chicken or egg: infrastructure design vs. train performance · chicken or egg: infrastructure...
TRANSCRIPT
Felix Schmid and Daisuke Hasegawa
The University of BirminghamWith many thanks to Central Japan Railway Company
Chicken or Egg:
Infrastructure Design
vs. Train Performance
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 2
F Schmid and D Hasegawa
Overview of Presentation
• Conventional thinking about high-speed rail systems;
• Why high-speed rail is so successful in Japan;
• Visit to Japan thanks to Central Japan Railway’s exchange programme:
– I had read lots about Japan’s railways and the Shinkansen;
– I was convinced that I would learn nothing new!
• And then I understood…
• I also brought back an excellent doctoral student from Central Japan Railway Company…
• He applied a lean approach to high-speed railways.
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 3
F Schmid and D Hasegawa
Is High-Speed Rail all about Speed?
Inte
rnat
ional
Unio
n o
f R
ailw
ays,
2009
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 4
F Schmid and D Hasegawa
Agreed High-Speed Rail Principles
• Minimise journey time by operating at high
speed;
• Minimise journey time by reducing number of
stops;
• Minimise journey time reducing duration of
dwells;
• Avoid speed reduction by minimising stops;
• Optimise station and station design principles;
• Use the best technology that money can buy!
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 5
F Schmid and D Hasegawa
High Performance Infrastructure Theory
• High-speed railways require high-speed infrastructure:
– ‘normal direction’: turnouts must allow operation at line-speed;
– ‘reverse direction’: turnouts must allow operation at a speed that is as high as possible.
• Long overtaking / stopping loops give best performance.
Top of the Range European High-Speed Rail Station – Planned
Station
230 km/h diverging
230 km/h converging95 km/h
95 km/h
Start of Braking for 95 km/h-5893 m from Platform End
Train
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 6
F Schmid and D Hasegawa
Scaled Length of High-Speed Turnouts
toe - 220 m – crossover centre
toe - 440 m – crossover centre – tangent track
toe - 160 m – crossover centre – tangent track
4.5
m
95 km/h turnout near station
240 km/h turnout to access station
toe - 80 m – crossover centre
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 7
F Schmid and D Hasegawa
GB High Speed 1,© Google Earth
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 8
F Schmid and D Hasegawa
Choose the Best Kit that Money can BuyPoints on Swiss Mattstetten-Rothrist Line
‘© Charles Watson
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 9
F Schmid and D Hasegawa
Turkish High Speed Line S&C (Polatlı)
Turkish High-Speed Rail © Vossloh-Cogifer
With Thanks to Central Japan Railway
Company and ATOC
A Case Study:
The Tokyo to Osaka
Shinkansen Line
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 11
F Schmid and D Hasegawa
Tokyo - Osaka is served by Shinkansen
• Tokyo-Osaka built to standard gauge rather than 1067 mm:– Avoids any chance of enforced mixed traffic operation;
– Greater loading gauge and greater roll stability;
– Standard gauge allows higher speeds;
– Electric multiple units, known as ‘bullet’ trains, from day one;
– Tokyo-Osaka commenced service in 1964, at 200 km/h;
– Accident free thanks to traffic separation and robust train control;
– Super Express (Nozomi), Express (Hikari), and ‘slow’ (Kodama).
• Today, Shinkansen services operate throughout Japan:– Network is over 2400 km long, more being built to remote areas;
– Covers most of Honshu and Kyushu;
– Links into other services and islands;
– Operates fastest station to station scheduled service in the world:• Hiroshima to Kokura, 192 km in 44 minutes (261.8 km/h);
– Frequency of service high, with as little as 2 minute headway;
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 12
F Schmid and D Hasegawa
Distribution of 126 M Japanese People
• Japan as a whole: 380,000 km2, 126m people
• Honshu Island: 230,897 km2, 105m inhabitants:– Kanto Plain with (Sendai 0.7 M) Tokyo 12 M, Yokohama 3.5 M;
– Nagoya 2 M, Kyoto 1.5 m Osaka 2.7m, Kobe 1.4 m, Hiroshima 1m.
– In total 20m people in city centres.
• Kyushu Island: 35,640 km2, 12m inhabitants:– Fukuoka 1.4m Nagasaki 0.5m;
– Tunnel link to Honshu , the ‘mainland’.
• Shikoku Island: 18,800 km2, 4.1m inhabitants: – Matsuyama 512,000;
– 9.4 km Seto Ohashi bridge system links Shikoku to Honshu.
• Hokkaido Island: 83,452 km2, 5m inhabitants:– Sapporo is the main city, link to Honshu through a sub-sea tunnel.
• Main corridor of Japanese commercial activity:– Japan: Tokyo to Nagasaki , corridor ca. 900 km long by 20-40 km wide.
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 13
F Schmid and D Hasegawa
Tokyo to Osaka, an Economic Corridor
• Land area: 23.7% of total land area of Japan;
• Population: 59.7% of total population of Japan;
• Share of Japan’s GDP: 64%. Source, CJR Data Book 2012
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 14
F Schmid and D Hasegawa
300km
Radius
Busan
Kita Kiyushu
Higoshima
Aomori
Fukushima
Sendai
Nagoya
Sapporo
Tokyo
Osaka
‘Japan by Night’ Photograph
© Dundee University
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 15
F Schmid and D Hasegawa
Views from Hotel in Shin Yokohama
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 16
F Schmid and D Hasegawa
The Japanese Choice
toe - 220 m – crossover centre
toe - 440 m – crossover centre – tangent track
230 km/h?
toe - 115 m – crossover centre – tangent track
70 km/h
toe - 160 m – crossover centre – tangent track
4.5
m95 km/h
230 km/h
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 17
F Schmid and D Hasegawa
Keep it Simple Infrastructure PracticeTop of the Range European High-Speed Rail Station – Planned
Typical Shinkansen Station (Shin Yokohama)
Station
230 km/h diverging
230 km/h converging95 km/h
95 km/h
Start of Braking for 95 km/h-5893 m from Platform End
Train
70 km/h
70 km/h
Start of Braking for 70 km/h-4789 m from Platform End
Train
Station
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 18
F Schmid and D Hasegawa
Train Motions and Following Trains
0
50
100
150
200
250
0
10
20
30
40
50
60
70
80
90
100
-6000 -5000 -4000 -3000 -2000 -1000 0
Braking Curve Europe
Braking Curve Japan
Elapsed Time Europe
Elapsed Time Japan
Following Train E
Following Train J
Station
PlatformTrain
BCE BCJ
ETJETE
FTJ
FTE
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 19
F Schmid and D Hasegawa
Turnout Clearing Behaviours
0
5
10
15
20
25
30
35
40
02004006008001000
Europe
Japan
Distance from Stopping Point [m]
Spe
ed o
f H
ead
of
Trai
n [
m/s
²]
Turnout Clearing Distance
Turnout Clearing Distance
Station Platform
Train
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 20
F Schmid and D Hasegawa
Japanese Approach to Railway Design
1. Identify demand:
– Quantity, journey time, competitors.
2. Design Rolling Stock:
– Traction power;
– Powered axles.
3. Design Infrastructure:
– Gradients, curves;
– Stations.
4. Define Timetable:
– Number of trains;
– Service pattern.
ROLLING
STOCK
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 21
F Schmid and D Hasegawa
Central Japan Railway: N700 Series
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 22
F Schmid and D Hasegawa
Highly Powered Rolling Stock
• Design best possible rolling stock:– Build it light: Shinkansen N700 weighs 730 t;
– Put in lots of seats: Shinkansen N700 offers 1400;
– Shinkansen N700 has 56 out of 64 axles powered, with motors rated at 300 kW / each – 17.1 MW of power:
• N700 reaches 300 km/ h over distance of 11 km;
• Allows all-electric braking down to ~15-20 km/h.
• Design rolling stock that copes with the job in hand:– Build it light: Pendolino (9 car) weighs 460 t;
– Put in a fair number of seats: Pendolino offers 480;
– Pendolino (9 car) has 14 out of 36 axles powered, with motors rated at ~360 kW each – 5.1 MW of power:
• Pendolino reaches 200 km/ over distance of 11 km;
• Requires blended braking at high speed.
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 23
F Schmid and D Hasegawa
Simplified Infrastructure
• Short stopping / dwelling tracks at stations;
• Short turnouts / straightforward S&C;
• Steep gradients (2%) / tight curves;
• High integrity railway control system:
– Unidirectional signalling only;
– Driver controlled with ATP on plain line sections;
– Automatic braking on approach to stations (ATC).
• Ballasted track to cope with seismic activity;
• Substantial maintenance window each night.
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 24
F Schmid and D Hasegawa
Mishima Station
at the Foot of
Mount Fuji...
Short S&C and Tight
Curves in Simple
Ballasted Track
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 25
F Schmid and D Hasegawa
Tokaido Shinkansen: Generalised Cost
• Fare is affordable compared to road transport;
• High speed results in short journey time;
• High proportion of comfortable seated travel:
– Company ensures spare capacity on most trains.
• Stations in all major towns and cities;
• Good integration with local public transport;
• High service frequency minimises waiting time;
• Average delay of 0.6 min per train results in perception of very high reliability.
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 26
F Schmid and D Hasegawa
Tokyo to Shin Osaka (Tokaido Services)
Type Journey Time Frequenc
y
Intermediate
Stops
Fare
(Yen)
Fare (£)
Nozomi 2:30 7 per hour 4 14,050¥ £117.00
Hikari 2:57 – 3:10 2 per hour 7 – 9 13,750¥ £114.60
Kodama 3:54 – 4:00 2 per hour 15 13,750¥ £114.60
Road Toll 6:11 n.a. n.a. 9,100¥ £80.00
Maglev Line
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 27
F Schmid and D Hasegawa
Benefits of Japanese Approach
• High ratio of motored axles increases acceleration and braking rates – latest trains are 100% powered;
• Large size traction machines allow purely electro-dynamic braking (270 km/h to 30 km/h);
• Low speed (70 km/h) turnouts have multiple benefits:– No need for swing nose crossings / moving frogs;
– Low complexity and cheaper to buy and renew;
– Reduced land take and lower track forces;
– Replacement in one 5 hour maintenance window.
• Unidirectional operation is simple and safe;
• Low technology maintenance is flexible.
A ‘Lean’ Approach to
High Speed Rail
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 29
F Schmid and D Hasegawa
WizzAir – The Ultimate Lean Approach
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 30
F Schmid and D Hasegawa
Value Stream Map: Passenger Journey
London to Birmingham
10 min
Walk
10 min
Queue
Buy a
ticket
10 min
Walk to gate
Ticket check
Walk to a
platform
60 min
Board train
Ride train
Alight from
train
5 min
Walk to
gate
Ticket
check
10 min
Walk
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 31
F Schmid and D Hasegawa
Value Adding Activities in H-S Rail
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 32
F Schmid and D Hasegawa
Value Stream Management in H-S Rail
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 33
F Schmid and D Hasegawa
Value Stream Mapping: Single Train
VANVA
Move
Alight Board
Move
Terminus
Terminus
Train path
VA: Value adding step Necessary activity
NVA: Non Value adding step Unnecessary activity
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 34
F Schmid and D Hasegawa
VSM: Turnaround at Terminus
VA
NVA
Move
Alight Board
Move
Terminus
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 35
F Schmid and D Hasegawa
VSM: Cumulative Negative Effect
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 36
F Schmid and D Hasegawa
VSM: Original vs. Lean Operation
VA
NVA
Terminus
Original
Lean
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 37
F Schmid and D Hasegawa
VSM: Reduced Negative Effect
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 38
F Schmid and D Hasegawa
VSM+JIT Running Time Optimization
JIT
How do we benefit from slower running?
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 39
F Schmid and D Hasegawa
Maximum Velocity HS2: 200-360 km/h
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 40
F Schmid and D Hasegawa
JIT: Journey Time and Energy Use
+ 3 min
- 11.7 %
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 41
F Schmid and D Hasegawa
Train Turnaround Activities
Alighting Boarding
Hand over Checking
Toilet cleaning
Wiping tables
Collecting rubbish Internal
Arrival Departure
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 42
F Schmid and D Hasegawa
SMED: Train Turnaround Activities
Alighting Boarding
Hand over Checking
Toilet cleaning
Wiping tables
Arrival Departure
Collecting rubbish
External
Internal
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 43
F Schmid and D Hasegawa
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 44
F Schmid and D Hasegawa
Britain might be better than Japan!
Voith TESSEI*
VehicleImaginary
Class 390 (18 car)E5 + E6 (17 car)
Train length (m) 434 401.7
Number of seats 916 1069
Turnaround time (min) 12 12
Cleaning time (min) 7 7
Coach cleaners 6 35
Toilet cleaners 3 9
On board cleaners 2 0
4411
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 45
F Schmid and D Hasegawa
Non-Optimised Terminus Operation
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 46
F Schmid and D Hasegawa
SMED+JIT+VSM at a Terminus
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 47
F Schmid and D Hasegawa
Operations Stability: Taguchi Analysis
• Taguchi Analysis:
– Variation of multiple factors;
– Identification of highly influencing factors;
– Minimisation of numbers of experiments.
• Factors used:
– 4, 5 or 6 station tracks;
– 4 station track layouts;
– 8, 7 or 6 min turnrounds;
– 40, 30 or 20 s signalling system headways.
• Signals, greatest influence.
1 1 1 2 3 4
Arrival 2 Arrival 2 1 × × ×
4 platform tracks Layout a 2 × × ×
3 3 3 〇〇 ×
Departure 4 Departure 4 4 〇〇 ×
1 1 1 2 3 4
2 2 1 × × ×
5 platform tracks Layout b 2 〇 × ×
3 3 3 〇〇 ×
4 4 4 〇〇 ×
5 1 1 2 3 4
2 1 × × ×
1 Layout c 2 〇 × ×
3 3 〇〇 ×
2 4 4 〇〇〇
3
6 platform tracks 1
4 1 2 3 4 5 6
5 2 1 × × × × ×
3 2 〇 × × × ×
6 Layout c 3 〇 × × × ×
4 4 〇〇〇 × ×
5 5 〇〇〇 × ×
6 〇〇〇〇〇
6
AD
D A
AD
AD
Chicken or Egg: Infrastructure Design vs. Train Performance
RAILcph Danis Railway Conference, 15 May 2017Slide No: 48
F Schmid and D Hasegawa
Summary and Conclusion
• Focus on simplicity rather than complexity;
• Improved control of resources and performance causes a leap in process predictability;
• This allows focused decision making for:
– More reliable outcomes for customers,
– More targeted allocation of operational resources; and
– More motivated process owners through achievable targets and measurable success criteria.
• This change can not be ‘acquired’ by just buying tools but also demands:
– Change in attitudes at all levels of management towards continued improvement of processes.