NOISE AND VIBRATION MEASUREMENTS OF CURVE SQUEAL NOISE DUE TO TRAMS ON THE TRACK

Transcription

1 19 th INTERNATIONAL CONGRESS ON ACOUSTICS MADRID, 2-7 SEPTEMBER 2007 NOISE AND VIBRATION MEASUREMENTS OF CURVE SQUEAL NOISE DUE TO TRAMS ON THE TRACK PACS: Lj Volz, Rudi 1 ; Feldmann, Joachim 2 1 Technical University of Berlin, Institute of Fluid Mechanics and Engineering Acoustics, Secretariat TA 7, Einsteinufer 25, D Berlin, Germany; rudi.volz@tu-berlin.de 2 Technical University of Berlin, Institute of Fluid Mechanics and Engineering Acoustics, Secretariat TA 7, Einsteinufer 25, D Berlin, Germany; joachim.feldmann@tu-berlin.de ABSTRACT Within Subproject "Low noise trains and tracks - curve squeal noise: development of effective practical applications to reduce curve squeal on wheel and track" funded by the German BMWi (Federal Ministry of Economics and Technology) the Institute of Fluid Mechanics and Engineering Acoustics has taken air borne and structure borne sound measurements on a curved and on a straight section of tram railway track in order to validate the simulation tool. The acceleration was measured at several positions along the rail, the sound pressure was measured near the track on both sides and the measurements were taken for different speeds of the tram. The results will be presented and discussed. MEASUREMENTS The measurements were taken at a curved section of track in Bielefeld (Germany) on the railway system of "mobiel" near the station "Luther-Kirche" in April At this location there were many examples of curve squeal and previously in January 2006 early measurements were taken at the same location. The curve radius was 30 m. The air borne sound was measured at several distances from the track at several positions. Also the acceleration was measured at several positions along the rail. The measurements were taken on the straight and the following curve. The curve is on an incline. The tram was of the type MS8C (dead weight: 38 tons) with four bogies. The first wheelset at the second bogie (in red in figure 1) had three attached acceleration sensors on every wheel. The Institute of Land and Sea Transport Systems at the Technical University of Berlin was responsible for these measurements. Both sets of wheels on the second bogie lacked absorbers while all the other wheelsets had attached wheel absorbers. Figure 1.- Tram MS8C, red: wheels with sensors, green: wheels without sensors, blue: normal wheels (with wheel absorbers) The train passed at three different speeds in both directions. The airborne sound was measured at a distance of 1.5 m from the middle of the track on both sides and at a height of 0.3 m above the upper edge of the rail. Also there was a microphone at a distance of 7.5 m from the middle of the track at a height of 1.2 m above the upper edge of the rail and inside the curve. The structure borne sound was measured at 15 positions along the rail both above and between the sleeper. On the straight section there was one microphone and three acceleration sensors. Figure 2 shows an overview of the curve and two microphones near the track. Figure 3 shows the photoelectric relay used to detect the wheels and some of the acceleration sensors.

2 Figure 2.- Overview: place of measurements Figure 3.- View from above: inner rail Figure 4 shows the positions of the acceleration sensors on the curve (black numbers: inner rail, blue numbers: outer rail). Figure 5 shows the microphone near the track on the inner side of the curve and the acceleration sensors on the outer rail. All channels were recorded at a sampling frequency of 25.6 khz. Figure 4.- configuration at the rail Figure 5.- View onto the outer rail RESULTS As an example some selected results of the measurements are shown. For the acceleration sensor at the head of the rail in the horizontal direction and the inner microphone near the track. The direction of the train was from the right in figure 1 and downhill to the right in figure 2 so that the wheelset with the instrumented wheels was the second bogie. It was observed that there was no curve squeal when the train drove in the opposite direction. The results shown here were measured at a speed of 10 and 30 km/h. The spectra of the measured time signals are shown as x-y plots and also contour maps of the spectra. Figure 6 shows the spectrum of the velocity level at the head of the rail in the horizontal direction, at the inner rail of the curve and a speed of 10 km/h. This corresponds to measurement point 6 in figure 4. The fundamental frequency of the curve squeal is at 1580 Hz and the upper harmonics up to 10 khz can also be seen. Comparing with figure 7 which shows the sound level from the microphone near the track on the inner side of the curve it can be seen that the lower frequencies are more attenuated because the radiation of sound is more efficient for the higher frequencies. So the dominant tone is the fundamental tone of the curve squeal and the level at the microphone is above 130 db. Also for the air borne sound the harmonics up to 8 khz are well pronounced. 2 19th INTERNATIONAL CONGRESS ON ACOUSTICS ICA2007MADRID

3 Figure 6.- velocity level spectrum, 10 km/h Figure 7.- sound level spectrum, 10 km/h To get a more detailed view in the time domain of the signal the next figures show contour plots. The horizontal axes show the time and the vertical axes show the frequency. The level is shown by colour. Similar to figure 6, figure 8 shows the velocity level at the head of the rail in the horizontal direction, at the inner rail of the curve with a speed of the tram of 10 km/h. The size of the tram is scaled down in the figure. Also the signal from the photoelectric relay of detecting the wheels is added (white curve). The low frequencies (below 500 Hz) are dominated by the rolling noise. Because of damping in the low frequencies the bogies can be detected and in some cases so can the wheelsets. When looking on figure 8 at the bottom one can see the red areas near 6.5, 8.5, 10 and 11.5 seconds that correspond with the four wheelsets and coincides with the signal from the photoelectric relay and the downsized picture of the tram. The fundamental frequency of the curve squeal can be seen as a horizontal line at 1580 Hz and also all the harmonics ( 3160, 4740, 6320, 7900 and 9480 Hz). Because of low damping in this frequency range no detection of bogies or wheelsets is possible. Figure 9 shows a section of Figure 8 from 1450 to 1700 Hz. In this figure it can be observed that the fundamental frequency of the curve squeal changed from 1587 to 1570 Hz when the second bogie passed the photoelectric relay. Figure 8.- velocity level, v = 10 km/h Figure 9.- Detail of spectrum, v = 10 km/h Similar to figure 8, figure 10 shows the sound level from the microphone near the track, at the inner rail of the curve and with a tram speed of 10 km/h. The rolling noise at lower frequencies is reduced due to less radiation. The fundamental frequency of the curve squeal can also be seen as horizontal line at 1580 Hz and the harmonics can be observed up to 8 khz. Figure 11 shows a section of figure 10. Similarly in the air borne sound a change in frequency from 1595 Hz to 1560 Hz occurs when the second bogie passes the photoelectric relay. 3

4 Figure 10.- sound level, v = 10 km/h Figure 11.- Detail of spectrum, v = 10 km/h The next figures show the results for a tram speed of 30 km/h. Figure 12 shows the velocity level at the head of the rail in the horizontal direction at the inner rail of the curve. At lower frequencies (below 500 Hz) the bogies and especially the wheelsets from the second bogie can be detected because of damping in this frequency range. At this speed the fundamental frequency of the curve squeal also lies at 1580 Hz. The upper harmonics are detectable up to 10 khz. Figure 13 shows a section from figure 12 where it can be seen that the fundamental frequency of the curve squeal changes from 1587 to 1558 Hz. At this speed the change of frequency occurs when the first wheelset from the second bogie passes the photoelectric relay. At this moment it can be seen that a vertical line in the spectra appears. This is the wheelset equipped with the acceleration sensors. Figure 12.- velocity level, v = 30 km/h Figure 13.- Detail of spectrum, v = 30 km/h Figure 14 shows the sound level from the microphone near the track at the inner rail of the curve with a tram speed of 30 km/h. Compared with the structure borne sound measurements the rolling noise at lower frequencies is reduced due to less radiation. The fundamental frequency of the curve squeal can also be seen as a horizontal line at 1580 Hz. The harmonics can be observed up to 8 khz. Figure 15 shows a section from figure 14. Showing a change in frequency from 1605 Hz to 1538 Hz associated with the passing of the second bogie at the photoelectric relay. 4

5 Figure 14.- sound level, v = 30 km/h Figure 15.- Detail of spectrum, v = 30 km/h Doppler-Effect The change in frequency of the fundamental frequency in the air borne sound is explainable by the Doppler-effect given by the equation f D = f 1 v ( Eq. 1) c with f D : shifted frequency due to the Doppler-effect, f: frequency of the moving source, v: speed of the moving source in the direction of the microphone and c: propagation velocity. f = Hz, c = 344 m/s and a speed of 2.78 m/s (10 km/h) gives a shifted frequency of f D = 1590 / 1565 Hz. Speed of 8.33 m/s (30 km/h) gives a f D = 1616 / 1539 Hz. These fit well with the measured shifted frequencies. The shifted frequencies in the structure borne sound may be also determined from the Doppler-effect but the propagation in the rail is more complicated than in the air because of more types of wave propagation. CONCLUSIONS Squealing generates strongly forced harmonics in the radiated sound as well as in the structure borne sound. Apart from the fundamental frequency no other eigenfrequencies seem to play a significant role from either the wheel or rail. The results show that curve squeal in the measured tram passings only accurs at the first inner wheel of the second bogie which are the only wheels without attached wheel absorbers. No squealing was measured from the acceleration sensors on the outer rail and when the tram went in the other direction. In the contour maps the frequency shift coincides well with the passing of the first wheel in the second bogie. References: [1] J. N. Vincent, J.R. Koch, H. Chollet. J.Y. Guerder: Curve squeal of urban rolling stock - Part 1: State of the art and field measurements. Journal of Sound and Vibration, Vol. 293 (2006) [2] F. Böhm, K. Knothe (eds.): Hochfrequenter Rollkontakt der Fahrzeugräder. Results from SfB 181 (in German), Deutsche Forschungsgemeinschaft DFG, Weinheim, Wiley-VCH, ISBN (1998) [3] D.J. Thompson, A.D. Monk-Steel, C.J.C. Jones: Railway Noise: Curve Squeal, Roughness Growth, Friction and Wear. Real Research UK, Report: RRUK/A3/1 (2006) [4] H.-J. Giesler: Geräuschemissionen von Straßenbahnen. Nahverkehr, Vol. 4/2000, (2000) 10 [5] U. Möhler, G. Prestele, H.-J. Giesler, W. Hendlmeier: Schallemissionen von Schienennahverkehrsbahnen. Zeitschrift für Lärmbekämpfung, Vol. 6/1998 (1998) 209 5