Prediction of interior noise in rolling stock by statistical energy analysis

Abstract

A model of statistical energy analysis (SEA) of railway carbodies was developed to predict interior noise in rolling stock. Since the carbody shell is made of double skin structures of extruded aluminum and their coupling loss factors (CLFs) cannot be easily estimated from theoretical equations, sectional finite element method (FEM) models of the carbody shell were used to calculate CLFs between each SEA subsystem. The whole carbody shell was modeled as SEA subsystems using these CLFs, and interior panels and the passenger rooms were also added as SEA subsystems to the carbody shell model. The input power of an operational device was then calculated from measured acceleration data to simulate interior noise in the passenger room due to the operational device located under the carbody. Calculated input power was applied to the SEA model, and it was found that the simulated vibration level in the carbody, as well as interior noise in the passenger rooms, agreed well with the experimental results.