Special edition paper

Research and Development on

Reducing Interior Noise in
Shinkansen Vehicles Shigenari Horikawa* Yoshihiro Fujii * Takeshi Kurita*

Noise inside Shinkansen cars increases with increase in train speed. We have conducted considerable development to reduce this
interior noise and successfully developed new rolling stock with interior noise levels that are equal to or less than that in previous
trains even when the maximum speed increased from 275 km/h to 320 km/h. In this paper, we discuss the direction of development
to further increase the speed of Shinkansen trains by reviewing past developments such as the effects of various types of products
developed and noise measurement results obtained through running tests.

•Keywords: Shinkansen rolling stock, High-speed train, Interior noise

1 Introduction
Air-borne noise Exterior noise (wind noise, rolling noise, etc.) [Measure example]
coming in from car body frame, windows, etc.
(transmitted noise)
Air-borne noise
Car body (transmitted noise)
frame • Attach sound absorbing material
Energy of noise in a train car generally increases in proportion to • Make air layer in window glass thicker
Window
the square to sixth power of running speed of that train. Also, Floor panel Structure-borne noise

sound insulation performance required to reduce that noise is Underfloor Duct

• Use rubber etc. for supporting floor
panel and interior material
equipment • Improve rigidity of floor frame and
proportional to mass of car body, floor, and windows. In other Airtight floor, interior material
Bogie underframe • Change method of installing
words, an effective measure to control increase of interior noise due underfloor equipment

to train speed increase is to increase rolling stock weight by making Structure-borne noise
materials of car body and windows thicker. On the other hand, Vibration of bogie or underfloor
equipment propagated via car
high-speed Shinkansen rolling stock must increasingly be faster body frame vibrates floor panel
and interior, causing noise
Duct noise (direct noise)
Duct noise (direct noise)
• Change shape of duct
Noise of air conditioner, etc. and • Attach sound absorbing material
air flow noise in the duct coming
in from duct space and gaps

(arrive at destinations in shorter time), operate with less power,

and have less ground vibration; and we have been aiming to reduce
body weight as an effective measures for those. Reducing interior
noise and reducing rolling stock weight are contrary to each other;
however, we need to achieve both of those to increase Shinkansen
running speed. In light of that, much research and development
on technology to reduce interior noise without increasing rolling
stock weight has been considered and that technology introduced
to actual trains in commercial operation.
In this paper, we review the results of research and development
the Research and Development Center of JR East Group has
conducted on reducing interior noise .
Causes of and Countermeasures for
2 Interior Noise
There are various sources of interior noise, including aerodynamic
noise from the car body, rolling noise from wheels, and noise
caused by bogie vibration that is propagated to and vibrates the
floor panel and interior. Noise generated from different sources
enters the cabin to become interior noise, so interior noise can
be categorized by transfer path into “air-borne noise (transmitted
noise),” “structure-borne noise,” and “duct noise (direct noise).”
Fig. 1 shows countermeasures against interior noise by cause and
transfer path.
By identifying the noise source and transfer path, we can take
specific measures to reduce interior noise. Particularly, in efforts
to reduce noise caused by speed increase, we need to minimize
increases in car body weight. Correctly identifying the noise
Fig. 1 Causes and Measure Examples of Interior Noise
transfer path from source to cabin and degree of contribution of
individual noises to the total noise and taking measures focusing
on individual targets based on that knowledge are requirements for
achieving interior noise reduction and car body weight reduction
at the same time.
3 Interior Noise Reduction Method
3.1 Flow of Development of Interior Noise Reduction Methods
There are many possible methods of reducing interior noise.
Fig. 2 shows an example of flow of development of a method we
have actually carried out. First, we measure noise and vibration
in stationary tests and running tests to identify actual noise in
detail. Then, according to the measurement results, we analyze
and evaluate degree of contribution to total noise and transfer
paths from sources of individual noises. Based on the analysis
and evaluation results, we focus on the target noises and develop
technical elements appropriate to the part and method that need to
be dealt with, and we then examine how to actually introduce the
items developed to rolling stock while taking into account weight,
cost, and the like. Finally, we return to the first step and check the
effect of the equipped items developed by measuring noise and
vibration in stationary and running tests again. By repeating this
cycle, we have been working on refining interior noise reduction
methods.

*Advanced Railway System Development Center, Research and Development Center of JR East Group JR EAST Technical Review-No.33 19
 Special edition paper
Examination of Element development Focusing on targets the vibration measurement results. Vibration level of floor and
method of installing • Degree of contribution
to rolling stock Part to be dealt with Method of noise to total noise
wainscot panels that are greatly affected by structure-borne noise
• Weight reduction
• Cost • Floor • Supporting method
• Transfer path
did not increase so much in tunnels. In contrast, vibration level
• Space • Wall, ceiling • Material rigidity of windows and ceilings increased greatly. The reason is that air-
• Window, etc. • Material thickness,
etc. borne noise (transmitted noise) increased due to noise reflection in
a tunnel, and this is a characteristic of the part that is affected by
Test/Measurement Acoustic analysis that transmitted noise.
Running test Data obtained Calculation of transfer function In addition to those running test results, we also gained noise
• Measurement of actual • Noise data [Structure-borne noise]
input/output values • Vibration data • Cabin acoustic sensitivity, etc. and vibration data by acoustic vibration and impact excitation
Stationary test
[Air-borne noise (transmitted noise)] methods in stationary tests.
• Car body damping factor, etc.
• Acoustic vibration test
• Impact vibration test
Degree of contribution to total
• Measurement of
noise, transfer path 3.2.2 
Results of Acoustic Analysis (Analysis of Degree of
reverberation time • BEM
• SEA Contribution to Total Interior Noise)
From the running and stationary test data, we calculated transfer
Fig. 2 Example of Development of Interior Noise Reduction Method
path functions that are the rates of outside noise and vibration
3.2 Case Example of Noise Measurement and Analysis Using propagated inside to become interior noise, and we also calculated
Series E2 Shinkansen Train degree of contribution to the total interior noise by transfer paths
3.2.1 Test and Measurement using a noise prediction model by statistical energy analysis (SEA)
Fig. 3 shows as the results of interior noise measurement carried and boundary element method (BEM). Here, Fig. 5 shows the
out in running tests from March to June 2003 using a series E2 calculation results of degree of contribution to total interior noise
Shinkansen train. Interior noise increased as train speed increased. per part.
It increased by approx. 2 dB(A) in an open section and approx.
Frieze Side ceiling
3 dB(A) in a tunnel section as the train speed increased by 40 km/h board Ceiling panel
from 280 km/h.
We also measured vibration of floor panels, interior panels,
windows and the like in addition to noise values. Fig. 4 shows

Open section Window

Tunnel section Wainscot panel
【 】
Interior noise level

Floor (air-borne noise) Floor (structure-borne noise)

* Open section [Tunnel section]

Fig. 5 Rate of Noise Coming into Cabin Directly
Above Bogie of E2 (at 275 km/h)

In the figure, values are the degree of contribution to the total

interior noise in an open section per part, and values in brackets
are those in a tunnel section. The results revealed that nearly 90%
Fig. 3 Results of Noise Level Measurement Directly Above Bogie of E2 of interior noise is transmitted noise and structure-borne noise
from the floor and windows in an open section, and noise from
Increase in tunnel section those two sources accounts for a large percent in a tunnel section
Open section
too, while the rates of the structure-borne noise from the floor and
Vibration acceleration level

transmitted noise from windows are reversed.

3.3 Verification of Interior Noise Reduction Measures Using

FASTECH 360
Based on the measurement results using a series E2 Shinkansen
train, we further worked on development of elements focusing on
some specific noises. Since June 2005, we carried out running
tests using the FASTECH 360 Shinkansen high-speed test train
Floor panel
Wainscot to verify noise reduction effects of individual elements developed.
panel Window
Frieze board
Ceiling panel

Fig. 4 Results of Noise Level Measurement Directly

Above Bogie of E2 (at 275 km/h)

20 JR EAST Technical Review-No.33

 Special edition paper

3.3.1 Measures for Noise from Floor material that was conventionally fixed to the car body frame of
Fig. 6 shows the measures for noise from floor verified using Shinkansen rolling stock.
FASTECH 360. We found that the measures were effective on The evaluation results in the running tests gave us information
transmitted noise for the floor panel structure. For comparison, on type and thickness of sound absorbing material required and
we carried out tests by installing aluminum honeycomb floor part to which it is attached as a measure for reducing transmitted
panels also used in series E2 trains, newly developed aluminum noise. In contrast, we could confirm almost no effect of the
honeycomb floor panels with a rubber plate, and aluminum floor method of elastic support of interior material.
panels with resin foam as the core material. As a countermeasure
against noise structure-borne noise, we adopted a floating floor 3.3.3 Countermeasures for Noise from Windows
structure where floor panels were supported by two types of rubber We adopted a cabin window structure with an air layer between
supports. Moreover, we made a cutout at the end of a joist to make the outer glass and inner glass, and based on the element
a structure where little vibration is propagated between airtight development results, we installed three types of outer glass to the
floor panels and floor panels. FASTECH 360 for verification. Those were multi-layer glass with
and without an air layer in between and polycarbonate glass. Fig. 8
Floor panel
Floor panel structure supporting method Joist structure shows the structures of conventional and tested window types. To
Aluminum honeycomb floor panel Elastic supporting method 1 Joist cutout structure
Aluminum panel Aluminum expand the frequency band where sound can be insulated, the air
honeycomb Floor panel
Section bar layer was made thicker. Window size itself was made smaller while
Rubber
maintaining cabin comfort.
Aluminum honeycomb floor
panel with rubber sheet
Rubber Aluminum Joist Conventional
Aluminum panel Developed structure
sheet honeycomb
Elastic supporting method 2
structure
Chip Multi-layer glass + Multi-layer glass Multi-layer glass Polycarbonate glass
Floor panel Air layer + (with air layer) + (without air layer) + Air layer + Inner glass
Inner glass Air layer + Inner glass Air layer + Inner glass
Aluminum floor panel with resin foam core Polycarbonate
Multi-layer Multi-layer glass Glass Multi-layer
Aluminum panel Rubber Joist glass Glass glass Glass glass Glass
Resin foam (with air layer)
Airtight floor panel

Joist

Floor panel
Air layer

Air layer

Air layer

Air layer
Outside

Outside

Outside

Outside
Inside

Inside

Inside

Inside
material

Joist Space
Airtight
floor
Sound absorbing and
Fixed to car Fixed to Fixed to car Fixed to Fixed to car Fixed to Fixed to car Fixed to
heat insulating material body frame interior material body frame interior material body frame interior material body frame interior material

Fig. 6 Countermeasures for Noise from Floor Fig. 8 Countermeasures for Noise from Windows

The evaluation results in the running tests showed that using
aluminum floor panels with resin foam was most effective of the
three floor structures on reducing transmitted noise. They also
confirmed that elastic support and cutout on the joist greatly
reduced structure-borne noise.
The running test results confirmed the effects of each type of
multi-layer glass and polycarbonate glass on reducing transmitted
noise. From that, we gained insight on the appropriate structure
and thickness of multi-layer glass windows in installing to rolling
stock in commercial use.

3.3.2 Countermeasures for Noise from Walls and Ceiling
Fig. 7 shows the countermeasures for noise from walls and ceiling.
The effect on transmitted noise of attaching sound absorbing
material was known from before, but using more sound absorbing
material leads to increase of car body weight and cost. In light of
that, we checked on the FASTECH 360 the effects of different
types and thickness of sound absorbing material that also works
as heat insulation. We also verified elastic supporting of interior
Car body frame Sound absorbing material
Space
Supporting rubber fixing screw
Fixing screw
Interior panel Interior panel supporting rubber
3.3.4 Other Measures
In addition to the measures explained above, we verified the
following measures using the FASTECH 360 based on the element
development results.
(1) Quiet air-conditioner ducts
In reviewing the structure of air-conditioner ducts, one of the
interior noise sources, we developed a quiet air-conditioner duct
with a simple structure and to which sound absorbing material was
added and confirmed its effects.
(2) Elastic support of underfloor equipment (main transformer)
As a measure to reduce harmonic vibration and noise, we tested
a structure where main transformers area supported with rubber.
The results of verification using FASTECH 360 confirmed
vibration insulation and noise reduction.

Fig. 7 Countermeasures for Noise from Wall and Ceiling

JR EAST Technical Review-No.33 21

 Special edition paper
Employment on the Series E5 and E6 Shinkansen
4 Rolling Stock and Noise Reduction Effects
Series E2
～

We employed interior noise reduction measures confirmed effective Tunnel section

A-weighted sound
using the FASTECH 360 on the series E5 and E6 Shinkansen

pressure level
～
rolling stock that runs at 320 km/h in commercial operation,
taking into consideration noise reduction effects, weight, cost, and
the like. Fig. 9 shows the noise reduction measures introduced.
Open section
Structure-borne Air-borne noise
Window structure Wall and ceiling structure Duct shape noise (transmitted noise)
• Multi-layer glass with • Sound absorbing material • Quiet air conditioner duct
thicker air layer attached where needed * E6 only
• Smaller window structure Frequency

Series E5 Spectrum
Total RMS dB(A) = 69.65
70.0 74.20
～
60.0

Floor panel structure Floor panel supporting method Underfloor equipment Tunnel section

A-weighted sound
supporting method

• Aluminum floor panel
with resin foam core
* Partly used on floor
above bogie, etc.
pressure level
• Elastic supporting method using rubber
• Joist cutout
• Elastic supporting method
～
* Partly used on floor above bogie, etc.

Fig. 9 Development Results Employed on Series E5 and E6 Rolling Stock

Open section
Fig. 10 shows a comparison of total interior noise level between Structure-borne
noise
Air-borne noise
(transmitted noise)
series E2 and E5 rolling stock. This confirms that interior noise 10.0
10 100 1000 2500
could be kept to a level equal to or less than before both in open 17 :1001X+ 1001X+ 1 Frequency 17 :1001X+ 1001X+ 1
09-Aug-13 17:04:52
and tunnel sections by introducing the items developed even when Fig. 11 Comparison of Frequency Analysis Results
speed was increased from 275 km/h to 320 km/h. (at 275 km/h, directly above bogie)

Open section Tunnel section

Less than interior

noise level of
5 Aiming for Further Speed Increases
Equal to interior noise level series E2 train
of series E2 train at 275 km/h at 275 km/h
Interior noise level

We could reduce interior noise by the technologies developed as

described herein even with train speed increase from 275 km/h
to 320 km/h while minimizing increase of rolling stock weight.
However, with further speed increase, rolling stock weight and
production cost will increase and cabin space will be smaller by
using current technologies only.
Aiming to reduce interior noise for the next generation of high-
speed Shinkansen rolling stock, we will work on developing new
Fig. 10 Comparison of Interior Noise Levels (directly above bogie)
analysis methods to identify noise transfer paths and degrees of
Next, Fig. 11 shows frequency analysis results. With the series contribution to total noise in more detail and completely new
E5 train, A-weighted sound pressure level of around the 100 to interior noise reduction methods that are not simply an extension
200 Hz frequency band, which determines total noise level, is kept of conventional rolling stock technologies. We will seek an ideal
low both in open and tunnel sections. As it is known that structure- rolling stock structure that has a good balance between the effects
borne noise such as from bogies is predominant in this frequency of such new interior noise reduction measures and the constraints
band, we could confirm that measures for reducing structure- of rolling stock design in weight, cost, space, and the like.
borne noise keep total noise level down. In the 300 to 1,000 Hz
frequency band, noise level difference with the series E5 train is
6 dB(A) between open and tunnel sections, while that is 10 dB(A) Reference:
with the series E2 train. As transmitted noise is predominant in 1) Daisuke Muto, Katsutoshi Horihata, Kazuhiro Makino, Masahiko
Horiuchi, Katsufumi Hashimoto, Hitoshi Shiraishi, “Experimental
this frequency band, this result shows that measures for reducing and Computational Analysis to Reduce the Noise in High-speed
transmitted noise are effective. We could reduce both structure- Train [abstract in English]”, Proceedings of the 2004 Symposium on
Environmental Engineering (July 2004): 96 - 99
and air-borne noise in a well-balanced manner, resulting in total
2) Masahiko Horiuchi, Katsufumi Hashimoto, Hitoshi Shiraishi,
interior noise reduction. Satoru Akiyama, Yasufumi Minamimoto, “Sound Insulation
Quality of the Windows for High Speed Train [abstract in
English]”, Proceedings of the 12th JSME Transportation and Logistics
Meeting (December 2003): 37 - 40

22 JR EAST Technical Review-No.33