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]

NUMBER OF WORDS: 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 . In addition, strain gage data was collected from a short bridge on a line carrying both intermodal and coal .

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 , also known as a double-stack car or stack car, is a type of specially designed to carry intermodal (shipping containers) used in intermodal (3). The "well" is a depressed section which sits close to the rails between the wheel trucks of the car, allowing a to be carried lower than on a traditional . This makes it possible to carry a stack of two containers per unit on railway lines (double-stack ) wherever the 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 . 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 A D E F Axles Trucks Loaded Loaded 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 70 60 50 40 30 20 10 0 Equivalent Cooper Loading (E) 0 10 20 30 40 50 60 70 80 90 100 End Shear 80 70 60 50 40 30 20 10 0 Equivalent Cooper Loading (E) 0 10 20 30 40 50 60 70 80 90 100 Floor Beam Reaction 80 70 60 50 40 30

20 10

Equivalent Cooper Loading (E) 0 0 10 20 30 40 50 60 70 80 90 100 Span length (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 14% 2014_4 12% 2015_1 10% 8% 2015_2 6% 2015_3 4% 2015_4 Percent of Percent sample 2% 2016_1

0% 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 2016_2

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% 14% 12% 10% 8% 6% 4% Percent of Percent sample 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_4 140 2015_1 120 100 2015_2 80 2015_3 60 2015_4

Truck weight, kips 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 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.

12 Loaded Coal Intermodal Train 10

8

6 Stress, ksi

4

2

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. 0 20 40 60 80 100 120 140 0.05 Time, sec 0

-0.05

-0.1

Deflection, inch -0.15

-0.2 North South -0.25

FIGURE 11. Deflection histories for middle beams under loaded unit coal train

0 20 40 60 80 100 120 0.05 Time, sec 0

-0.05

-0.1

Deflection, inch -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. 12 SS_mid NS_mid 10

8

6

Stress, ksi 4

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).

140 NSMIDstress 120 SSMIDstress 100

80

60

Number of cycles 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).

12 SS_mid NS_mid 10

8

6

Stress, ksi 4

2 Time, sec 0 0 20 40 60 80 100 120 -2 FIGURE 15. Stress histories for six beams under intermodal train

35 NSMIDstress 30 SSMIDstress 25

20

15

Number of cycles 10

5

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.

Articulated Double Stack Car Effects on Bridges

Anna Rakoczy, Duane Otter, David Linkowski, Stephen Dick Transportation Technology Center, Inc. Articulated Double Stack Car Effects on Bridges

• Motivation • Car Configurations • Analytical Findings • Wayside Data • Analytical Evaluation of FAST Bridges • Evaluation of Bridge in Revenue Service • Summary and Conclusions Articulated Double Stack Car Effects on Bridges Motivation: • Increases in double-stack intermodal traffic • Nominal truck vertical load for articulated double-stack cars is 157.5 kips — about 10% increase higher than interchange cars with a maximum truck load of 143 kips. Articulated Double Stack Car Effects on Bridges

Number of Weight (lbs.) Length dimensions (ft.) Truck Wheelbase (ft.) Railcar Type Truck Centers Axles Trucks Platform Loaded Total Loaded A D E F B C 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. Articulated Double Stack Car Effects on

Bridges Bending - Mid Span 80 70 Analytical Findings - 60 double-stack car effects 50 are highest only on 40 shorter spans and floor 30 systems. 20 10 0 Equivalent Cooper Loading (E) 0 10 20 30 40 50 60 70 80 90 100 Span length (ft)

6-Axle Loco 53'0" Coal Car International Stack - 3 Platforms International Stack - 5 Platforms Domestic Stack - 3 Platforms Domestic Stack - 5 Platforms Articulated Double Stack Car Effects on Bridges End Shear Floor Beam Reaction 80 80 70 70 60 60 50 50 40 40

30 30 20 20 10 10

0 Equivalent Cooper Loading (E) Equivalent Cooper Loading (E) 0 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Span length (ft) Span length (ft) 6-Axle Loco 53'0" Coal Car International Stack - 3 Platforms International Stack - 5 Platforms Domestic Stack - 3 Platforms Domestic Stack - 5 Platforms Articulated Double Stack Car Effects on Bridges Wayside Data • Wayside truck force data obtained from four different sites: Bagdad (CA); Gothenburg (NE); Vine Creek (IN), and Goodeve (Canada). Frequency Histogram - Bagdad In-trucks • Two years of data 16% 2014_3 14% 2014_4 12% 2015_1 • Bagdad contains the highest 10% 8% 2015_2 number of records. 6% 2015_3 4% 2015_4 • No significant seasonal of Percent sample 2% 2016_1

0% 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 differences 2016_2 Truck weight, kips Articulated Double Stack Car Effects on Bridges

• Average interior truck weight is 105 kips — below the nominal value of 157.5 kips. • About 3% of all truck weights exceeded the nominal truck weight.

Frequency Histogram - Bagdad In-trucks Bagdad In-trucks 16% 200 14% 180 2014_3 12% 160 2014_4 140 10% 2015_1 120 8% 100 2015_2 6% 80 2015_3 4% 60 2015_4 Percent of Percent sample 2%

Truck weight, kips 40 2016_1

0% 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 20 2016_2 0 Truck weight, kips 68% 95% 97% 99.50% 99.95% Articulated Double Stack Car Effects on Bridges Analytical Evaluation of FAST Steel Bridges: • 3-D finite element models of four steel spans at FAST developed in LUSAS™ software • 4-axle bulk commodity cars govern Articulated Double Stack Car Effects on Bridges Analytical Evaluation of FAST Bridges: • For all analyzed spans, articulated intermodal cars produced lower stresses than 42-foot sand cars or 53-foot coal cars. • Testing shorter span or bridge with short span floor system members in revenue service was recommended. Articulated Double Stack Car Effects on Bridges Evaluation of Bridge in Service • Short span bridge near Kirkland, Texas, on BNSF Red River Valley Subdivision • Span carries unit coal and grain traffic as well as intermodal traffic. • Tangent track. Articulated Double Stack Car Effects on Bridges • Six rolled beams 16 feet long (end to end) with open deck. • Beams are 15 inches tall and 6 inches wide (S 15×60) • Strain gages installed on all six beams • Deflection measured from middle beams at mid span. Articulated Double Stack Car Effects on Bridges Articulated Double Stack Car Effects on Bridges • Data collected for 3 coal trains, 11 intermodal trains • Maximum measured stress ~10 ksi • Maximum measured deflection ~0.2 inch. Articulated Double Stack Car Effects on Bridges Articulated Double Stack Car Effects on Bridges Articulated Double Stack Car Effects on Bridges Articulated Double Stack Car Effects on Bridges Articulated Double Stack Car Effects on Bridges Summary and Conclusions • Primary concern is short spans, floor systems, deck ties. • Double-stack cars are often not loaded to their full weight capacity. • Double-stack car effects may be comparable to or less than coal car effects. • Stress history under intermodal trains is more variable than under coal trains. Articulated Double Stack Car Effects on Bridges Summary and Conclusions • In terms of fatigue: • The number of accumulative fatigue stress cycles (>6 ksi) is about 80% lower. • The equivalent stress range is about 40% lower under a typical intermodal train as compared to a coal train. • Note: short railroad bridges are often built from rolled beams that are less prone to fatigue than built-up riveted girders. Thank you!