Five Years’ Monitoring Data on Rail Damper Performance Wilson HO 1 , Max YIU 2 , Ron WONG 3 Jabez Innovation Limited Unit 601, Block A, Shatin Industrial Centre, Sha Tin, N.T., Hong Kong

ABSTRACT Roughness in rail surface accumulates with train operations and grows faster on curved tracks than tangent tracks. Higher roughness leads to higher noise radiation. In Hong Kong, railway noise pre- diction includes a 3dB correction for rail surface deterioration. Depending on rate of deterioration, different rail grinding cycles (typically between 3 and 24 months) are scheduled to ensure train op- erations without excessive noise radiation. Before rail dampers installation, noise levels were moni- tored for 1.5 years and around the noise limit with variation +/- 3dB(A) due to rail roughness varia- tion. After installation of rail damper, railway noise was generally maintained at 7dB(A) below the noise limit with variation +/-1.5dB(A) for more than 3 years. Noise reduction at dominant frequen- cies 630 Hz and 800 Hz were more than 10dB in 1/3 Octave analysis. The rail dampers were installed at alternative spaces such that additional rail dampers can be installed to double the amount in case further noise reduction is required. This paper examines the five years’ monitoring data in detail, showing the noise levels with smaller variations over grinding cycles after rail damper installation, and explores rail damper potential ability to slow down the growth rate of rail roughness levels.

1. INTRODUCTION

Since May 2017, noise from railway passby is monitored regularly for a section of curved track closed to a residential premises. Nine months after the noise monitoring started, the track was in- stalled with Shearing Tuned Mass Damper (STMD). This paper examines the five years’ monitoring data in detail, showing the noise levels with smaller variations over grinding cycles after rail damper installation, and explores rail damper potential ability to slow down the growth rate of rail roughness levels and may achieve long term noise reduction.

2. SITE INFORMATION

2.1. Monitoring Location

A microphone equipped with weather kit was installed at the balcony of a residential building facing the railway track at 6th floor on top of a podium. The closest slant distance is ~30m from the track, as shown in Figure 1 below.

2.2. Track Configuration and Rail Damper Installation

The railway track is a viaduct and 10m above ground level. It is a floating slab track and installed with high resilient base plate (Delkor Egg) at 0.6m interval. The residence reported experiencing high noise level during train passby. The train operating company installed Shearing Tuned Mass Damper (STMD) as a trial to reduce noise level, as shown in Figure 1 . The STMD had reported reducing rail corrugation growth rate [1-3].

1 who@jabez.hk 2 max.yiu@jabez.hk 3 ron.wong@jabez.hk

worm 2022

worm 2022

Rail dampers are installed for 80m long track at up track at alternative rail mounting spacing, i.e., half of the maximum number per unit length, on 5 th Nov 2018 and 9 th Nov 2018. They are tuned at 650Hz resonance frequency in both

Figure 1: Microphone Location and Damper Design

No Dampers

——_ mt tot, la UN ies! .

Dampers

Figure 2: On-site Photos of STMD

2.3. Different Rail Grinding Method Throughout the Monitoring Period

Typical rail grinding was conducted before June 18. Special rail surface milling treatment was conducted between June 18 and right before damper installation on Nov 2018. Typical rail grinding resumed after rail dampers are installed.

Rail milling treatment produce a very smooth rail finishing without grinding marks, this would not induce noise from the rail. However, rail milling is a costly process and did not adopt very frequently. Rail roughness would be increase and normalized if rail grinding treatment resumes.

Table 1: Rail Treatment Schedule and STMD Condition

Period Rail surface treatment type STMD condition

May 2017 – Jun 2018 Normal Grinding No STMD

Jun 2018 – Nov 2018 Special Milling No STMD

Nov 2018 – Mar 2022 Normal Grinding STMD Installed

3. MONITORING RESULTS

3.1. Monitoring Period and Parameters

Average up track passby sound exposure level (SEL) of most common type of train, SP1900, was extracted weekly on the first day of the week that is not public holiday, i.e., mainly Mondays, from 0630 to 0700. The frequency is 7-8 train passby per data point. Double passby and down track passby data were neglected. Up track passby duration remains at 19-22s throughout the monitoring period.

A typical time history, extracted on 29 th Mar 2018, is shown in Figure 3 below. Time correction of ~13dB(A) were applied to the averaged passby L eq,event . All data are presented in relative noise level based on the highest SEL noise level on 29 th Mar 2018, as shown in Figure 4 below.

0

Up track Down Track Double passby

-5

Relative Noise Level, Leq,1s, dB(A)

-10

-15

-20

-25

-30

-35

06:30 06:40 06:50 07:00

Time

Figure 3: Typical Passby Time History

3.2. Five Years Monitoring Results

Average up track passby SEL measured over five years are shown in Figure 4 below. Before rail dampers installation, noise levels were monitored for 1.5 years with variation + 3dB(A) due to rail roughness variation. After installation of rail damper, railway noise was generally maintained at 7dB(A) below the noise limit with variation + 1.5dB(A) for more than 3.5 years. When compared between similar rail surface condition under normal rail grinding cycle, the STMD provide 7dB(A) noise reduction ( Table 2 ). The immediate 2dB(A) noise reduction by STMD is limited by special rail milling.

Table 2: Rail Treatment Schedule and STMD Condition

Period Rail surface treatment

STMD condition

Noise Level Range

type

(Relative)

May 2017 – Jun 2018 Normal Grinding No STMD 0 to -4.6 dB(A)

Jun 2018 – Nov 2018 Special Milling No STMD -4.7 to -6.6 dB(A)

Nov 2018 – Mar 2022 Normal Grinding STMD Installed -6.7 to -11.9 dB(A)

worm 2022

worm 2022

0May 17Jul 17Sep 17Nov 17Jan 18Mar 18May 18Jul 18Sep 18Nov 18Jan 19Mar 19May 19Jul 19Sep 19Nov 19Jan 20Mar 20May 20Jul 20Sep 20Nov 20Jan 21Mar 21May 21Jul 21Sep 21Nov 21Jan 22Mar 22

Each data point represents average SEL of 7-8 SP1900 up

track passyby that lasted for 19-22s from 0630 to 0700

Train Passby Average Noise Level (Relative), SEL, dB(A)

-2

Normal rail grinding cycle

-4

Damper installation (half)

-6

Normal rail grinding cycle

-8

-10

iy

-12

Figure 4: Up Track Noise Monitoring Results of Past 5 Years

3.3. Spectrum Analysis

Spectrum of each passby on 29th Mar 2018, where noise level peaked without dampers, and 7th Dec 2018, around 1 month after damper installation are shown below. It is observed that the dominant frequencies of 630Hz and 800 Hz had reduced over 10 dB.

-10

No Dampers (29th Mar 18)

-15

With Dampers (7th Dec 18)

Noise Level, Leq, Event, dB(A)

-20

-25

-30

-35

-40

-45

-50

-55

31.5 63 125 250 500 1000 2000 4000 8000 16000

1/3 Octave Band Centre Frequency, Hz

Figure 5: 1/3 Octave Noise Spectra of Passby

4. OBSERVATION AND DISCUSSION

4.1. Normal Rail Grinding Cycle Before Rail Damper Installation

As shown in Figure 4 , during the first normal grinding cycle without STMD between May 17 to June 18, the noise varies by + 2dB(A) and noise level peaked on 29 Mar 2018. One of the major factors on noise level variation during train passby is rail surface roughness. There are two types of rail surface roughness, corrugation and rail grinding mark.

Corrugation would be developed over time from train operation which increases rail surface rough- ness and result in increased noise level [3]. Corrugation is a wave-like deformation on rail surface and produce tonal noise which is more prominent at a curved track. One of model on corrugation formation is stick-slip model [4].

Periodic rail grinding treatment is a typical measure to reduce rail surface roughness and conducted once per 0.25 - 1 year. Typical grinding mark has 12-28mm interval. Noise level would be increased immediately after rail grinding due to increase rail roughness from grinding mark. Around 2 months after rail grinding, the rail surface would be smoothed out from train operation until corrugation de- velops again [4].

4.2. Rail Milling Treatment Before Rail Damper Installation

It is observed that during the rail milling treatment period, the noise level gradually drops, and noise variation is more contained. This may be due to rail milling treatment produce a smoother rail finishing while corrugation did not have enough time to develop before damper installation. Com- bined with previous normal rail grinding cycle session, the noise varies by + 3dB(A) before rail damper installation due to rail roughness variation.

4.3. Normal Rail Grinding Cycle After Rail Damper Installation

After rail damper installation, normal rail grinding cycle resumes. The noise variation after rail damper installation remains + 1.5 dB(A), and 90% of the time the variation is contained + 1dB(A), which may be due to STMD is able to reduce corrugation growth [1-3]. The highest measured noise level was measured on Oct 2019 and it is 7dB(A) lower than the highest noise level measured without STMD.

After normal rail grinding cycle resumes which normal rail surface condition is recovered, the noise level also remains 7dB(A) lower that normal rail grinding cycle before damper installation ( Table 2 ).

4.4. Rail Damper Performance

Noise reduction of 7dB(A) is achieved by STMD when compared between the two normal rail grinding cycle before and after rail damper installation. Even comparing the peak noise level meas- ured between with and without STMD at normal rail grinding cycle, i.e., 29 th Mar 18 and 30 th Sept 19, the noise reduction of 7dB(A) is also achieved. Their respective spectrum is shown in Figure 6 .

Although the immediate noise reduction of STMD is ~2 dB(A), the true performance is masked by the special rail milling treatment before STMD installation and it is not the normal rail surface condition. Noise monitoring results of the last 3.5 years with STMD shows the noise level trend and variation remain in the range of -8 to -12dB(A) relative noise level ( Figure 5 ). Should the true rail damper performance be ~2dB(A), the relative noise level should gradually increase to the range of - 4 to -6dB(A) after returning to normal rail surface condition.

worm 2022

It is anticipated that 10 dB(A) noise reduction could be achieved if the STMD are installed at each rail mounting spacing, i.e., double the number at the same unit of track length, since halving the energy in the system would results an additional 3dB reduction.

-10

29th Mar 18 (Peak noise level w/o damper) 26th Oct 18 (Rail milling w/o damper) 7th Dec 18 (with damper) 30th Sept 19 (Peak noise level with damper)

Average Noise Level, Leq, all

-15

-20

passby, dB(A)

-25

-30

-35

-40

-45

-50

-55

31.5 63 125 250 500 1000 2000 4000 8000 16000

1/3 Octave Band Frequency, Hz

Figure 6: Noise Spectrum at Different Stages of Noise Monitoring

5. CONCLUSIONS

Noise monitoring was conducted over 5 years for an 80m railway track section with excessive noise peaked at 630-800Hz. After 1.5 years of noise monitoring starts, Shearing Tuned Mass Damper (STMD), tuned to 650Hz, was installed in alternating rail mounting spacing for the 80m section (half of the normal number of STMD per unit length).

When compared between similar rail conductions during normal rail grinding cycle, the STMD reduced 7dB(A) noise and maintained the noise level since the STMD installation 3.5 years ago. Noise variation due to rail roughness is contained from + 3dB to + 1dB for 90% of the time after STMD installation, which may be due to STMD is able to suppressing rail corrugation growth.

Special rail surface milling treatment was conducted fore half year period before the STMD in- stallation. It is believed the measured immediate 2dB(A) noise reduction of STMD cannot truly re- flect its true performance since the milling treatment had produced anormal rail conditions.

It is anticipated that 10 dB(A) noise reduction could be achieved for full installation of STMD, i.e., double the number of currently installed STMD per unit length.

6. REFERENCES

1. W. Ho, B. Wong, D. England, A. Pang, and C. W. S. To, "Tuned mass damper for rail noise and

corrugation control", presented at the Acoustics 2012 Hong Kong, 2012. 2. W. Ho, B. Wong, D. Tsui, and C. Kong, "Reducing Rail Corrugation Growth by Tuned Mass

Damper", presented at the The 11th International Workshop on Railway Noise (IWRN11), 2013. 3. W. Ho, D Tse, P. Wong, W. Wong, M. Ip, G. Soltanieh, “Suppression of Corrugation Growth by

Rail Damper”, presented at the The 27th International Congress on Sound and Vibration (ICSV27), 2021 4. P. A. Meehan, R. D. Batten, and P. A. Bellette, "The effect of non-uniform train speed distribution

on rail corrugation growth in curves/corners," Wear, vol. 366–367, pp. 27–37, 2016, doi: 10.1016/j.wear.2016.05.009.

worm 2022