fatigue loading behavior of general mixed freight traffic

8/3/2019 Fatigue Loading Behavior of General Mixed Freight Traffic

1/23

Stephen M. Dick 1

FATIGUE LOADING BEHAVIOR OF GENERAL MIXED FREIGHT TRAFFIC

STEPHEN M. DICK, P.E., S.E., PH.D., SENIORRAILROAD ENGINEER, HANSON-WILSON,INC.,

KANSAS CITY, MISSOURI

ABSTRACT

Unit trains of have been the standard for development of fatigue design recommendations. While

unit-trains in various forms are representative of a significant portion of railroad traffic in various

forms, a large portion of traffic still moves in manifest trains consisting of a mix of empty and

loaded equipment. This is recognized in the AREMA recommendations although no specific

guidelines are included.

This study analyzes various configurations of empty and loaded equipment using new

formulation and compares the results of the analysis against unit-train behavior. The analysis is

based on new fatigue theory that utilizes specific car dimensions along with the relationship between

the railcar dimensions and the specific span length. The results are compared against the unit trains

consisting of similar equipment with a variety of span lengths on girders in the typical range of span

lengths. The results can be used for both practical application and inclusion into the design and

analysis recommendations where applicable.

INTRODUCTION

While freight traffic on railroads increasingly moves in unit-train consists, a major portion of traffic

still moves in general manifest trains. The traffic mix for this type of train includes many types of

freight that does not lend itself to unit-train movements. Materiel included in this category is lumber

and other construction materials, chemicals, general boxcar traffic, and a very necessary railroad

function, backhaul of empty railcars. Movement of this traffic results in mixed train consists that

have been the mainstay of railroad traffic since the beginning.


2/23

Stephen M. Dick 2

Fatigue analysis of railroad traffic has concentrated on unit-train consists. This work has

resulted in the current fatigue provisions contained in Chapter 15, Steel Structures (1). Those

provisions have been based on the railcar configuration resembling the unit coal train railcar. Given

the high unit weight of this railcar and the axle loadings that are at the maximum allowable for

unrestricted interchange, the magnitude of the bending moments are among the highest of any

railcar equipment. Combined with the number of unit-train movements for coal and grain, the

choice is appropriate when examining maximum bending moment magnitudes.

Fatigue depends not only upon an overall magnitude of bending moment. Fluctuation in

bending moment from the change in live-load bending moment creates potentially damaging cycles

without the necessity of the highest magnitude of bending moment. The fluctuation of bending

moment is provided with combinations of loaded and empty railcars, with the possibility of high

magnitude stress ranges potentially leading to damaging fatigue cycles.

While unit coal and grain trains accumulate tonnage at a fast rate, mixed trains will

accumulate tonnage at perhaps half that rate given the assumption that about half of the railcars in

mixed train are empty. If the potential for greater numbers of fatigue cycles is realized from mixed

train traffic, the assumptions in AREMA Chapter 15 for design and analysis may require revision

from both an estimated number of cycles along with the assumptions of total tonnage and its

relationship to fatigue damage.

A distinction is made in this report between design and analysis. New girder construction

generally consists of either bolted or welded construction, conforming either to a Category B or

Category C designation for potential fatigue damage. With a Category C fatigue detail, the minimum

stress range creating damage is 10 ksi (kips per square inch, or thousand-pounds per square inch).

Rating of existing bridges includes dealing with riveted girder spans that have been in place for quite

a long time. For fatigue analysis of riveted girders, live-load stress ranges in excess of 6 ksi are


3/23

Stephen M. Dick 3

assumed to create fatigue damage. That difference is significant, and an examination of the

differences in potential for fatigue damage between design and analysis are worth exploring.

RESEARCHAPPROACH

The range of girder lengths in this study is the same range specified in Chapter 15: 30 feet to 150

feet. As is usual practice for railroad spans, the span lengths are assumed to be simply supported.

The necessary data to examine the potential fatigue effects is the live-load moment time history for

arrangements of railcars, along with assumptions of span data to be able to calculate the potential

stresses. Given the variability of railcar equipment, a variety of equipment was needed to examine

the effects. Assumptions of span data are also required for calculation of potential stresses.

Railcar Equipment

For railcar equipment, conceptual dimensions were assumed that are reflective of current equipment

in use presently. Conceptual railcar lengths chosen were 50 feet, 75 feet, and 100 feet. Figure 1

shows the typical dimensions for railcars along with the analytical dimensions needed for bending

analysis. Table 1 shows the dimensions for the specific railcars chosen for this study. For all

railcars, a value of 0.75 was assumed for the ratio of the truck center distance to the overall length.

This value is in line with current equipment configurations, although this value can vary. For all

equipment the axle spacing on the trucks was assumed to be six feet, with a gross maximum weight

of 286,000 pounds, and a light weight of 66,000 pounds.

The rationale for the railcar lengths is to use factored multiples of railcar lengths that closely

represent prototype equipment. This allows mixing of the equipment in combinations that reflect

actual operating practices for analysis. With the use of the same overall weight for the various

configurations, the comparison of bending moments due to either loaded or empty railcars shows


4/23

Stephen M. Dick 4

LO

FIGURE 1. Dimensions Used For Analysis

LO - Overall length of railroad car measured over the pulling face of the coupler.

TC - Length between the center pin on the trucks, known as the truck center distance.

SI - Inboard Axle Spacing, the distance between the inside axles of the railroad car.

SO - Outboard Axle Spacing, the distance between the outside axles of the railroad car.

ST - Truck Axle Spacing, the distance between the adjacent axles of a truck.

n - number of axles

P - axle load

LO

TC TC

ST ST ST STSO/2 SO/2SI SO SI

All dimensions are in feet and kips.


5/23

Stephen M. Dick 5

TABLE 1. Railcar Dimensions Used in the Analysis.

Overall Truck Axle

Length Centers Spacing LO SO SI ST

50' Car 50 37.5 6 50.00 6.50 31.50 6.00

75' Car 75 56.25 6 75.00 12.75 50.25 6.00

100' Car 100 75 6 100.00 19.00 69.00 6.00

All dimensions are in feet.

All railcars are assumed the following weights:

Loaded 286,000 pounds on four axles

Empty 66,000 pounds on four axles


6/23

Stephen M. Dick 6

the effects of railcar length. The rationale for the assumed weight is the current allowable maximum

allowable weight for unrestricted interchange of rolling stock. The empty weight of the railcars is

based on the assumption of a payload of 110 tons, or 220,000 pounds. This leaves a remainder of

66,000 pounds for the light weight.

The 50-foot railcar is similar to the standard unit-train coal car and produces similar results.

The 75-foot railcar is similar to center-beam bulkhead flatcars used in lumber and building material

service. This railcar model also approximates chemical and tank cars. The 100-foot railcar is longer

than any current railcar in service. The longest railcars currently used are approximately 95 feet long

represented by TOFC flatcars and standard auto racks. The assumption of 286,000 pounds for a

gross maximum weight is above the maximums for the majority of that equipment, but 95-foot

flatcars with an allowable gross weight of 263,000 pounds exist in the Trailer Train fleet. Also,

construction of a long railcar with an allowable maximum of weight of 286,000 pounds is possible,

even if that has not been accomplished at this time.

The railcars were modeled in several combinations in multiple locations on the girder span

lengths of 30 feet to 150 feet in ten-foot increments. The locations included for calculation of

moment were the quarter point, 3/8 point, and at midspan. All three railcar configurations were

first run in a unit-train format to calculate the maximum bending moments for each type of railcar

on all span lengths. The bending moments of the 50-foot railcar were used to verify the data in

Table 15-9-1 of AREMA Chapter 15 which is the basis for the number of cycles assumed per train

crossing for design and rating. With verification of the design table, comparisons of the different

combinations of railcars were used to determine the adequacy of Table 15-9-1 for analysis.

Besides the typical unit-train configuration assuming all loaded railcars, the load-empty

combination was used. The general combination used was two loaded railcars followed by two

empty railcars, repeating that combination as necessary. The choice of two loaded railcars followed


7/23

Stephen M. Dick 7

by two empty railcars allows a full live-load moment range to be developed on all span lengths. For

most of the railcar combinations this allowed placement of the axle loads such that when the two

loaded railcars are placed on the span, the live-load moment that is generated is the same as if the

loading was a unit train. When the empty railcars are placed on the span such that they are the only

loading, a live-load moment that is due to the empty railcar weight is produced. With the disparity

in weight between loaded and empty railcars, a great difference in magnitude in live-load moment is

created, which has the potential for developing a damaging fatigue cycle.

The 50-foot railcar does not follow that convention, however. For a span length beyond the

length of twice the railcar, axles from the empty railcars will load the span. When those axles are on

the span, maximum live-load moment will not be generated, but the greater variation in live-load

moment will be apparent. Even though the overall moment is lower than the all loaded unit-train

moment, the variation in moment may still generate a damaging fatigue cycle.

This loading pattern is recognized by AREMA in Article 9.1.3.13r describing the potential

for one damaging fatigue cycle for every two railcars with combinations of loaded and empty

railcars. The article further states that the assumption of two loaded followed by two empty railcars

as a typical pattern was not considered a likely combination and the reduced number of 3 load-

empty combinations was assumed for Table 15-9-1. Three cycles is the assumed number of

damaging fatigue cycles in that table for all span lengths in excess of 100 feet.

For each railcar configuration, the loaded version of the railcar was run with empty versions

of all other railcars. This ensured that all combinations of loads and empties were established and

checked the potential effects of shorter loaded railcars with longer empty railcars and vice versa.

The shorter loaded railcar (50 feet) provides the highest overall bending moment of any of these

combinations while the longer railcars, when empty, will generate the lower magnitudes of moment.


8/23

Stephen M. Dick 8

Sample Girder Bridge Span Sizing

For examination of the potential for creating fatigue cycles, the magnitude of live-load moment

needed to create a damaging cycle had to be estimated. Comparison against spans from various

design loads can create disparities since the loadings have changed over time. While the magnitude

of the Cooper E loading (2) has been modified over time, allowable stresses and impact equations

over time have changed as well.

To examine the potential for this disparity, the design moments with impacts were calculated

for a series of design moments from Cooper E40 through E80, taking into account the impact

formulae in use at that time along with the allowable stresses. As an example, in 1920 the design

loading was Cooper E60 with steam locomotive (hammer blow) impact and a maximum allowable

stress of 16 ksi. Currently, the design loading is Cooper E80 with rolling impact replacing hammer

blow, and a maximum allowable stress of 20 ksi.

The ratio of the allowable stresses from Cooper E80 to Cooper E60 is slightly smaller than

the ratio of the design loads (1.25 versus 1.33). The differences in impact (hammer blow in 1920

versus rolling impact in 2004) reduce the discrepancy between loads and stresses. Rolling impact is

approximately 75 percent of hammer blow impact depending upon span length and formula used.

Application of rolling impact eliminates the discrepancy between loads and stresses and makes the

overall design requirement for current Cooper E80 actually somewhat less than 1920 Cooper E60.

Figure 2 graphically shows this relationship between design loading, impact, and allowable stresses.

The assumption for normalization of the loadings is the Cooper design moment with impact

applied, then adjusted by the ratio of the allowable current allowable stress to the historic allowable

stress corresponding with that design loading. Figure 2 demonstrates that while the E40 design

loading of 1900 is lighter than the other loadings, the requirements for support of live load are

essentially the same since the 1920 upgrade to Cooper E60 with impact and allowable stresses


9/23

Norm

alizedBendingMoment

1920 Coo

2004 Cooper E80

30 40 50 60 70 80 90 100 110 120

Span Length (ft)

FIGURE 2. Comparison of Normalized Design Loadings


10/23

Stephen M. Dick 10

modified at the same time that the Cooper loading was increased. The overall effect is that the span

requirements have changed little in the last 80 years.

The estimate of dead load for each span length was based on historical data used for

estimation of steel weight for girder spans of a specific length along with inclusion of deck weight

for ballasted deck bridges. Combining the dead load with the design live load plus impact gives a

total moment required for design. With the total design moment calculated, the assumption of

maximum allowable stress can be applied, so that the individual stresses for dead load and live load

plus impact can be identified and quantified.

Overall design impact and fatigue impact for design or analysis are different values. It has

been recognized that impact experienced in actual operations is well below design impact for the

majority of equipment. For different types of spans, the results of testing on Canadian National

demonstrated that average impacts are sometimes significantly below design impact. AREMA

Chapter 15 takes this into account with reductions for design impact for both design and rating of

fatigue details based on those findings. With a reduced impact a threshold of 6 ksi for rating of

existing bridges and 10 ksi for design, the live load moment for each span length can be calculated

that exceeds that stress threshold. The moments generated by each equipment combination are

then checked against the threshold moments to identify the potentially damaging combinations.

Calculation of Bending Moments

The moments for the combinations of loaded and empty railcars were calculated using an influence

line computer program, MOMENTS, which provides a time history live-load moment trace as the

load is passed over a specified span length. The output of the program is the history of the

calculated live-load bending moment at an incremental step while the load is passing over the

specified span length. The uniform increment is specified by the user, along with the loading


11/23

Stephen M. Dick 11

pattern, span length, and the location on the span that is under investigation. With that, the bending

moment can be calculated in very close increments for each loading, allowing greater precision for

the calculation.

To check the accuracy of the bending moment calculations, a second computer program,

RANGE, was used. The second program calculates the maximum and minimum moments at any

number of points for a specified railcar configuration on a given span length. This output of this

program was used to verify the output of the MOMENTS program. The RANGE program is also

useful for determining the maximum moment range that is encountered from unit train loadings.

For moment range calculation with the mixture of loaded and empty railcars, the MOMENTS

program proved more useful by examination of the time history of the bending moment as the loads

were passed over the span length.

ANALYSIS OF RESULTS

Samples of the output from the MOMENTS program are shown in Figures 3 and 4. Figure 3 shows

the bending moment versus time for a mixture of loaded and empty railcars where the span length is

short enough that the span is loaded and then completely unloaded by the passage of each car. The

magnitude of bending moment is related to axle weight as the main criteria with the span completely

loading and unloading for each pair of railcar trucks that passes over the bridge. The potential for a

damaging fatigue cycle is based on the overall magnitude of the bending moment which is mainly

dependent upon the magnitude of the axle loading.

Figure 4 shows the time history of bending moment for two train configurations. One

configuration is for a unit train of loaded railcars while the other configuration is a combination of

loaded and empty railcars. The behavior of equivalent magnitudes of maximum moment is shown

in Figure 4 while the phenomenon of empty railcars clearly shows the potential for damaging fatigue


12/23

BendingMoment

Time

FIGURE 3. Bending Moment Versus Time For Railcar Loadings Over Short-Span Brid


13/23

Loaded Cars Only

Loaded and Unloaded Cars

BendingMoment

Time

FIGURE 4. Bending Moment Versus Time For Railcar Loadings Over Long-Span Bridg


14/23

Stephen M. Dick 14

cycles. The empty railcars produce a much lower magnitude of bending moment and the difference

clearly has potential for creating fatigue damage. What can also be seen from that moment trace is

that another peak in bending moment can be developed at the locale of a loaded railcar next to an

empty railcar. In the case shown, the magnitude of the peak is not significant, but can be significant

on some span lengths. If all peaks from the bending moment had sufficient magnitude, the

potential number of damaging cycles is 75 percent of the total number of railcars in the train.

The data for all loading combinations was analyzed to determine if the peaks were of

sufficient magnitude to develop a damaging fatigue cycle. Adjustments to the data were made so

that the calculated bending moment and bridge capacity were being compared appropriately. The

bending moment from the car combinations was calculated assuming static loading, i.e., no impact.

The live load stress capacity calculated for the spans assumes full design impact. For fatigue

analysis, design impact is reduced to reflect what is experienced by bridges under actual loading

conditions, which is generally much lower than design for girder spans. To ensure that the bending

moments and allowable stresses were comparable, the overall live load plus impact design moment

was assumed to require all of the available live-load stress. The calculated live-load bending moment

capacity was then prorated to a stress of either 6 ksi for examination of existing riveted spans, or to

10 ksi for examination of spans with welded stiffeners; useful for either analysis or design of such

spans. Assuming the lower impact used for fatigue analysis, the moment capacity was further

reduced by factoring impact revised for fatigue, leaving a live-load bending moment capacity that is

strictly for statically calculated bending moments, as is generated by the computer programs. The

bending moments generated by the computer programs were then checked against the calculated

bending moment capacities for each span length.

The checks of bending moments were for three separate conditions. The first condition was

design of new spans or analysis of spans with welded stiffener details, the second condition was for


15/23

Stephen M. Dick 15

fatigue analysis of existing spans using the 1900 design load of Cooper E40, and the third condition

was using 1920 Cooper E 60 design conditions, which is equivalent to later design loads. The first

condition assumed a threshold of 10 ksi for fatigue damage potential (Category C) while the second

and third conditions assumed a threshold of 6 ksi for fatigue damage potential (used for fatigue

analysis of riveted details).

DISCUSSION OF RESULTS

The bending moment time histories were compared against the bending moment capacities for each

span length assuming the conditions previously described. For design, the check resulted in the

reaffirmation of the information provided in Table 15-9-1. Beyond this point, the check for fatigue

damage potential from loaded and empty cars did not produce any results that would modify that

table, at least from the basis of railcars with 286,000 pounds on four axles. The assumptions for

design of new spans and analysis of welded spans with Category C (or better) fatigue details appear

to be sufficient under current conditions.

Analysis of existing riveted spans shows a different condition, however, with results that may

need to be taken into account for future fatigue analysis of that type of bridge. Tables 2 and 3

display the results of analysis for potentially damaging fatigue cycles for the 1900 Cooper E40 and

for Cooper E60 and higher, respectively. An examination of those tables shows that loaded railcar

trains do behave similarly to the assumptions used for design. Both the 75-foot and 100-foot cars

show that each railcar can produce a damaging cycle for assumed unit-train conditions. The results

of the 100-foot railcar should not become a center of focus, however, as a 286,000-pound car of

that length does not exist. The results of the other railcar configurations combined with an empty

100-foot railcar, however, do provide results that need to be taken into consideration.


16/23

Stephen M. Dick

TABLE 2. Number of Potential Cycles for Combinations of Loaded and Empty Railcars for Coo

Railcar

Configuration 30 40 50 60 70 80 90 100 110 120

50' Loaded Car Combinations

50L 100 100 100 1 1 1 1 1 1 1

50L 50E 75 75 25 25 25 25 25 25 25 1

50L 75E 75 75 25 25 25 25 25 25 25 25

50L 100E 75 75 50 25 25 25 25 25 25 25


75L 100 100 100 100 100 1 1 1 1

75L 50E 75 75 75 75 25 1 1 1 1

75L 75E 75 75 75 75 25 25 25 1 1

75L 100E 75 75 75 50 25 25 25 25 1


100L 100 100 100 100 100 100 1 1

100L 50E 75 75 75 75 50 25 1 1

100L 75E 75 75 75 75 50 25 1 1

100L 100E 75 75 75 75 25 25 1 1

Number of cycles are based on 100-car train lengths, 286,000 pounds maximum weight, locomotive effect

Categories are coded for loaded and unloaded cars. The following is an example:

50L 75E indicates two loaded 50' railcars with two 75' empty railcars.

Span Length


17/23

Stephen M. Dick

TABLE 3. Number of Potential Cycles for Combinations of Loaded and Empty Railcars for Coo

Railcar

Configuration 30 40 50 60 70 80 90 100 110 120


50L 100 100 100 1 1 1 1 1 1 1

50L 50E 75 50 25 25 25 1 1 1 1 1

50L 75E 75 50 25 25 25 25 25 25 25 1

50L 100E 75 50 25 25 25 25 25 25 25 25


75L 100 100 100 100 1 1 1

75L 50E 75 75 50 25 1 1 1

75L 75E 75 75 75 25 25 25 1

75L 100E 75 75 75 25 25 25 25


100L 100 100 100 100 100 1

100L 50E 75 75 75 25 25 1

100L 75E 75 75 25 25 25 1

100L 100E 75 75 25 25 25 25

Number of cycles are based on 100-car train lengths, 286,000 pounds maximum weight, locomotive effect

Categories are coded for loaded and unloaded cars. The following is an example:

50L 75E indicates two loaded 50' railcars with two 75' empty railcars.

Span Length


18/23

Stephen M. Dick 18

The significance in Tables 2 and 3 is in the results of the load-empty combinations and the

potential number of damaging cycles. For Cooper E40 bridges (Table 2), the 50-foot mixed train

combinations show that beyond a span length of 50 feet, the number of potentially damaging cycles

represents 25 percent of the number of cars for the combination of two loaded and two empty

railcars. Though the range of spans is not as great, the same behavior is shown for 75-foot loaded

railcars with the same empty combinations. The number of potentially damaging cycles goes

beyond what is shown in Table 15-9-1 for span lengths in excess of 75 feet.

Similar results are apparent for Cooper E60 and later designs as shown in Table 3. The

difference between the two tables is in the range of span lengths that are affected by the load-empty

combinations. The same percentage of cars represent potentially damaging fatigue cycles for both

Cooper E40 and Cooper E60 designs.

Tables 4 and 5 display the same data with emphasis on the maximums for each loaded railcar

length. A distinction is made between unit-train and mixed-train conditions so that an assumed

number of cycles can be applied to those train types. Again, for each type of span design, the

number of potentially damaging cycles is greater than what is assumed in Table 15-9-1. For mixed

train conditions, the number of potential cycles is not as great as for loaded unit trains on the

shorter spans, but the potential for multiple cycles above the span length of 75 feet is more

pronounced than for unit trains.

INFERENCE FORFATIGUE REQUIREMENTS

Fatigue analysis has historically used the same assumptions for both design of new spans and

analysis of existing spans. With the requirements for existing riveted details being significantly

below a Category C detail, evaluation criteria for existing spans need to be developed more fully to

reflect the potential for fatigue damage when those criteria are beyond what is needed for design.


19/23

Stephen M. Dick

TABLE 4. Maximum Number of Potential Cycles for Combinations of Loaded and Empty Railcar

Railcar

Configuration 30 40 50 60 70 80 90 100 110 12

Maximum number of fatigue cycles for loaded unit trains of each car length.

50' Loaded Cars 100 100 100 1 1 1 1 1 1 1

75' Loaded Cars 100 100 100 100 100 1 1 1 1

100' Loaded Cars 100 100 100 100 100 100 1 1

Maximum number of fatigue cycles for unit trains for all loaded car lengths.

Maximum cycles 100 100 100 100 100 100 1 1 1 1

Maximum number of fatigue cycles for mixed trains for each loaded car length.

50' Loaded Cars 75 75 50 25 25 25 25 25 25 2

75' Loaded Cars 75 75 75 75 25 25 25 25 1

100' Loaded Cars 75 75 75 75 50 25 1 1

Maximum number of fatigue cycles for mixed trains for all loaded car lengths.

Maximum cycles 75 75 75 75 50 25 25 25 25 2

Number of cycles are based on 100-car train lengths, 286,000 pounds maximum weight, locomotive effects n

Span Length (ft)


20/23

Stephen M. Dick

TABLE 5. Maximum Number of Potential Cycles for Combinations of Loaded and Empty Railcar

Railcar

Configuration 30 40 50 60 70 80 90 100 110 12

Maximum number of fatigue cycles for loaded unit trains of each car length.

50' Loaded Cars 100 100 100 1 1 1 1 1 1 1

75' Loaded Cars 100 100 100 100 1 1 1

100' Loaded Cars 100 100 100 100 100 1

Maximum number of fatigue cycles for unit trains for all loaded car lengths.

Maximum cycles 100 100 100 100 100 1 1 1 1 1

Maximum number of fatigue cycles for mixed trains for each loaded car length.

50' Loaded Cars 75 50 50 25 25 25 25 25 25 2

75' Loaded Cars 75 75 75 25 25 25 25

100' Loaded Cars 75 75 75 25 25 25

Maximum number of fatigue cycles for mixed trains for all loaded car lengths.

Maximum cycles 75 75 75 25 25 25 25 25 25 2

Number of cycles are based on 100-car train lengths, 286,000 pounds maximum weight, locomotive effects n

Span Length (ft)


21/23

Stephen M. Dick 21

This is demonstrated by the number of potentially damaging cycles on existing spans that are

postulated as shown Tables 2 through 5. The comparison of these figures to the assumptions for

design on new spans indicates that existing spans are possibly experiencing a greater percentage of

damaging cycles than previously envisioned.

A better understanding of this behavior also requires a better understanding of railroad

operating practices which can vary between companies. Mixed train composition depends upon the

loading patterns of loaded and empty railcars. Some railroads will group loaded railcars into one

block while the empty railcars are separated into another block. This reduces the number of

potentially damaging fatigue cycles considerably and modifies the results of this research. Blocking

of trains in this fashion promote better and safer train handling than a general mixing of loaded and

empty railcars.

Other railroads, however, will group railcars by destination without regard for separating the

loads and empties. This is done where swapping of blocks of railcars at intermediate points occurs

between trains based on destinations. While this method requires more careful train handling

procedures, it provides convenience in train makeup at the origin and facilitates block swapping at

the intermediate points. Mixed railcar consists of this variety will certainly generate more potentially

damaging fatigue cycles than separation of loads and empties. Better definition of the problem is

needed to determine an accurate number of potentially damaging cycles, and that definition may

vary from railroad to railroad, or even to different territories within the same railroad.

CONCLUSIONS AND RECOMMENDATIONS

This study demonstrated that mixed train traffic has the potential to generate more potentially

damaging fatigue cycles on longer girder spans than is currently reflected in Table 15-9-1 of Chapter

15 of the AREMA Manual for Railway Engineering. For shorter span lengths, unit-train loadings


22/23

Stephen M. Dick 22

still provide a potential for more fatigue cycles, but the ramifications for longer girder spans are

more pronounced with these findings.

A significant difference was noted for requirements of design of new spans versus analysis of

existing spans, especially existing riveted spans. The recommended values for damaging cycles in

Table 15-9-1 are adequate for new designs and analysis of welded girders with Category C details or

better. For analysis of riveted girder spans, especially those spans of light designs such as Cooper

E40, the table may understate the potential number of damaging cycles. New criteria to

demonstrate the potential for damage should be developed to evaluate existing riveted girder spans.

Operating practices of railroads should be studied to determine the loading characteristics of

mixed trains. Those railroads that mix loaded and empty railcars are more susceptible to generating

potentially damaging fatigue cycles than those railroads that separate loaded and empty railcars into

separate blocks. A study of those patterns would provide a better understanding of the nature of

load-empty railcar loadings along with more accurate estimation for the potential fatigue damage.

This study focused on the current maximum loading standard of 286,000 pounds on four

axles. If the maximum allowable weight is increased, these results should be reviewed to make

recommendations for fatigue damage potential for the increased weight. It can be expected that the

damage potential will increase to include more cycles on longer span lengths.

REFERENCES

1. American Railway Engineering and Maintenance of Way Association (AREMA) (2004),

Manual for Railway Engineering, Chapter 15, Washington, D.C.

2. Cooper, T. (1894), Train Loadings for Railroad Bridges, Transactions of the American Society of Civil

Engineers, Vol. 31, pp. 174-184.


23/23

Stephen M. Dick 23

List of Figures

FIGURE 1. Dimensions Used For Analysis

FIGURE 2. Comparison of Normalized Design Loadings

FIGURE 3. Bending Moment Versus Time For Railcar Loadings Over Short-Span Bridges

FIGURE 4. Bending Moment Versus Time For Railcar Loadings Over Long-Span Bridges

List of Tables

TABLE 1. Railcar Dimensions Used in the Analysis

TABLE 2. Number of Potential Cycles For Combinations of Loaded and Empty Railcars for

Cooper E40 Designs

TABLE 3. Number of Potential Cycles For Combinations of Loaded and Empty Railcars for

Cooper E60 Designs

TABLE 4. Maximum Number of Potential Cycles for Combinations of Loaded and Empty Railcars

for Cooper E40 Designs

TABLE 5. Maximum Number of Potential Cycles for Combinations of Loaded and Empty Railcars

for Cooper E60 Designs

fatigue loading behavior of general mixed freight traffic

Documents