Railway and ICT (Information and Communication … Japan Railway & Transport Review o 67 s Mar 2016 Railway and ICT (Information and Communication Technology) result of changing from

22Japan Railway & Transport Review No. 67 • Mar 2016

Railway and ICT (Information and Communication Technology)

Innovation in Railway Maintenance utilizing Information and Communication Technology (Smart Maintenance Initiative)

Mitsunobu Takikawa

post-war population increase

and high-economic-growth era.

But changes to the economic

structure with falling population,

d i ve r s i f y i n g va l u e s , a n d

globalization, mean conventional

ways of doing business no

longer apply, and not even large

corporations can rest easy.

Companies that are growing

recent ly are creat ing new

business models using the latest

ICT. For example, a manufacturer

of photocopiers successfully

switched from manufacturing

and selling photocopiers to a

new business leasing model,

making large profits (Figure 1).

This new business model was

successful because advances

in ICT allow the company to

identify the operational status

of individual photocopiers at

low cost, enabling maintenance

before failures.

This can be said to be the

Introduction

Information and communication technology (ICT) is

advancing at a tremendous pace in recent years, and is

set to change the social environment and composition

of industry. The ‘Industry 4.0’ national project started

in Germany in April 2013 is attempting to create a fourth

industrial revolution to change the composition of Germany’s

manufacturing industry using the Internet of Things (IoT) and

‘big-data’ technology. In March 2014, the Industrial Internet

Consortium was launched by the private sector in the USA

with participation of big companies, such as GE, IBM, Intel,

Cisco, and AT&T. This plan also uses IoT and big data-

related technologies to make industry as a whole smarter,

and maintenance more efficient.

Manufacture and sell products

Collect fees according to usage and maintenance

Figure 1 Change of Business Model

With major changes to global society using ICT,

Japan’s railways too must change proactively to become

the transportation mode chosen by society in this new

environment. The Research and Development Center of the

JR East Group has proposed a ‘Smart Maintenance Vision’

so as not to fall behind in these changes, and is working

on R&D for innovation of railway maintenance. This article

introduces trends in ICT and the Smart Maintenance Vision

along with case examples of R&D currently under way.

Recent Social Conditions in Japan

Japan’s population is already declining. Its industry

experienced many past successes in high-quality

monozukuri (manufacturing) by mass-production amidst the

23 Japan Railway & Transport Review No. 67 • Mar 2016


result of changing from tangible manufacturing to intangible

services, utilizing ICT at the core.

Looking at railway maintenance from the same

perspective, the function of conventional R&D has been to

break from 3D (dangerous, difficult, dirty) work and reduce

equipment maintenance costs. For example, in the area of

tracks, low-maintenance tracks (Figure 2) and the like have

been developed to mechanize inspection and repair so as to

reduce the need for maintenance. This has reduced failures

and human errors using a tangible approach. However,

railways face a major change in the social environment, so

maintenance must change too. Looking at case examples in

other industries, intangible approaches will be needed.

Latest Trends in ICT

This section covers trends in ICT and discusses ideas for

future railway maintenance.

Internet of Things (IoT)IoT is a method for enabling ‘things’ to exchange information

via the Internet. In the future, we may exchange information

based on data gathered via sensors while being unaware

that the other side is a machine; Machine to Machine (M2M)

is another term with similar meaning but the IoT concept

includes people as ‘things’. The main purpose of M2M is

to increase efficiency by automatic control of machines,

aiming for a ‘smart’ society. IoT has come to the focus of

attention due to the increase in devices and sensors that

can be carried by people and the creation of a ubiquitous

communications environment. For example, smartphones

have many sensors, such as GPS (Global Positioning

System) receivers and accelerometers, in addition to

telephone and email functions. It is no exaggeration

to say they are full-fledged all-purpose sensors with

communications functions, and it is amazing that such

devices can be purchased for less than ¥100,000

(approximately US$850).

IoT is already being used for purposes such as

monitoring aircraft flights and automobile status. For

example, in the aviation industry, sensors throughout the

aircraft monitor the status during flight. The large volumes

of data are analyzed and used to identify signs of failures in

advance. In railways, embedding sensors in rolling stock and

mechanical equipment would enable similar maintenance.

Big Data AnalyticsBig-data analytics has been used in recent years in areas

such as online shopping recommendation functions and

social network analysis. In IT companies and among

experts, ‘big data’ means more than just lots of data, and

is interpreted as data leading to knowledge that is helpful

in corporate management and business and in people’s

lives. Big data includes diverse types as shown in Figure

3, including text information typified by ‘tweets’ and similar

social networking services (SNS), photos and videos,

logs generated when accessing websites and the like,

and information from various sensors. The data volume

amounted to 2.8 Zettabytes (ZB) (2.8 × 1021 bytes) in 2012,

and is expected to grow 14-fold to 40 ZB in 2020.

In addition to services, there are expectations for big

data technologies in the maintenance field. The problem

of aging infrastructure built during Japan’s high economic

growth era has come into focus, and the Ministry of

Wide sleeperCement-filled layer

Nonwoven fabric

Figure 2 Low-maintenance Track


Internal Affairs and Communications (MIC) estimates that

the economic impact of preventive maintenance on road

bridges will be ¥270 billion. Bridges, tunnels, and other

structures are maintained for railways as well because they

are part of the social infrastructure. Like roads, applying

big-data analytics technologies to maintenance work should

bring about new industries and help make maintenance

more efficient.

Artificial Intelligence (AI)Research on AI started in the 1950s, and rule-based expert

systems were in the spotlight in the 1980s. These expert

systems were an attempt to express the complexity of the

real world by rules alone, but they ended in failure. Reflecting

the lessons learned, current AI has become closer to the

thought patterns of humans by ignoring rules for data input

and speculating on results by statistical processing. Recent

examples of AI are Watson, which beat the champion of a

famous American TV quiz programme in 2011, and Ponanza,

which beat a professional shogi (Japanese chess) player in

an official match in 2013.

Current AI systems are based on machine learning

where humans teach answers. That machine learning

evolved further with Deep Learning developed by Stanford

University and Google in the USA in 2012. Deep Learning

successfully obtained the concept of a ‘cat’ from images

on a network without being taught by humans. In other

words, AI became able to learn on its own without relying on

humans to build knowledge. This Deep Learning technology

is used not only in image recognition, but also as voice

recognition technology in smartphones and tablets.

The knowledge of AI including Watson and Ponanza is a

database where past knowledge is accumulated. In railway

maintenance, past data gained from periodic inspections

and reports of incidents is stored on many systems. If this

data could be utilized effectively, there is a high possibility

that we could build AI with abilities greater than veteran

engineers.

Environment Surrounding Railway Business

This sect ion descr ibes issues re lated to fu ture

maintenance in terms of the environment surrounding JR

East’s railway business. First, we organize issues from the

standpoint of maintenance costs, and second, from the

makeup of personnel.

Railway Maintenance ExpensesFigure 4 shows the changes over time in railway equipment

maintenance costs and Figure 5 shows the changes in

capital investment. Maintenance costs have not changed

much since 1988 after the privatization of Japanese National

2012 2.8 Zettabytes

Smartphones

Sensors

E-money

Locations

POS Systems

Emails

Tweets

Videos

Photos

Tablets

202040 Zettabytes

90% to 95% 5% to 10%

Temporarydata

(discarded)

Permanent stored data

5247 Gigabytes per person

Figure 3 Future Data Volumes



0

50

100

150

200

250

300

350

400

450

Inve

stm

ent c

ost

Financial year

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

2007

2009

2011

2013

System change

Safety Renewal

(billon yen)

Others

Lifestyle services

Station services

Better transport

Figure 5 Investment Costs

0

200

400

600

800

1000

1200

1400

1600

1800

2000

Cos

t

Financial year

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

2007

2009

2011

2013

Operating cost

Maintenance cost

(billon yen)

Figure 4 Maintenance Costs


Railways (JNR), making up between 25% and 33% of overall

railway operating costs. Overall railway capital investment

has been around ¥200 billion since privatization, but has

been increasing in recent years. Capital investment declined

in 2011 as a result of the Great East Japan Earthquake

and tsunami. The amount for replacement of dilapidated

infrastructure-related equipment is included in the item

for safety and renewal. However, new maintenance-free

equipment in the future is being introduced at equipment

updates, so those expenses are included in the system

change item. The low-maintenance track shown in Figure 2

is an example.

However, in current society with a decreasing population,

there is no guarantee that today’s expense structure and

capital investment amounts can be maintained. With the

anticipated fall in the number of passengers (and resultant

decrease in railway revenues) coupled with increases in

personnel costs as a result of a workforce shortage, the time

to develop a new maintenance structure is coming.

Age Composition of PersonnelJR East had a total of 59,240 employees at 1 April 2014.

Of these people, approximately 20% (11,530) are involved

directly in maintenance of rolling stock and wayside

equipment. Figure 6 shows the age composition for

all personnel. Veteran experts age 50 or older who bear

a central role in railway operation make up 43% of all

employees, and they will be retiring within 10 years. Almost

the same trend can be seen in maintenance, with few people

in the mid-level segment in their 40s, making passing-on

skills to new hires a pressing issue.

Smart Maintenance Initiative

The Research and Development Center of JR East

Group has proposed the Smart Maintenance Initiative to

create innovation in future railway maintenance from the

perspectives of major changes in Japan’s social environment

with falling population, rapid progress in ICT, and the

management environment. The Smart Maintenance Initiative

is not a tangible approach to mechanizing maintenance

work and enhancing equipment. Instead, it is an intangible

approach of proposing a new structure for maintenance.

The Smart Maintenance Initiative is made up of the four

challenges shown in Figure 7: achieving condition based

maintenance (CBM), introduction of asset management,

work support by AI, and database integration. The following

explains these four challenges.

Age: –19 .... 1%

Age: 20–24 .... 7%

Age: 25–29 .... 14%

Age: 30–34 .... 14%

Age: 35–39 .... 13%

Age: 40–44 .... 6%Age: 45–49 .... 2%

Age: 50–54 .... 20%

Age: 55– .... 23%

Figure 6 Age Composition of Personnel



Achieving Condition Based Maintenance

Introduction of Asset Management

Work Supported by Artificial Intelligence

Introduction of Integrated Database

Decision-making based on data analysis and prediction

Carrying out 'smart' repairwork based on

previous results

Monitoring with commuter train

Repair at best timing

Visualization costs and level of service

Strategic budget planning

Condition (risk)

Past

Future

Risk Cost

Cost

Para

met

ers

Time

Limit level

Independent DB

Linked DB

A turnout is not switched!

Check point?

You should check ………

Knowledge Experience

Rules Inspected Data History Environment -----

Clarify relation between risk and cost

Analysis by entire maintenance system

Suggestion for new maintenance method!

LegendM = Maintenance C = Civil Engineering E = Electrical Engineering

S = Signalling R = Rolling Stock

Gathering lot of inspecting data

Tracing after repair

Figure 7 Four Pillars of ‘Smart Maintenance Initiative’


Achieving Condition Based Maintenance (CBM)Achieving Condition Based Maintenance means changing

the basis of maintenance from Time Based Maintenance

(TBM) to Condition Based Maintenance (CBM). This

entails a major change in the maintenance philosophy. The

difference between the two is explained as follows using an

example of maintenance for tracks, a typical type of railway

equipment (Figure 8). Conventionally, in TBM, inspections

are conducted at regular cycles (once every 3 months for

narrow-gauge track) to obtain data on deformations (track

irregularity). Track repairs are conducted when a threshold

decided based on experience (for example, 23 mm) is

exceeded. That threshold is decided based on maximum

deterioration taking into consideration track irregularity (for

example, 40 mm) at which derailment could occur based on

past data and the 3-month inspection cycle. Inspections are

assumed to be performed at a set cycle, so the threshold is

uniform regardless of the track environment and a sufficient

leeway must be set.

Conversely, CBM is based on monitoring the status of

facilities and equipment rather than performing inspections

at set intervals. In the example of tracks, obtaining track

displacement data from trains in operation would allow such

displacement data to be obtained every day. Analyzing

that data would allow detailed identification of the speed at

which track deteriorates per 1 m (per track environment).

As a result, track deterioration could be predicted, allowing

repairs at the optimum timing. Decision-making based

on predictions would be a fundamental change in railway

maintenance.

Moreover, if displacement data could be obtained every

day, the effects of repairs would be instantly evident. In other

words, with CBM, the cycle shown in Figure 9 could be

followed on a daily basis. That cycle entails obtaining data,

analyzing data, making decisions, conducting repairs, and

evaluating the results (obtaining data). The more data is

accumulated, the smarter the important decision-making

process becomes. Conversely, in TBM, justification for

decision-making is prescribed by rules (internal regulations,

etc.), so there is inevitably little awareness about responding

flexibly according to budget status, etc. In fact, maintenance

criteria for track repair have not changed in about 50 years.

Measuring IrregularityBy inspection car (every 3 months)

Inspection for Track MaterialBy visual inspection (annually)

Periodic inspection Standards

Present (Time Based Maintenance)

Jan Apr July

Future (Condition Based Maintenance)

Irregularity and Track Material

Condition monitoringJan Apr July

Inspection and maintenance based on standards

Warning in real time Planned maintenance based on projection

Inspection by monitoring equipment installed on commercial train (daily)

Decision-making support system

Figure 8 Time Based Maintenance and Condition Based Maintenance



Introduction of Asset ManagementAsset management is the concept of considering

equipment as assets and efficiently managing them from

a long-term view. There are many case studies on this for

bridges, tunnels, and other civil-engineering structures

where deterioration is slow and the scale of repair work

is large. However, there are currently few examples as an

actual business practice. A possible reason is because

identifying equipment and evaluating deterioration over the

long term are difficult. If these issues could be overcome

by smart maintenance, maintenance engineers could

compare various repair methods and periods based on

the deterioration status of equipment. This would enable

optimum repair plans to be proposed over a long-term

span. Like CBM, this concept is thought to be a method of

supporting decision-making by maintenance engineers.

Work Support by AIThe purpose of work support by AI is to back-up evaluation

work currently done by maintenance engineers. Here, AI

is more than a support system using unstructured data,

such as text and images; it can be applied to discovering

new relationships beyond human experience by big-data

analysis using structured data.

The following introduces a support method for when

equipment failure occurs as a past case example using

text data. When equipment failure occurs, maintenance

engineers need to quickly identify the cause of the failure.

Experienced engineers can accurately infer the failure cause

based on past experience, the status of equipment, and

various environmental conditions to quickly recover from

the failure. However, it is difficult for engineers with little

experience to identify the cause of a failure. If we create a

Forecast of deterioration

Time

Dete

riora

tion

Data base

Data gathering

Data analysis

Decision making

Repair work

Evaluation

Cost

Min Cost

Min Total

Min Risk

Risk

- Maintenance at optimal intervals- Projected maintenance- Minimized accident risk and

maintenance cost

Smart repairs based on past repair results

Evaluating timing method of repairs based on deterioration after repairs

Evaluation of repair work Gathering various inspection data by monitoring, etc.

Cost

Figure 9 Condition Based Maintenance Cycle


database of the tacit knowledge and experiences of veteran

engineers and utilize AI-related technologies, we should be

able to infer the cause of failures.

Database IntegrationTo make CBM a reality, introduce asset management, and

achieve work support by AI, we need an environment where

data is handled freely. JR East currently has various systems

for individual divisions. These were established to optimize

divisional work, so each has a completely independent

system structure; the individual systems are not linked, and

there is no common data strategy.

To make quick management decisions and provide

better support to front-line divisions, it is important to

Figure 10 Integrated Platform for Smart Maintenance

Integrated Platform for Smart Maintenance

Social Data

Database for analysis

Support system for maintenance

Independent Systems

Enterprise service bus

This platform supports various kind of work!

Dashboard for executive

Data can be integrated while existing systems work.

Infrastructure data Signal data

Rolling stock data

Security system data

Track data

Electric power data

Operation data Failure data

SNS



Control center

Operation control

Maintenance depot

Trouble prediction

Engineer onsite

Trouble recovery

Catenary monitoring system

Track monitoring system

Figure 11 Monitoring Train: Series E235

manage the vast amount of data held by the company in

a centralized manner. In other words, we need to integrate

the separate internal systems and construct a large system

(common platform based on individual databases) that can

also link with external data (weather information and SNS

data). Figure 10 shows this concept. Using this platform

everyone from front-line organizations to management will

be able to use the same data, allowing quick and accurate

decision-making in the company.

Efforts in R&D

We are conducting various R&D to achieve smart

maintenance. Three case studies are introduced here.

Monitoring Train: Series E235Figure 11 shows Series E235 commuter rolling stock

manufactured in 2015 for Tokyo’s Yamanote Line. This is

the world’s first rolling stock to monitor its own onboard

devices as well as wayside equipment (track equipment and

power equipment) while operating. Monitoring data captured

onboard is sent to the wayside by wireless communications

and distributed to dispatch and maintenance worksites. Data

received by an individual worksite is analyzed and can be

used to plan later maintenance.

One example of monitoring data analysis is track

irregularities captured by a track monitoring system on other

rolling stock. Data for the same location is captured up to

five times. Figure 12 shows the change in track irregularities

after repairs. We can see that track subsidence after


Signal facilities

Track irregularity (left)

Track irregularity (right)

Catenary height

Figure 13 Example of Maintenance Support System

Repair

Repair

Machine tamping

Manual tamping

2013/06/26 2013/09/25 2013/12/25 2014/03/26

(mm)0

-2

-4

-6

-8

-10

-12

-14

-16

-18

-20

2013/06/26 2013/09/25 2013/12/25 2014/03/26

(mm)0

-2

-4

-6

-8

-10

-12

-14

-16

-18

-20

Figure 12 Analyzed Monitoring Data (Track Irregularity)



manual repairs occurs earlier than after mechanical repairs.

By constantly conducting such analysis, we can discover

deterioration details by location according to differences

in repair method and environment, supporting efficient

maintenance.

Platform for Big-Data AnalyticsAs much data as possible must be collected for big-data

analytics and analysis by AI. We are studying a method

to integrate inspection and specification data for wayside

equipment managed by independent systems in each

department. We first built a prototype

virtual environment and support tool

where equipment and inspection

information separated by system

are displayed on a single screen to

extract issues to solve. Figure 13 is a

case study of a chart where power

and track maintenance equipment is

inspected and the location status of

communications equipment is displayed

as a list. This study showed that location

information and time information are

impor tant in coordinating rai lway

maintenance. Some equipment does

not have location information, and we

confirmed the necessity of providing it

efficiently.

We a re c o n s id e r i ng u s e o f

Geographic Informat ion Systems

(GIS) already introduced by railways

in other countries as a method for

efficiently managing equipment location

information. Coordinates (longitude and

latitude) are necessary to use GIS. For

this reason, we need to build a database

of coordinate information for equipment

without location information. A possible

solution is to use Mobile Mapping

Systems (MMS) . Wi th MMS, the

surrounding trackside equipment can

be acquired as 3D data while running a

maintenance vehicle. Figure 14 shows

an example of an image obtained by

measurements. Although the figure

looks like a photograph, it is actually a

collection of points, and each point has

coordinate information. Using this data,

we can efficiently obtain coordinate

information about equipment.

Passing-on SkillsThe know-how of veteran engineers is very important when

conducting maintenance. However, these veterans in their

50s or older, who make up about 40% of all employees,

are approaching retirement and their know-how must be

passed-on to younger engineers. We have started studying

the possibility of using amassed maintenance data for

smoothly passing-on skills.

The first case study is a method of expressing past

information so that it can be used in decision-making work

(Figure 15). This expresses the status of track irregularities

35

32

33

32

31

28

20

21

17

20

15

14

8

19

20

19

18

4

34

32

32

27

25

22

00

21

20

20

10

22

22

17

3

2

2

2

1

1

2

1

0

1

2

2

0

6

6

5

11

10

10

7

7

3

0

1

1

1

0

0

0

0

0

0

31

29

27

25

25

23

21

21

17

14

14

13

9

18

15

15

25

24

19

15

16

13

12

9

8

5

5

4

3

7

7

6

25

23

22

20

17

23

25

24

21

21

20

19

16

17

22

22

29

22

34

32

33

32

38

34

40

39

40

39

39

22

29

26

25

21

27

28

28

27

38

34

25

23

26

25

23

21

24

23

27

14

33

29

28

26

28

23

29

28

26

25

24

18

35

35

14

7

17

17

18

20

26

23

24

23

23

19

17

23

28

28

23

13

35

32

30

28

30

26

25

20

17

14

10

17

25

24

34

25

43

39

40

37

45

43

35

31

29

28

26

23

33

34

31

17

39

37

37

36

39

38

31

34

25

25

24

21

36

32

41

31

43

45

45

43

54

51

39

37

35

35

31

29

50

45

35

21

40

35

41

43

54

53

58

60

48

46

44

46

48

42

38

31

36

34

35

35

40

35

36

34

36

36

31

33

41

40

22

14

26

25

26

26

36

25

17

16

13

12

11

10

25

21

29

40

41

38

36

40

42

42

27

32

27

29

25

26

29

27

31

48

44

44

44

39

31

29

13

11

9

7

5

2

19

12

42

53

44

44

44

44

16

49

42

42

37

37

32

32

45

41

36

56

55

55

57

53

23

42

26

26

18

23

18

15

30

29

24

44

43

37

50

47

22

33

40

39

37

34

35

30

43

41

17

38

34

27

36

33

9

37

26

23

23

21

20

16

32

31

Present

48 months ago

Location (100-m intervals)

Good

Bad

Bad

Figure 14 Point Cloud Data with Mobile Mapping System

Figure 15 Heat Map Showing Track Condition History


Input image

Head checks Corrugation Squat Other defects

Machine learning

Surface state

classifier

Figure 16 Image Recognition using Machine Learning



every 3 months for 4 years as a heat map. The horizontal

axis is the location on the line (per 100 m) and the vertical

axis shows the track status history. The darker locations

show where the deterioration status is good (the actual

colour is green) or poor (the actual colour is red). By viewing

this figure, the engineer can see the track deterioration

status and changes at a glance, so it can be used as a

reference in repair planning.

The second case study is image analysis using machine

learning, a form of AI technology (Figure 16). This is an

attempt to use images showing the rail head obtained by

track monitoring devices to discover faults based on the

experience of veteran engineers. This image processing

implements machine learning and can reflect the know-how

of veteran engineers. Even using only part of the data, rail

head defects could be discovered with a success rate of

about 98%.

Conclusion

The essence of maintenance—accurately perceiving

deterioration and deciding the optimal timing and method

for repairs—can be strongly supported by the Smart

Maintenance Initiative, making CBM a reality, introducing

asset management, and supporting work using AI.

Furthermore, integration of databases of fers a new

mechanism for making the most of data that is now ‘buried’

in the system and is not being used effectively.

I strongly believe that achieving smart maintenance will

lead to major reforms in railway maintenance.

ReferencesAtsushi Yokoyama, ‘Innovation in Railways: Platform for Maintenance Innovation Using ICT’ JR East Technical Review No. 29 (summer 2014)

Mitsunobu Takikawa, ‘Impact of ICT Trends on Future Railway Maintenance’ JR East Technical Review No. 29 (summer 2014)

Mitsunobu TakikawaDr Takikawa is a chief researcher at Technical Center, Research and Development Center of JR East. He joined JR East in 1990. He earned his doctorate in mechanical engineering from Tokyo University of Science in 2013.

Documents

Railway and ICT (Information and Communication … Japan Railway & Transport Review o 67 s Mar 2016 Railway and ICT (Information and Communication Technology) result of changing from