Pilot study on numerical prediction of sound reduction index of double window system: Comparison of finite element prediction method with measurement

The development of windows with high sound insulation performance is essential for preventing the infiltration of traffic noise and the leakage of room noise. A numerical prediction is an effective means of reducing sound insulation testing and development costs to develop a quietness window. As a numerical prediction method for the sound reduction index, the finite element method (FEM) is useful in dealing with structure-acoustic problems. This study was conducted as a pilot study toward developing an accurate numerical model to predict the sound reduction index of a double window. We discussed the accuracy of an FEM model for predicting the diffuse incidence sound reduction index of double windows through a comparison with measured values for a simplified realistically scaled double window. The FE results were compared with measured ones for eight cases with and without a frame absorber. Results showed that the best match to measured values is obtained when using a frame absorber in all the perimeters inside the air cavity. Also, a better agreement is obtained at frequencies of 160-2,000 Hz in other cases. However, a marked discrepancy is found at frequencies above 2,000 Hz and below 160 Hz. Possible reasons for the discrepancies are also discussed.