【正文】 3 supported to form a continuous deck under the track. To find the best elasticity for the rail fasteners, predictions of the bridge noise were made using the Norbert model. Measurements were made on the bridge with the track in its original state to provide parameters for the model. These included rail and sleeper vibration as well as passby noise from service passenger and freight trains at different speeds. For the two tracks, elastic rail fasteners from two suppliers were installed. The measurement after installation showed a clear noise reduction for the frequency range from 80 to 400 Hz of about 10 dB. However the reduction in Aweighted overall noise level is in the range of 2 to 4 dB, as indicated by the model. The results show similar reduction for both systems. 1 Introduction The main reason for the increase in noise as trains cross a bridge is the vibration of the structure. Here the rolling noise radiated by the track and bridge is studied for the twin track bridge over the River Emme at Burgdorf, Switzerland. Two photographs of the 57m long bridge are shown in Fig 1. The track on the bridge takes an unusual form in that steel sleepers have been added between the wooden sleepers on which the track is supported in order to form a continuous deck. The steel sleepers are hung from the rail and are not otherwise supported by the bridge structure. Resilient rail fasteners were installed and the steel sleepers replaced with wooden sleepers in October 2021. The reconstruction was acpanied with noise measure ments in May, October, December 06 and January 07. One track was equipped with resilient fasteners from Pandrol, the other, from Vossloh. The elasticity of the fasten ers in both cases is about 20MN/m. To identify the effect of the hanging steel sleep ers, additional noise measurements were carried out on one track after installing the resilient fasteners but before replacing the steel sleepers. To find the best elasticity for the rail fasteners, predictions of the bridge noise were made using the ISVR software Norbert (Noise of Railway Bridges and Elevated Structures) [1]. Measurements were made on the track in its original state to provide 畢業(yè)設(shè)計(jì)外文文獻(xiàn)及譯文 4 parameters for the model. Fig. 1. Left: The Burgdorf bridge and track. Right: Arrangement of wooden bearing sleepers and suspended sleepers. 2 Modelling ISVR has developed a bridge noise prediction model called Norbert based on the bination of an analytical model of the track and a statistical energy analysis method (SEA) for the bridge [2, 3]. In this method, the bridge structure is described using a single subsystem type。 right side, decay rates of the rail. An estimation of the bined effective roughness (. bined wheel and rail roughness with the contact filter already accounted for) is needed for the predictions. A method to determine the bined roughness from rail vibration measurements under traffic is described in reference [9]. This method uses the spectrum of vibration during a train passby and three correction factors. It has been applied to rail vibration measured by the SBB. The roughness is a function of the brake type of the train. Here, the most important trains to consider are the cast iron block treadbraked freight trains. Two measurements of vertical rail vibration are available from trains travelling at steady speeds of 77km/hr and 72km/hr. Three more are available for lower speeds that vary during the measurements from about 40 to 60km/hr. The estimates of the bined effective roughness from these records are presented in Fig. 3. They are plotted as a function of frequency corresponding to a train speed of 100km/hr. These roughness spectra are plotted in parison with the typical spectra for smooth rail and either cast iron block treadbraked, or discbraked trains. These are the standard roughness spectra used in the Silent Freight and Silent Track EU projects [10]. The roughness assumed for the present calculations is also shown on Fig. 3. 畢業(yè)設(shè)計(jì)外文文獻(xiàn)及譯文 7 Fig. 3. Combined effective roughness spectra derived from measurements pared with those used in the Silent Freight and Silent Track Projects (──, calculated from the two faster freight trains。 ── mean spectrum assumed for current calculations) 3 Noise Measurements The microphone positions were, for both tracks: beneath the bridge。 (a) north track。 ? , predicted after, ?? (a) measured at atgrade track, (b) measured at the bridge with the steel sleepers and resilient baseplates) Fig. 5. Average of measurements of passenger trains。 (b) south track (──, before。 Part