Driving Efficiency, Preserving Infrastructure and Creating Capacity With Long Trains S.Bell

Canadian Pacific, Calgary, Canada

Summary: Building on the distributed power work presented to the IHHA in 2009 and 2011,

Canadian Pacific (CP) has continued to systematically introduce long trains that are driving efficiency, preserving infrastructure as well as creating capacity. CP extensively leverages multiple-remote distributed power trains, where remote locomotives are placed in up to three locations, which permits much longer and heavier trains. This multiple-remote capability has become the cornerstone of Cp's long train strategy, resulting in transcontinental intermodal trains up to 4300m (14,000 feet), 170-car, 22,700 tonne potash trains and most recently, the introduction of 152-car coal trains. These trains are not only more productive; longer, heavier and more fuel efficient, but are also less destructive in terms of reduced lateral forces. CP has a total of 8 lateral force detectors on the network, which are being leveraged to confidently introduce more productive and less destructive trains. An example that will be featured in the paper is the results from a detector recently installed in the coal route, near the site of two separate coal train derailments over the last 5-years. Analyses and simulations have revealed this area is subjected to high lateral forces from repetitive run-in behaviour, exacerbated by the relatively uniform train sizes and similar train braking patterns.

The paper will examine the forces imparted by different train

makeup and locomotive configurations, including the recently implemented 152-car coal sets, as well as ECP (Electronically Controlled Pneumatic Brakes) equipped coal sets, and will build on this to demonstrate how to drive systematic improvement to train/track interaction and improve safety. In summary, the paper will update the IHHA on the recent long train progress made at CP and how the continued introduct�on of long trains is driving efficiency, preserving infrastructure as well as creating capacity. Index Terms: Distributed power, multiple remote controlled locomotives, long trains, ECP

The objective of this paper is to update the IHHA on the

1.

continuing distributed power work at CP that has led to

INTROD UCTION

the introduction and operation of more productive, less

Canadian PacifIC is North America's 6th largest railway,

destructive and safer trains. The paper will discuss how

operating over 25,800 route-kms from Vancouver to

distributed locomotive power (DP) placement on CP is

Montreal in Canada, with US operations in the northeast

both an enabler and neutralizer for the impact of long,

through mid-west, including Chicago, Minneapolis and

heavy trains, operated with AC power on challenging

Kansas City.

geography with a tight horsepower/ton ratio (HP/T).

789

1 0th International Heavy Haul Association Conference, 2013

2.

through a train, to ascend the much tougher 2.2% eastward

�T COmuDOR OVER�EW

grades.

CP's West Corridor continues to be the focus and testing ground for DP train model advancements. The

subdivisions on CPo

West Corridor is the geographic territory from Calgary, Alberta to Vancouver, British Columbia (BC).

From a maintenance and capital perspective,

the Mountain Sub is roughly triple the cost/km of other

Horsepower per ton (HPIT) is typically around 0.9

There

are almost 700 kms (400 miles) of mountain territory in

for loaded bull," trains heading west, while the tougher

the West Corridor, involving three mountain ranges; the

eastward grades require almost 2.0 HPIT at Albert Canyon

Rockies, Columbia and Selkirk mountains, requiring the

through Field Hill. AC 4400 locomotives are, by far, the

addition of a pusher locomotive to the headend of trains

dominant locomotives in the West Corridor, which has the

transiting the corridor, both eastward and westward.

highest train density on the network, with about 40% of

At Golden, BC is where the coal route meets the

CP's total business transiting the corridor.

transcontinental mainline and train density increases by

Throughput consistency is critical as there is limited

about 10 more trains/day west to Vancouver. Corridor

capacity to handle any surges or backlogs, particularly with

capacity between Golden and Vancouver is 36-38 trains/

the added complexity of coordinating train movements

day.

with West Coast ports.

The three mountain ranges present steep grades and stall

Finally, sidings lengths are typically 2134m (7,000 feet),

risks. Eastward, there is a tough 2.2% ascending grade

network-wide, however, we have demonstrated there is

at three separate locations on the Mountain and Laggan Subdivisions. \Vestward, a 1% ruling grade challenges

sufficient long siding infrastructure to operate some over­ length trains, bi-directionally, with precision planning and

loaded bulk trains up to 22,730 tonne (25,000 ton) on the

execution to leverage the long train meet locations.

3 major ascending grades.

3.

The route is predominantly single track with almost 50%

DRIVING EFFICIENCY, PRESER �G

of the routing traversing curves tighter than 3493m radius

INFRASTR UCT URE AND CREATING

(1/2 degree) and 1 33 km (80 miles) of curves tighter than

CAPACITY WITH LONG TRAINS

290m R ( > 6 degrees). Maximum curvature is 150m

The

radius (>11 degrees). Sixty percent of the trains are unit

journey

that

began

in

2007

to

improve

our

understanding of train/track dynamics has led to the ability

bulks (carrying coal, potash, grain and sulphur), as heavy

to confidently introduce more productive, less destructive

as 22,700 tonnes (25,000tons), powered by up to 5 AC

and safer trains on CPR, network-wide.

traction 4400HP diesel-electric locomotives, distributed in

Below is a table that illustrates the train models that we

as many as 3 locations through the train.

have leveraged DP to significantly increase productivity,

CP continues to power eastward intermodal trains, with

from 2007 to present;

up to 5 AC 4400's, distributed in as many as 4 locations

Long Train Model Evolution: 2007 2012 •

Intermodal:

2012

2007

Coal

3

locos (2-0-1) 124 cars 6,800 ft 17,500 tons 0.75 hp/t

4

locos (2-1-1) 152 cars 8328 ft 21,400 tons 0.82 hp/t

Potash

4

locos (2-2-0) 124 cars 6,120 ft 17,900 tons 0.99 hp/t

5

locos

(2-2-1)

cars 8,350 ft 24,500 tons 0.90 hp/t

3

Westwards

2

locos (2-0-1) 380 TEU's 7,000 ft 6,300 tons 2.10 hp/t locos (2-0-0) TEU's 7,000 ft 6,300 tons 1.40 hp/t

380

170

Figure 1: Evolution of Long Train Models at CP: 2007 - 2012

790

2012

2007

Eastwards

5

locos (2-1-1-1) 760 TEU's 14,000 ft 12,000 tons 1.83 hp/t 3

locos (2-1-0) TEU's 9,500 ft 8,600 tons 1.53 hp/t

515

10th International Heavy Haul Association Conference, 2013

Clearly, potash, coal and intermodal trains are significantly

% Change Baseline vs Current. by Train Series

more productive, and given not all trains can be over-length in our largely single track network, the next question is are

40%

they less destructive?

20%

Below is a chart which illustrates the low rail forces

0%

of various westward train models, their train sizes and locomotive configurations, from Fall 2007 to Fall 2008, a period that coincided with the implementation of our

�40%

I a Waighl .. Length C Low R"il I

first lateral force detector, just west of Revelstoke, on a 1% ascending grade, 6-degree curve as well as just

Figure 4: Comparison of Productivity and Low Rail Lateral

prior to the introduction of top-of-rail lubrication in late

Force Changes

2008.

Westward intermodal trains are 40% longer and heavier, but low rail forces dropped almost 40% with the 2-1-0

Baseline low Rail Force at Revelstoke Fall 2007 • Fall 2008

locomotive configuration. Grain trains were unchanged at 1 1 4 cars, and at 7,000 feet are essentially a "design staple"

14

for the high density of 7,000 feet sidings in the corridor, however, of note, the low rail forces dropped with the I P

10

change to mid-train remote instead of tail-end. Manifest

8

trains are a bit biggel� with the same conventional configuration but experienced a drop in forces, likely due to top-of-rail lubrication. 11M )·11·1) x

(;1.111 ).1). Ix

MM ,.1).1) x

7k

114

5k

�ul"ltul ,.1). Pnld\1t ,.,. 1 x 115

IIMI ).(1.

Ox124

with the 1-1-1 locomotive configuration. P otash and coal

lx124

both experienced significant productivity lifts of almost

Figure 2: Low Rail Lateral Forces - Westward TI'ains:

40% and 25% respectively, while low rail forces dropped

2007 - 2008

almost 1 5% with the multiple-remote footprint of these

Loaded bulk train forces all approached,

two trains.

and even

It is worth mentioning as well that the 1 70-car potash

exceeded 10 kips, the threshold at which Engineering

train, discussed in a 2011 paper and presentation to the

get concerned with the cumulative wear and fatigue to

IHHA in Calgary [2], has become the dominant potash

infrastructure from these sustained force levels. Figure

3

plots

current

low

rail

Sulphur trains too were

unchanged, size wise, however, low rail forces dropped

forces,

train on CP, with over 300 1 70-car trains operating over

locomotive

the last 12-months.

configuration and train size and gives

In terms of average train lengths and how long trains have contributed to preserving train capacity, below is an illustration of train lengths through Revelstoke, BC on the

Current low Rail Force at Revelstoke Jan/Zoll- July/12

western mainline.

14

Average Train Length

12

K

P

-

R evelstoke

------. 6800 ,---.--10 8

e e

G

t

-

I/M 1

1 (I x (;1.iIl2 1 (l x

9k

114

MM 1 0 (I X

5ulpltm 1 1

Pola;1t 21

(oalll

5.5k

1 x 115

lx170

lx152

6600

i------

6400

7------

6200 6000 2005

Figure 3: Low Rail Lateral Forces: IVestward Trains:

2006

2007

2008

2009·2012

Figure 5: Average Train Length at Revelstoke,

Janl2011 - July/12

BC: 2005 - 2012

As can be seen in Figure 5, average train lengths have been

clear evidence of how the current menu of trains are not

relatively steady through Revelstoke until the long train

only more productive but less destructive.

breakthrough in 2009 and have increased an average of

In fact, the Figure 4 summary chart illustrates the % change

400 feet, which on a train population of 1 1,000 trains per

in Train Weight, Train Length and Low Rail Forces;

year has resulted in the equivalent of 2 more trains per day.

791

1 0th International Heavy Haul Association Conference, 2013

To further demonstrate this point, the number of trains

Considering the menu and frequency of relatively uniform

through Revelstoke has actually remained steady at about

2134 m (7,000-foot) loaded bulle trains; approximately

30 trains per day ... so, long trains are more productive, less

1 600 export coal as well as 250 potash and 300 grain trains

destructive and have preserved capacity in this key corridor.

destined to the UP, the infrastructure was being stressed by sustained run-in events through this location, illustrated

The 1 52-coal train is the newest bulle train improvement on

immediately below in Figure 6. A number of proactive

CP, with about 1/3 of the coal sets operating at this length.

steps were undertaken to both strengthen track as well as

Some of the pre-work on the 1 52-car model focused on

improve our understanding of the forces in this area.

simulations descending through mile 29 to 30 on the

Given our success in leveraging data from lateral force

Cranbrook Sub-division, down near the mines in south­

detectors, we installed the first commercial application of

eastern BC, approaching Fernie, which was the site of a loaded coal train derailme�t in March, 2011 ... the second

an LB Foster Wireless LN Module at MP 29.9 to allow us to track the behavior of bulle trains, including the 1 52-car

loaded coal train in 5-years to derail near this location.

coal, once introduced.

Simulations revealed repetitive run-in events as the various portions of these relatively uniform bulk trains descended

The LN detector' went into service in early 2012, on the

the staggered grade, with similar train braking patterns

high rail of the 1 .2% descending grade on the 6-degree

The in­

curve. To-date, fOr'ces fr'om approximately 800 westwar'd

train forces manifested in repetitive run-ins, appr'oaching

trains have been captured, of which 1 60 were 1 52-car coal.

356kN (80 kips), and tended to concentrate in particular,

The high r'ail fOr'ces, averaging 8.14 kips, for these 1 52-car

in the portion of the train approaching Mile 30, near' the

coal trains, configured 1 -1 -1 are illustI'ated in Figure 7;

and operated over the flat spot at Mile 29.2.

290m R (6-degree) reverse curves.

------����-- 40------�/ 2

o

.2 28.89 STEPHENSON Rd

0&011t�

� ;fqj .'"

11.20

STATION 9011 CAPACITY 8559fl Mile 28

'SMER 'lEEK

Figure 6: Illustration of Curves and Grades on Cranbrook Subdivision

152... car coal on HR at Fernie: Feb ... July/12 14 12 K I 10 P

8

s

6 4 Figure 7: High Rail Forces of 1-1-1 x 152 coal at MP 29.9 Cranbrook Sub

792

10th International Heavy Haul Association Conference, 2013

There is actually a significant range of forces within the

we were able to identiJY the ECP sets by virtue of their

distribution, from 5 kips through 13 kips, which suggests

dedicated locomotive numbers and permit a comparison of

there may well be optimal train handling strategies for

126-car coal, non-ECP vs ECP.

Locomotive Engineers to follow through this location to

First, the 126 non-ECP high rail forces are show in Figure

minimize forces and undesirable in-train behavior. It is an

8. This population of 275 non-ECP trains actually had

area for further research on CP to zero-in on via detailed

higher average forces, at 8.41 kips, than the larger 152-

examination of train handling by leveraging locomotive

car model, and again, another fairly wide distribution from

download information.

5 kips through 13, but have two potential explanations

For context, and as validation of the new LN device,

for the higher forces and variation. In addition to train

we compared the high rail descending forces from Mile

handling, lilce the 152-car observations, the non-ECP

29.9 to those from a traditional strain gauge LN detector

trains operate with only 2-locomotives and therefore

at Albert Canyon on the Mountain Sub, for loaded coal

single DP placement of either mid or tail, We do not have

trains descending the 2.2% grade in a 2-1-1 configuration

an AEI feed and match process at this site, but do know

and found that population had an average force of 29.5

there is a mix of both mid and tail operated and expect

kN (6.63 kips), at an average speed of 32 kmlh (20 mph)

that tail-end only trains account for higher average force

vs Fernie's 48 km/h (30 mph) ... so forces are considered

levels than mid configured trains. Once again, an area

relevant and accurate at Mile 29.9 Cranbrook Sub.

for additional research for us, however, we do expect the aluminum coal fleet will move

Another common coal train size is 126-cars, both ECP (Electronically Controlled Pneumatic Brakes) and non­

as well as better understanding ECP forces, which is the

ECP. Only two of our sets are ECP equipped, however,

next chart in Figure 9.

126-car coal at Fernie - Non-ECP Feb - July/12 14 12 10 8 6 4 Figure 8: 126-car coal non-ECP at MP 29.9 Cranbl'ook Sub

ECP Coal 1-1-1 X 126 Feb July/12 -

14 12 10

.

-

8 6 4

I

to 152-cars this year, so

research will focus more on train handling opportunities

I

I

I

Figure 9: 1-1-1 x 126 ECP Coal at MP 29.9 Cranbrook Sub

793

10th International Heavy Haul Association Conference, 2013

behavior from these uniform size trains, occupying the same

The ECP sample produces very interesting results: a very

kN (7.38

geographic footprint, where double-mid forces would be

kips), or 12% lower than the 126 non-ECP trains above,

particularly vulnerable to Emding a weak spot in the track,

Again, this lower force behavior is consistent with results

or cumulatively creating weak spots for subsequent trains.

tight distribution, with average forces of 32.8

separately collected at the Albert Canyon LN site and

For example, a 1-2-0 coal train on the Cranbrook Sub

attributed to the increased train handling control provided

developed severe forces, particularly in dynamic braking,

to locomotive engineers with the graduated release feature

at even relatively benign descending locations, where the

of ECP and how this permits trains to better navigate non­

double-mids are not only pushing back and holding the

uniform descending conditions. Much

as

we

consider

portion of train behind them, but also pulling the portion

multiple-remote

placement

of train ahead of the double-mids.

a

So, the double-mids

are essentially winning a "tug-of war" with the single

"neutralizer" for undesirable forces in ascending situations, we believe ECP is showing strong evidence of being a

lead engine by both pushing backwards and pulling

"geographic neutralizer" for the significant grades and

backwards. The result is severe swings between draft and

curves in Cp's west corridor in descending situations. We

buff, in relatively close proximity within the train to risk

originally anticipated, at the time of abstract submission

separation, and producing run-in events that approach

to the IHHA, having both ECP sets stretched to 152-cars,

and can exceed safe limits.

however, that did not occur. We did, however, operate

Furthermore, it would also be very challenging for

two 152-car ECP trains and the results, though hardly

the

statistically significant, produced forces at the LN site of

engineering

department

to

adequately

reinforce

fastenings, at not just the chronic, repetitive unavoidable

15.6 kN (3.5 kips) and 31 kN (7 kips), both promising and

spots, but even relatively "benign" locations will require

intriguing, and we still expect to lengthen both ECP sets

strengthening to withstand the extreme force swings of

to 152 to enable a more robust examination of 152 ECP vs

double-mids on loaded bulk trains.

non-ECP, and what fleet implications that may produce as

The conclusion from these double-mid simulations was

a result.

that train size on CP had reached a point where multiple­

Another area worth discussing is a robust process we

remote placement was essential to effectively manage and

recently undertook to re-examine our multiple-remote

mitigate the impact of long, heavy trains, controlled by

configurations and specifically considered simulations of

AC power that are capable of high tractive and retarding

the operationally easier to assemble double-mid placement

forces, operated under tight horsepower per ton, over very

for; 152 coal, 142 potash as well as 3048m (lO,OOO-foot)

challenging geography and within the limits of the existing

eastward intermodal trains ... so 2-2-0 instead of 2-1-1

fastening system.

trains. CP used the same tool, NYAB Train Dynamics Analyzer,

4.

widely used in the industry, that initially revealed we could

Building on the 2011 work shared with the IHHA in

safely operate the present multiple-remote configurations

Calgary [2] that examined curve rail replaced and re­

of longer, heavier trains, which produced simulated forces well within industry accepted tolerances of: Draft kN (300k), Buff and LN Ratio

<