modelled using SEA is proposed. The low frequency vibration of the floor due to impact is analyzed by modelling the floor asa clamped uniform plate while the propagation of vibratory energy to adjacent walls is analyzed by determining CLF and DLF values. The proposed SEA-FEM hybrid analysis method is capable of providing accurate estimation of the impact noise for low-, medium- and high-frequencies. Compared with a FEM−based method, the proposed method is found to yield results that are similar while reducing computation time by 70%.