【文章內(nèi)容簡介】 beneath the bridge。 to the side, above the rail head and 25m to the side, 2m above the rail head. These were all in a plane 6m from one end of the bridge. An additional measurement was made in the position adjacent to the track at grade. The details of the measurements can be found in a report [11] for Swiss Railways. Both freight and passenger trains were measured. Only freight trains fully equipped with cast iron brakes are included. All passenger trains were either equipped with positeblock or disc brakes. All passenger trains stopped at the nearby station of Burgdorf resulting in a large range of velocities and some changes of velocity (up to20% in some cases) during measurement. 畢業(yè)設(shè)計(jì)外文文獻(xiàn)及譯文 8 4 Results and Discussion Fig. 4 presents the measurements and predictions for the freight trains for the north and south tracks before and after the installation of the resilient baseplates. Fig. 5 presents the corresponding measurement results for the passenger trains. No predictions were made for the passenger trains. The measurements are the average of freight trains travelling between 66 and 72km/hr in each case. The predictions are the nearest available at 80km/hr. It can be seen that both baseplate types perform similarly resulting in 5 to 10dB reduction of noise in the 80Hz to 400Hz onethird octave frequency bands. However, around the peak of the noise spectrum near 500Hz, smaller reductions are achieved. It is at this frequency and above that the rail noise dominates over the bridgestructure radiated noise. At and above the wheel is the dominant noise source and there is very little variation before and after the change was made to the track. Additionally, Fig. 4(a) shows the mean of measurements on the track atgrade. These were made at the slightly lower average speed of 60 km/hr. The noise from the wooden sleepers, which was identified by Twins modelling [7] to dominate below 800Hz, is in this case ‘baffled’ by the ballast and the ground reflection is expected to be fairly absorbing. This is in contrast to the open bridge structure and the reflectio n of the water surface. Additionally there is no control of the rail roughness accounted for in the parison. Despite these differences, it can be observed that the bridge noise has been lowered to levels similar to those of the atgrade track in the frequency bands up to 250Hz. The rail noise ponent is clearly much lower from the atgrade track than from the bridge (800Hz to ). 畢業(yè)設(shè)計(jì)外文文獻(xiàn)及譯文 9 Fig. 4. Comparison of predictions with measurements of average of freight trains。 (a) north track。 (b) south track (, measured before。 ??, measured after, ──, predicted before。 , 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。 (a) north track。 (b) south track (──, before。 , after, right graph, measured at the bridge with the steel sleepers still in place but with resilient baseplates installed) Fig. 4(b) presents the noise measured when the resilient baseplates had been in stalled but the steel sleepers had not yet been replaced. It shows that the reduction achieved by the baseplates in the 80 to 400Hz range is promised by the steel sleepers by around 2 to 3dB. The results for the passenger trains shown in (Fig. 5) indicate the same trends as 畢業(yè)設(shè)計(jì)外文文獻(xiàn)及譯文 10 the freight trains. The 1kHz peak in the spectrum of noise from the south track is probably equipment noise from the trains travelling in this direction. 5 Conclusions The bridge is shown to give much higher levels of noise than nearby track at grade. The installation of resilient baseplates has reduced the overall level difference from about 11dBA to about 8dBA. However, the baseplates make of greater noise reduction of 5 to 10dB where the bridge structureradiated noise dominates between 80 and 400Hz. The Norbert model has predicted the reduction in noise in the 80 400Hz bands reasonably well although the measured spectra are smoother than those predicted. In the 630 and 800Hz bands, where the rail noise dominates, Norbert has predicted a reduction that is greater than that actually achieved. This leaves the overall noise reduction to be only about 3dB (4dB predicted) for the freight and passenger trains alike. The removal of the hanging steel sleepers was worthwhile to gain the full benefit of the structure noise reduction Promising locations for similar bridge treatments will be identified according to their cost benefit.