PERFORMING RAILWAY TRAFFIC VIBRATION FORECAST USING IN SITU VIBRATION MEASUREMENTS

Transcription

1 PERFORMING RAILWAY TRAFFIC VIBRATION FORECAST USING IN SITU VIBRATION MEASUREMENTS Vitor Rosão 1, C. César Rodrigues,3, Eusébio Conceição 1 1 Faculdade de Ciências e Tecnologia da Universidade do Algarve { vitortadeiarosao@gmail.com ; econcei@ualg.pt } Instituto Superior de Engenharia de Lisboa { crodrigues@deea.isel.ipl.pt } 3 MRA Instrumentação, S.A. { ccesar@mra.pt } Abstract As international normalized methods or perorming railway traic vibration prediction are not available yet, the authors o this paper will analyze the easibility and applicability o an in situ vibration measurement method in accordance with the guidelines establish in ISO The authors will present the achieved results o several vibration measurements, perormed near the railways and the related transer unctions that will allow predicting the vibrations near dierent receiver locations. Keywords: Vibration Forecast; Vibration Measurements; Railway Traic Vibrations. 1 Introduction In this paper the vibration is characterized by each second RMS (Root Mean Square) value o vibration velocity and the maximum value associated with the event is used. Such methodology ollows reerence [1] and reerence []. The unit used is mm/s, in line with the LNEC (Portuguese National Laboratory o Civil Engineering) Criteria explained in reerence [1]. Since the eect o vibrations can occur at 4 levels: 1. Human disturbance due to vibration sensation.. Human disturbance due to noise produced by the structural vibration. 3. Damage to buildings. 4. Intererence with the operation o equipment. 1

2 And since in accordance with the requirements o ISO [3], the typical requencies o interest or each o these eects are as ollows: 1. 1Hz to 80 Hz.. 16Hz to 50 Hz. 3. 1Hz to 500 Hz. 4. 1Hz to 00 Hz The determination o the total value o vibration associated with a particular event should only use the above requency range o interest. This study directs to the human disturbance due to vibration sensation and to human disturbance due to structural noise produced by vibration, so that will be analyzed, or each event, the requency ranges 1 Hz to 80 Hz and 16Hz to 50Hz. This constraint justiies the characterization o RMS values, more suitable or human disturbance [] [4], instead o characterization o peak values, more suitable or the disruption o structures [5]. To determine the global value (v G ) or the ranges 1 Hz to 80 Hz and 16 Hz to 50 Hz, there are used the ollowing equations, where v corresponds to the vibration velocity o each requency band: G,180Hz 80 v v (1) 1 G,1650Hz 50 v v () 16 Objectives It is intended to evaluate the easibility and applicability o a orecast method, based on the use o vibration attenuation actors, determined rom in situ vibration measurements. 3 Method To determine the vibration attenuation actors, the method uses: 1. A mass, which is driven into the ground, to generate an impulse o vibration (Figure 1).. vibrometers (Figure ), with capacity or registration RMS values at every second, in the 1/3 octave bands between 1 Hz and 50 Hz, and able to detect values suiciently small (ideally less than 0,001 mm/s), or simultaneous registration o the impulse o vibration at points.

3 Figure 1 Example o using mass to generate an impulse o vibration Figure Example o a vibrometer consisting o an accelerometer and a signal analyzer When the impulse occurs, it is measured or each second the RMS value o vibration velocity, on the 1/3 octave bands between 1 Hz and 50 Hz (v ) simultaneously at two points (P 1 and P ) at known distances (d 1 and d ), and is ound the maximum values (v max, ) associated with the impulse. Assuming the ollowing theoretical relationship between the vibration velocity in P 1 and P [3] [6]: v k d P v P max,, max,, 1 (3) d1 One can determine the attenuation actors k based on the ollowing equation (applied to each requency band): vmax,, P log vmax,, P1 k (4) d log d1 3

4 The place o the impulse must coincide, or be as close as possible, to the place where the new railway line is expected to be installed, P 1 must coincide, or be as close as possible, to the reerence distances where the uture vibration can be known, and P must coincide, or be as close as possible, to where one wants to predict the vibration. The test should be repeated at least 3 times and should be used, or security reasons, the maximum values o k (k max, ). As these values correspond to attenuation actors or a point source, it is necessary to transorm them into attenuation actors or linear sources (K line, ) using the ollowing equation (applied to each requency band): K linear, log l v d x d max,, P l0 l v d 1 x, P1 d d log d1 max, l0 1 kmax, kmax, dx dx (5) Using a table o integrals [7], it appears that, in a generic way: n1 n ax b ax b dx or n -1 (6) n 1a Using this general equation and considering l =, we can demonstrate the ollowing important relation: K k 0,5 (7) linear, max, 4 Tests perormed The tests were carried out along the Railway Line in Algarve (Portugal) at the ollowing sites: Site 1: Next to the Teatro das Figuras in Faro, corresponding to a paved area where trains have lower velocity, given the proximity o the station (Figure 3). Site : In rural area, not paved, where trains have a higher velocity (Figure 4). Thus, Measurements belonged to two kinds o vibration sources: Impulse induced: 1 mass o 13 kg (Figure 1). Dierent trains running on the line. Were used the ollowing equipment: 1 vibrometer including an IMI 66A04 accelerometer and a signal analyzer Svantek vibrometer including a PCB 393B31 accelerometer and a signal analyzer Svantek 946A. 4

5 Figure 3 Photo illustration or Site 1 Figure 4 Photo illustration or Site 4.1 Impulse induced The results obtained or the vibration attenuation actors k max are presented in Table 1. It is noted that the mass and distances used only allow the vibrometers to distinguish the occurrence o the impulse or requency bands greater than or equal to 8 Hz. However, such limitation is not a problem or the requency range associated with normal passage o trains, as will be demonstrated on the next chapter. 5

6 Table 1 Speeds and vibration attenuation actors obtained in situ (impulse). Frequency v max, [mm/s] k max, Band Site 1 Site [Hz] P 1 P P 1 P Site 1 Site 8 0,0005 0,0004 0, , ,16-0, ,0007 0,0005 0, , ,34-0,17 1,5 0,0008 0,0005 0, , ,38-0, ,0014 0,0006 0, , ,67-0,64 0 0,003 0,0009 0, ,0006-1,01-0,87 5 0,0061 0,000 0,0085 0,000-0,89-1,1 31,5 0,015 0,0033 0,0040 0, ,09-1, ,051 0,0035 0,0084 0, ,60-1,3 50 0,053 0,0038 0,0138 0,0009-1,55-1, ,0054 0,0010 0, , ,40-1, ,0075 0,0005 0,0330 0,0004 -, -, ,0083 0,0005 0,0377 0, ,33 -, ,0048 0,0003 0, ,0001 -,7-1, ,0047 0,0001 0, ,0001-3,57-1,7 00 0,0015 0,0000 0,0030 0, ,34-1,7 50 0,0005 0,0000 0, , ,61-1, Hz 0,0395 0,0067 0,0404 0,0010-1,45-1, Hz 0,0409 0,0067 0,0567 0,0009-1,48-1,80 4. Trains In Table and Table 3 are presented the measured values in P 1 due to passing trains, the orecasted values in P, based on attenuation actors determined by the method and presented in Table 1 (corrected or linear attenuation actors), and the values actually measured in P due to passing o trains, or Site 1 and Site. Table 4 shows the dierences between predictions and measurements. In the graphs o Figure 5 and Figure 6 are shown the spectrum o vibration velocity o the trains, at P 1, or Site 1 and, being visible the irrelevance o the values associated with the requency bands below 8 Hz. 6

7 Frequency Band [Hz] Table Comparison o measurements and orecasts (Trains, Site 1). measured in P 1 [mm/s] expected in P [mm/s] measured in P [mm/s] Type o train ,0077 0,0009 0,0011 0, ,0019 0,0037 0,0017 0,0009 1,3 0,0008 0,00 0,005 0, ,0009 0,0016 0,0015 0,00 1,6 0,0005 0,0006 0,0005 0, ,0008 0,0013 0,0006 0,0005 0,001 0,001 0,0004 0, ,0008 0,003 0,0004 0,0006,5 0,0007 0,0004 0,0007 0, ,0004 0,003 0,0004 0,0004 3,1 0,0003 0,0006 0,0005 0, ,0003 0,0014 0,0004 0, ,0006 0,0008 0,000 0, ,0007 0,0006 0,0005 0, ,0005 0,001 0,0005 0, ,0007 0,0011 0,000 0,001 6,3 0,0006 0,0011 0,0006 0, ,0004 0,0005 0,0004 0,00 8 0,0008 0,0007 0,0008 0,0003 0,0011 0,0006 0,0007 0,0003 0,0006 0,001 0,000 0, ,001 0,004 0,001 0,0005 0,0014 0,008 0,001 0,0006 0,0006 0,001 0,0007 0,0009 1,5 0,0018 0,003 0,0013 0,008 0,00 0,0034 0,0015 0,0031 0,0009 0,0041 0,0044 0, ,0035 0,005 0,0093 0,0019 0,003 0,004 0,0079 0,0016 0,0015 0,007 0,009 0, ,013 0,0154 0,0089 0,0144 0,0075 0,0094 0,0054 0,0087 0,006 0,0069 0,0098 0, ,098 0,0668 0,0961 0,055 0,003 0,0456 0,0656 0,0377 0,0139 0,016 0,065 0, ,5 0,0748 0,0985 0,113 0,085 0,04 0,0553 0,0681 0,0463 0,039 0,0531 0,036 0, ,074 0,119 0,1301 0,0666 0,047 0,0406 0,0444 0,07 0,0133 0,057 0,075 0, ,0446 0,0847 0,0608 0,0365 0,016 0,0303 0,018 0,0131 0,0187 0,0104 0,056 0, ,0419 0,07 0,0591 0,0571 0,0174 0,098 0,045 0,037 0,0094 0,0037 0,0104 0, ,047 0,0931 0,155 0,0688 0,0046 0,0173 0,033 0,018 0,0017 0,001 0,0059 0, ,0145 0,0641 0,0971 0,0377 0,004 0,0107 0,016 0,0063 0,0013 0,0019 0,004 0, ,005 0,0108 0,019 0,0068 0,0009 0,0019 0,0034 0,001 0,0006 0,001 0,001 0, ,0048 0,0059 0,016 0,0051 0,000 0,0003 0,0006 0,0003 0,0005 0,001 0,0005 0, ,005 0,0047 0,0095 0,005 0,000 0,0003 0,0006 0,0003 0,000 0,0007 0,000 0, ,0007 0,0015 0,0054 0,0048 0,0001 0,000 0,0007 0,0006 0,0001 0,000 0,000 0,0001 v G,1-80Hz 0,177 0,8 0,53 0,1544 0,0586 0,095 0,114 0,071 0,0378 0,065 0,0608 0,0503 v G,16-50Hz 0,185 0,3 0,71 0,1593 0,0586 0,0955 0,1136 0,0714 0,0378 0,0646 0,0607 0,05 Type o train: 1: Ala, : Intercity, 3: Regional; 4: Freight. 7

8 Frequency Band [Hz] Table 3 Comparison o measurements and orecasts (Trains, Site ). measured in P 1 [mm/s] expected in P [mm/s] measured in P [mm/s] Type o train ,013 0,0007 0,000 0, ,0009 0,0009 0,0014 0,0011 1,3 0,0015 0,0015 0,000 0, ,0015 0,0016 0,0008 0,0009 1,6 0,008 0,0003 0,0069 0, ,0011 0,0005 0,0005 0,00 0,0018 0,0007 0,006 0, ,0004 0,0007 0,0004 0,0005,5 0,0011 0,0001 0,0017 0, ,0008 0,0007 0,0004 0,0004 3,1 0,001 0,001 0,0095 0, ,0005 0,0007 0,0003 0, ,0035 0,0011 0,008 0, ,0011 0,0017 0,0003 0, ,0088 0,0018 0,0034 0, ,0017 0,0014 0,0006 0,00 6,3 0,0109 0,015 0,0041 0, ,007 0,004 0,003 0, ,006 0,0134 0,0058 0,0156 0,010 0,07 0,0098 0,064 0,0016 0,0156 0,0057 0, ,0197 0,013 0,0097 0,0167 0,0345 0,031 0,017 0,093 0,0053 0,0154 0,004 0,0136 1,5 0,014 0,057 0,0155 0,013 0,015 0,075 0,0166 0,0139 0,0041 0,0074 0,0045 0, ,006 0,0094 0,0061 0,0087 0,016 0,0074 0,0048 0,0069 0,0104 0,0065 0,00 0, ,0134 0,0176 0,0099 0,017 0,0071 0,0094 0,0053 0,009 0,004 0,0111 0,0083 0, ,065 0,057 0,006 0,037 0,009 0,009 0,007 0,0083 0,0088 0,011 0,0049 0,008 31,5 0,0374 0,0319 0,033 0,0379 0,0144 0,013 0,009 0,0146 0,01 0,016 0,0116 0, ,0557 0,065 0,0398 0,055 0,0138 0,0155 0,0099 0,013 0,0199 0,051 0,0138 0, ,0573 0,177 0,0866 0,0888 0,0084 0,0318 0,017 0,013 0,011 0,008 0,0088 0, ,119 0,6 0,0781 0,165 0,0147 0,089 0,0101 0,014 0,0098 0,009 0,004 0, ,3004 0,08 0,1733 0,149 0,0181 0,0133 0,0105 0,009 0,0039 0,0047 0,00 0, ,377 0,1555 0,140 0,0656 0,0149 0,0061 0,0055 0,006 0,0037 0,0047 0,0014 0, ,0795 0,0495 0,081 0,064 0,0065 0,0041 0,003 0,00 0,0017 0,005 0,001 0, ,036 0,041 0,068 0,014 0,003 0,0051 0,0034 0,0018 0,0016 0,0034 0,0008 0, ,0084 0,096 0,0155 0,0071 0,0011 0,0037 0,0019 0,0009 0,003 0,0016 0,0007 0, ,003 0,0171 0,0106 0,0048 0,0006 0,0033 0,00 0,0009 0,0019 0,0008 0,0005 0,0008 v G,1-80Hz 0,3363 0,3911 0,166 0,516 0,0541 0,0667 0,0363 0,0551 0,033 0,0477 0,046 0,0484 v G,16-50Hz 0,511 0,456 0,605 0,603 0,041 0,054 0,067 0,036 0,0317 0,0419 0,09 0,0405 Type o train: 1: Ala, : Intercity, 3: Regional; 4: Freight. Table 4 Dierences between predictions and measurements (global values). (v Gexpected v Gmeasured )/v Gmeasured Site Ala Pendular Intercity Regional Freight 1-80Hz 16-50Hz 1-80Hz 16-50Hz 1-80Hz 16-50Hz 1-80Hz 16-50Hz 1 55% 55% 46% 48% 85% 87% 4% 43% 64% 9% 40% 5% 48% 17% 14% -11% 8

9 Figure 5 Spectrum o vibration velocity o trains in P 1 (Site 1) Figure 6 Spectrum o vibration velocity o trains in P1 (Site ) 9

10 5 Conclusions The analysis o the previous tables shows that, overall, the estimates are mostly higher than the measurements, which put us in a position o saety. The deviations ound in terms o signiicant global values range between -11% and 87%. These deviations may seem high, but are o the order o magnitude o the deviations ound when using the calculation o attenuation actors corresponding to the passage o a particular train - without the correction o linear source - to predict the values corresponding to the passage o other trains, which demonstrates the suitability o the method and the diiculty o obtaining more precise methods. Acknowledgments The authors thank the ollowing entities and persons: MRA Instrumentação: the availability o Svantek analyzer 958 and IMI 66A04 accelerometer. SCHIU, Engenharia de Vibração e Ruído: the availability o Svantek analyzer 946A and PCB 393B31 accelerometer. Carlos Alberto Rosado: the availability o mass used or the impulse o vibration. Zélia Ramos and João Rosado: assistance in the measurements. Reerences [1] Schiappa, F.; Patrício, J.. Annoyance and damage in buildings caused by vibrations. Considerations or a vibration control good practice. Guimarães, Acústica 004. [] United States o America Federal Transit Administration. Transit noise and vibration impact assessment [3] ISO, International Standard : Mechanical vibration. Ground-borne noise and vibration arising rom rail systems. Part 1. General guidance [4] ISO, International Standard 631-1: Mechanical vibration and shock. Evaluation o human exposure to whole-body vibration. Part 1: General requirements [5] ISO/DIS, Drat International Standard 4866: Mechanical vibration and shock. Vibration o ixed structures. Guidelines or the measurement o vibrations and evaluation o their eects on structures [6] Suhairy, Sinan Al. Prediction o ground vibration rom railways. Swedish National Testing and Research Institute [7] 10