Takashi TSUTSUI

Research on SEA parameter estimate in a noise prediction of a railway bridge

Yasuyuki MIYAKI

A sound due to oscillation of a bridge extends over the audio frequency area (20Hz - 20 kHz). Therefore when we assume a high frequency area an analysis object, computational complexity increases because a lot of modes are included in an object. In addition, calculation precision deteriorates with it. For such a problem, Statistical Energy Analysis (following SEA method) attracts attention because SEA method is practicable on vibration noise analysis of relative high-frequency.
SEA method expresses oscillation with general quantity of energy and calculates transmission from the equilibrium. Input data of SEA method have to use vibration acceleration of a rail in a railway bridge. Because acceleration level of a rail can do prediction because it is not different with a bridge. Therefore we assumed input power a rail in last year. We evaluated a protection against oscillation effect of rail-pat and slab-pat with an energy transmission factor calculated by an actual survey value. This method calculates one of the dominant parameters of SEA method, an energy transmission factor with measured oscillation acceleration value. However, it is clear that the analysis that does not depend on actual survey value is desirable as possible in a noise prediction.
Therefore this research estimates an SEA parameter with orbit part FEM model, and we suggest FEM-SEA method which we do not use actual survey value for. We make an FEM model of track part and calculate a transmission loss ratio of oscillation between track- floor provided by a frequency response characteristic of the model. This transmission loss ratio expressed a vibrational energy transmission factor of track part and calculated as one of the loss factor of SEA parameter, energy transmission factor ?.
As a result, We was able to obtain improvement of prediction precision with the two-dimensional FEM model that modeled orbit section in a plane distortion element in low frequency area. In addition, calculation precision in high frequency area improved by the three-dimensional model that modeled a rail in a beam element.

Please use "previous botton" to return the previous page