Articulated Double-Stack Car Effects on Bridges

Anna M. Rakoczy, Ph.D., Principal Investigator Duane Otter, Ph.D., P.E., Scientist
Transportation Technology Center, Inc. Transportation Technology Center, Inc.
55500 DOT Rd 55500 DOT Rd
Pueblo, CO 81001 Pueblo, CO 81001
Telephone: 719-584-0782 Telephone: 719-584-0594
[email protected] [email protected]

David Linkowski; Engineer Stephen Dick, PhD, SE; Senior Research Engineer
Transportation Technology Center, Inc. Purdue University, Bowen Laboratory
55500 DOT Rd 1040 S. River Road
Pueblo, CO 81001 West Lafayette, Indiana 47905
Telephone: 719-584-0634 Telephone: 816-401-5605
[email protected] [email protected]

Word Count: 2,430 + 17*250 = 6,680

ABSTRACT
Transportation Technology Center, Inc.’s (TTCI) research on the effects of articulated double-stack cars on
bridges indicates that they generate larger loads on short bridge spans and floor system members as
compared to typical bulk commodity cars. Most articulated double-stack cars have a nominal maximum
truck weight of 157,500 lbs., which is higher than the nominal truck weight of 143,000 lbs. for a four-axle
bulk commodity car. An analytical investigation indicates that double-stack cars cause larger maximum
moments on spans shorter than approximately 15 feet. To confirm the analytical results, datasets were
collected from four Wheel Impact Load Detectors (WILD) located on different intermodal corridors in North
America. In addition, strain gage data was collected from a short bridge on a line carrying both intermodal
and coal trains.
1. INTRODUCTION
With recent increases in double-stack intermodal traffic on a number of lines, railroad bridge engineers
requested a study on the effects of those double-stack cars on bridges (1, 2). This study focused on
articulated double-stack cars, with nominal truck loads on the intermediate trucks that exceed the truck
load on typical four-axle freight cars. The nominal truck load for articulated double-stack cars is 157.5 kips
— about a 10 percent increase compared to the 286-kip freight cars with truckloads of 143 kips.
This paper summarizes an investigation of net truck vertical forces produced by articulated double-stack
cars in revenue service and calculations of internal forces for various bridge span lengths. Results
indicate that some short bridge spans may experience a load effect close to the current E-80 design load.
In addition, a summary of test results of a revenue service bridge under intermodal trains and coal trains
is presented. The measurements include strains from all six beams and deflections from the center
beams of each side. The analysis focuses on comparison of peak stress and stress range cycles.

2. ARTICULATED DOUBLE-STACK CARS

A well car, also known as a double-stack car or stack car, is a type of railroad car specially designed to carry
intermodal containers (shipping containers) used in intermodal freight transport (3). The "well" is a depressed
section which sits close to the rails between the wheel trucks of the car, allowing a container to be carried lower
than on a traditional flatcar. This makes it possible to carry a stack of two containers per unit on railway lines
(double-stack rail transport) wherever the loading gauge assures sufficient clearance. Each unit of a double-
stack car is constructed with a single well, but are often constructed with multiple units of three to five units,
connected by articulated connectors (3). Articulated connectors (sharing wheels between the car’s units) are
supported by the centerplate of a single truck, which reduces slack action and improves the ride quality for fragile
cargo.
Intermodal double-stack cars come in different configurations. As illustrated in Figure 1, common cars are five-
unit, articulated railcars with platforms for carrying 40-foot international containers; and three-unit, articulated
railcars for transporting 48-foot and 53-foot domestic containers (3). The nominal weights and car dimensions
for five-unit articulated double-stack cars are presented in Table 1.

FIGURE 1. Dimensions of Five-Platform Intermodal Cars

TABLE 1. Dimensions and Weights of Articulated Cars (3)
Weight Length dimensions Truck Wheelbase
Number of (lbs.) (ft.) (ft.)
Railcar Type
Platform Total Truck Centers
Loaded Loaded
A D E F
Axles Trucks B C
TOFC Spine Car,
12 6 97,000 485,000 263.75 51.50 47.83 6.79 5.50 5.50
Five Platforms
International Stack,
8 4 157,500 482,000 165.42 50.00 50.00 6.79 6.00 5.67
Three Platforms
International Stack,
12 6 157,500 800,000 264.67 50.25 50.17 6.79 6.00 5.67
Five Platforms
Domestic Stack,
8 4 157,500 485,000 203.25 63.50 63.08 6.79 6.00 5.67
Three Platforms
Domestic Stack,
12 6 157,500 800,000 307.25 58.83 58.58 6.79 6.00 5.67
Five Platforms
Note: (A) - Overall Length, (B) - Truck Centers Interior, (C) - Truck Centers End, (D) - Coupler Overhang, (E) - Interior Trucks, (F) -
End Trucks.

3. EQUIVALENT COOPER LOADING FOR VARIOUS BRIDGE SPAN LENGTHS

The equivalent Cooper loading is based on the design loading recommended by the American Railway
Engineering and Maintenance-of-Way Association (AREMA) (4). It is current practice to design railroad
bridges for Cooper E-80 loads, which have maximum axle loads of 80 kips. By comparison, the nominal
maximum axle load for an articulated double-stack car is 78.75 kips. Many bridges currently in service were
originally designed for lesser loads, such as E-60, but with a higher dynamic load (impact) allowance for
steam locomotives.

The results are presented only for spans up to 100 feet long, since the double-stack car effects are most
visible on shorter spans. Figure 2 shows double-stack cars have equivalent Cooper loads greater than
common 53-foot coal cars for spans up to 15 feet long. In addition, end shear is higher for double-stack
cars on spans shorter than 15 feet and the floor beam reaction is higher for spans up to 5 feet long (Figure
2).

Bending - Mid Span

80

Equivalent Cooper Loading (E)

70
60
50
40
30
20
10
0
0 10 20 30 40 50 60 70 80 90 100
Span
Endlength
Shear(ft)
80
Equivalent Cooper Loading (E)

70
60
50
40
30
20
10
0
0 10 20 30 40 50 60 70 80 90 100
Bending
Span
Floor - Mid Span
length
Beam (ft)
Reaction
80
Equivalent Cooper Loading (E)

80
Equivalent Cooper Loading (E)

7070
6060
5050
4040
3030
2020
1010
0 0
0 0 1010 2020 3030 40
40 50
50 60
60 70 80
80 90
90 100
100
Spanlength
Span length(ft)
(ft)

6-Axle Loco 53'0" Coal Car

International Stack - 3 Platforms International Stack - 5 Platforms
Domestic Stack - 3 Platforms Domestic Stack - 5 Platforms

FIGURE 2. Equivalent Cooper Loading up to 100 feet – Bending Moment, End Shear and Floor
Beam Reaction
4. WAYSIDE DATA

Wayside detectors can gather data from a large number of passing trains on different types of equipment.
Wayside detectors are currently in use at many locations throughout North America. One type of wayside
detector, a Wheel Impact Load Detector (WILD), measures the vertical forces on the rail. The maximum force,
also known as the peak force, represents a combination of the weight of the car carried by a single wheel and
the dynamic loads generated by surface imperfections. TTCI used wayside data from WILDs on tangent track
to estimate truck weight of articulated double-stack cars. This study estimates truck weight as the sum of
average vertical force of each truck wheel. The average vertical force of each truck wheel is calculated as the
peak force minus the dynamic force.
Wayside truck force data was obtained from four different sites: Bagdad, California; Gothenburg, Nebraska;
Vine Creek, Indiana, and Goodeve, Saskatchewan, Canada. Results are mainly presented for Bagdad since
that location contained the highest number of records. Data was analyzed by quarter starting with the third
quarter of 2014 through the second quarter of 2016. Figure 3 presents a frequency estimated truck weights
for the considered time periods. The end trucks of the cars have been excluded. The data is only for the
interior trucks at the articulated connections.

Frequency Histogram - Bagdad In-trucks

16% 2014_3
Percent of sample

14% 2014_4
12%
2015_1
10%
8% 2015_2
6% 2015_3
4% 2015_4
2% 2016_1
0%
2016_2
20
30
40
50
60
70
80
90
100
110
120
130
140
150
160
170
180
190
200
210

Truck weight, kips

FIGURE 3. Frequency Histogram of Net Truck Weight for Considered Time Periods
The distribution looks consistent for all time periods; therefore, the changes of the truck weight due to
seasonality can be treated as minimal. The frequency histogram of estimated truck weight calculated as an
average of all periods of time is presented in Figure 4. The coefficient of variation is about 30 percent for all
recorded data, including empty cars.
 Frequency Histogram - Bagdad In-trucks
16%
Percent of sample

14%
12%
10%
8%
6%
4%
2%
0%
20
30
40
50
60
70
80
90
100
110
120
130
140
150
160
170
180
190
200
210
Truck weight, kips
FIGURE 4. Frequency Histogram of Net Truck Weight Calculated as an Average of All Considered
Time Periods
The average interior truck weight is 105 kips — this is below the nominal value of 157.5 kips. However,
there are some exceptions where the estimated truck weights exceeded the nominal value.
As shown in Figure 5, estimated truck weights were further calculated for five probabilities of occurrence:
68, 95, 97, 99.5, and 99.95 percent. About 3 percent of all truck weights exceeded the nominal truck weight,
but 97 percent were below that value.

Bagdad In-trucks
200
180 2014_3
160 2014_4
Truck weight, kips

140
2015_1
120
100 2015_2
80 2015_3
60 2015_4
40 2016_1
20
2016_2
0
68% 95% 97% 99.50% 99.95%

FIGURE 5. Net Truck Weight for Five Probabilities of Occurrence

5. EVALUATION OF FAST BRIDGES

This section provides results from a three-dimensional FE model of four steel spans at FAST developed in
LUSAS™ software. Details of the bridge span at FAST can be found in a previously published Technology
Digest, TD-15-024 (5).
The three-dimensional analysis provides more valuable information for structural members not only at the
center of the span, but also at other locations — the model includes all members and the track structure.
The maximum stresses were checked at mid-span, quarter span, and at cover plate transitions under
various loads including: double-stack cars, typical 53-foot cars, and short, 42-foot cars. For all span lengths,
the intermodal cars did not produce higher stresses than 42-foot short cars or 53-foot standard cars.
Figure 6 presents bending stress comparisons at various locations along the four FAST spans due to four
considered car types. As predicted previously in Figure 2, intermodal cars show higher bending stresses
only for very short spans (≤ 15 feet); therefore, testing shorter spans or spans with short floor system
members in revenue service is recommended.

FIGURE 6. Bending Stresses at Mid-span, Quarter-span, and at Cover Plate Transition

6. MEASURED EFFECTS OF ARTICULATED DOUBLE-STACK CARS ON BRIDGE IN SERVICE
The analysis of spans of various lengths, using simple supported beam assumptions and various types of
double-stack cars, indicates that double-stack cars should cause larger maximum moments only on spans
shorter than 15 feet. Finite element modeling of four spans at FAST indicates that intermodal cars should
not govern the loading for the FAST steel spans, which range in length from 24 to 65 feet. Therefore, testing
shorter spans or spans with short floor system members under unit train traffic with different types of double-
stack cars in revenue service was needed.
In conjunction with BNSF Railway, TTCI measured the effects of both articulated double-stack railcars as
well as coal cars on a short span bridge near Kirkland, Texas, on the BNSF Red River Valley Subdivision
(Figure 7). The span is on a line that carries unit coal and grain traffic as well as intermodal traffic. The
bridge is built from six rolled beams 16 feet long (end to end) with an open deck. The beams are 15 inches
tall and 6 inches wide (S 15×60).

FIGURE 7. Bridge 209.22 I-Beam span near Kirkland, Texas

Strain gages were installed on all six beams as presented in Figures 8 and 9.

FIGURE 8. Scheme of gage locations

FIGURE 9. Photos of gage locations
Data sets were collected under revenue service trains traversing the bridge. Loaded articulated double-
stack cars were of most interest. For comparison purposes, loaded unit coal or grain trains also were
important. Data was collected for two days and during that time 14 trains passed over the bridge. Among
these, three were loaded unit trains and 10 were intermodal trains. Figure 10 presents maximum peak-
stress recorded under train passages.
 Maximum Stresses
12
Loaded Coal Intermodal Train
10

8
Stress, ksi

0
NS_out NS_mid NS_in SS_in SS_mid SS_out
FIGURE 10. Maximum peak-stress recorded under train passages
The maximum stresses measured under the loaded coal trains were generally as high as or higher than
the peak stresses measured under the intermodal trains.
Deflection of the bridge was measured from the middle beams beneath each rail at mid span. The maximum
measured deflection was around 0.2 inch. Figure 11 presents deflection under a unit coal train and Figure
12 presents deflection under an intermodal train. The deflections under the coal train were consistently near
the maximum for each car in the train, as expected due to uniform loading of each car. The deflections
under the intermodal train only had a few instances near the maximum value, indicating that many of the
trucks are more lightly loaded. Chart Title
0 20 40 60 80 100 120 140
0.05
Time, sec
0
Deflection, inch

-0.05

-0.1

-0.15

-0.2
North South
-0.25

FIGURE 11. Deflection histories for middle beams under loaded unit coal train
 Chart Title
0 20 40 60 80 100 120
0.05
Time, sec
0
Deflection, inch

-0.05

-0.1

-0.15

-0.2
South North
-0.25

FIGURE 12. Deflection histories for middle beams under intermodal train
The peak stresses vary from car to car. In order to use the data from a typical train pass for a fatigue life
estimate, the stress cycles should be counted using a rain flow cycle counting method (6).
The stress history for the center beams under a loaded unit coal train is presented in Figure 13.
Chart Title
12
SS_mid NS_mid
10

8
Stress, ksi

2
Time, sec
0
0 20 40 60 80 100 120 140
-2
FIGURE 13. Stress histories for center beams under loaded unit coal train
Distribution of the stress ranges is shown in Figure 14. This distribution shows that majority of the stresses
(100-120 counts) are in the range of 8 to 9 ksi. However; there are several cycles in the range of 9 to 10
ksi. The equivalent stress range for the south center beam is 8.8 ksi, including only stress ranges above 6
ksi (129 cycles).
 Stress Cycle Distribution - EB_LOCO_8523
140
NSMIDstress
120
Number of cycles

SSMIDstress
100

80

60

40

20

0
<1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10
Stress range, ksi

FIGURE 14. Cycle counts for center beam under unit coal train
The stress history under an intermodal train is presented in Figure 15. As with the deflection data, it is much
more variable than the stress history under the loaded unit coal train. Distribution of the stress ranges is
shown in Figure 15. This distribution shows that the stresses are broadly distributed but many of the cycles
(~75 counts) are in the range of 3 to 6 ksi. However; there are also higher cycles in the range of 9 to 10 ksi
and 10 to 11 ksi. The equivalent stress range for south center beam is 7.9 ksi, including only stress ranges
above 6 ksi (24 cycles).

Chart Title
12
SS_mid NS_mid
10

8
Stress, ksi

2
Time, sec
0
0 20 40 60 80 100 120
-2
FIGURE 15. Stress histories for six beams under intermodal train
 Stress Cycle Distribution - WB_LOCO_7086
35
NSMIDstress
30
Number of cycles

SSMIDstress
25

20

15

10

0
<2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10 10-11
Stress range, ksi

FIGURE 16. Cycle counts for center beam under intermodal train
In terms of fatigue, the equivalent stress range was about 10 percent lower under a typical intermodal
train as compared to a coal train. The number of accumulative stress cycles was about 80 percent lower.

7. SUMMARY AND CONCLUSIONS

Overall, the tested bridge span experienced comparable maximum load effects under articulated double-
stack cars and coal cars. There was one train in which the double stack cars produced higher stresses.
In terms of stress cycles, the stress ranges under an intermodal train are more variable than the stress
ranges under a loaded unit coal train. The loaded unit coal train produces 100-120 cycles per train in the
range of 8 to 10 ksi. While, the intermodal train produces only up to 20 cycles in the range of 8 to 11 ksi.
Note that short railroad bridges, including the one tested, are often built using rolled beams that are less
prone to fatigue than built-up riveted girders.

ACKNOWLEDGMENTS
The authors acknowledge BNSF Railway for allowing TTCI to perform a test on their bridge in revenue
service.

REFERENCES
1. Rakoczy A.M., D. Otter, and S.M. Dick. “Effects of Articulated Double-Stack Cars on Bridges”
Technology Digest TD17-020, AAR/TTCI, Pueblo, CO, August 2017.
2. Rakoczy A.M., D. Linkowski, D. Otter, and S.M. Dick. “Measured Effects of Articulated Double-Stack
Cars on Bridges” Technology Digest approved for publication, AAR/TTCI, Pueblo, CO, 2019.
3. Dick, S. “Legacy Train Configurations for Fatigue Life Evaluation of Steel Railway Bridges,”
Proceedings of the AREMA Annual Conference, Orlando, FL, 2016.
4. American Railway Engineering and Maintenance of Way Association (AREMA), Manual for Railway
Engineering, Chapter 15, Washington, D.C., 2015.
5. Otter, D., A.M. Rakoczy, and S.M. Dick. “Steel Bridge Life Extension for Riveted Steel Girder Spans at
FAST.” Technology Digest TD-15-024, AAR/TTCI, Pueblo, CO, August 2015.
6. Fisher, J.W., Kulak, G.L., and Smith, I.F.C.. “A Fatigue Primer for Structural Engineers.” National
Steel Bridge Alliance. May 1998.