Effects of extra long trains on track ?

5990 views
22 replies
1 rating 2 rating 3 rating 4 rating 5 rating
  • Member since
    December, 2007
  • From: Georgia USA SW of Atlanta
  • 8,888 posts
Effects of extra long trains on track ?
Posted by blue streak 1 on Monday, July 09, 2018 10:32 PM

We have note that if we take a small piece of steel and bend it many times that it will heat up and finally break.  However if we do the same number of bends but just  a few bends at a time the steel will not break unless doing a lot more bends.  Although not often we have noted that there has been several derailments due to rail breaks under the rear part of the derailed train and not under leading locos that weigh more than most cars.   

If the structure under a track has several  adjaecent cross tie sags would the rail more likely break with the constant bends and returns ?  It might be temperature sensors on rails at sags would provide answers ?  Anyone know if research has been done ?  Would this accelerate metal fatigue over say 2 - 500 axel trains ? Maybe at Pueblo ?  A  1000 axel train certainly has a lot of bends of a track .    Are we off base ?

What about bridges ?

  • Member since
    December, 2005
  • From: Cardiff, CA
  • 2,925 posts
Posted by erikem on Monday, July 09, 2018 11:41 PM

When you bend a piece of steel to the point of it taking a permanent set, you have exceeded the elastic limits of the steel. When this happens, the steel will typically work harden and when stretched or compressed too much will start cracking. Rail and bridges are almost never loaded past their elastic limit in normal operation, but fatigue is thought to be caused by progression of micro-cracks from loading near the elastic limit (also known as the yield limit).

  • Member since
    October, 2006
  • From: Allentown, PA
  • 9,409 posts
Posted by Paul_D_North_Jr on Tuesday, July 10, 2018 8:05 PM

Bridges definitely have a limit on how many loads of varying weight they can carry - governed by fatigue as you surmise.  It depends on their 'history'.  The formulas are pretty complex.  Heavy unit trains are usually the worst, depending on the axle loading and spacing, and where you're looking at in the bridge. 

Track is much more complicated, but what Erik says is a good summary.

- PDN.   

"This Fascinating Railroad Business" (title of 1943 book by Robert Selph Henry of the AAR)
  • Member since
    February, 2013
  • 34 posts
Posted by bratkinson on Wednesday, July 11, 2018 8:30 PM

This discussion on bridge fatigue issues reminded me of when I was a contractor, developing software at an internationally known crane manufacturing company about 1982 or so.  In particular, the program was for their 'smaller' line of overhead traveling cranes that ride on a set of rails maybe 40-50 feet off the ground with the operator in a cab on the travelling boom to which the motorized pully lifting equipment is attached.  We've all seen pictures of such cranes at locomotive manufacturing and/or repair facilities that pick up an entire locomotive or heavy component and move it somewhere.

In particular, I was provided specs by the engineering department of how to create the program.  The 'user' input to the program was data regarding the height, weight, and load limits of the rail system on which the crane was to travel.  The distance between the rails (gauge) was critical as the further apart they are, the less load that can be picked up unless stronger bridge steel between the rails is used...up to a limit. 

Of course, the most 'major' criteria in designing the crane, whether done via slide rule, calculators, or computers, is what's the maximum weight to be picked up, how far apart are the rails, and how fast does it need to travel down the rails.  The program would then calculate a range of values for steel thickness, length, trusses if needed, etc, and provide a materials requirement list for each quality of steel needed (tensile strength/hardness).  The engineers would 'price it out' for each steel combination and ultimately get the crane built and installed for the customer.

One of the anecdotal stories the engineers told me was about a customer that had one of their cranes in their factory for over 50 years and wanted new one just like the old one for their new factory.  Of course, there was nothing common between the cranes as steel quality/strengths had been improved, and there was nothing 'common' between dimensions, weights, etc between the old and new cranes.  The new factory opened up with the new crane, and about 6 months later, the engineers were called to find out why the new crane had a very noticible 'sag' in the traveling boom.  The engineers triple and quadruple checked their designs for the crane along with the weight/speed/height requirements from the customer and everything was more than sufficient for the 150% load carrying factor designed into the crane.

So, they travelled to the new factory to see what was going on.  It turned out that like the old factory, they used the crane to pick up still red hot 5-10 ton metal castings and give it a couple of 'shakes' by releasing the hoist brake and then slamming on the brakes just before the casting hit the ground to shake loose any left over slag on the casting...just like they always did at the old factory for over 50 years!  

The difference was that in the 1920s when the original crane was built, metalurgy technology was still very limited as was their knowledge of tensile strengths of steel, etc.  Everything was done by slide rule calculations back then, too.  So, back then, stuff was designed for a 200% overload factor, but as they didn't know the limitations of the steel being used, it turned out that the cranes of that era really had close to a 300-400% overload capability!

In effect, the 'shake' technique created a very momentary apparent weight of well over 2-3 times of the casting being picked up.  It's comparable to a head on car accident where the 'crush' stresses far exceed the design limits of the car and it crumbles.  The result of using 'new' technology, lighter steel could be used, etc, and still have a 50% overload safety margin.  Unfortunately, the 'old' technology had roughly a 300% overload safety margin, which was being 'entered' every day.  They didn't tell me who picked up the tab for a new crane...

The same is true with railroad and highway bridges.  When they were built in the late 1800s/early 1900s, the engineers unknowingly were building them with exceedingly large overload handling factors.  The Smithfield Street Bridge in Pittsburgh PA comes to mind.  Built in 1883 and widened twice (per Wikipedia), it still has 4 lanes of traffic plus light rail street cars crossing it every day.  The 'built in' overload factor is a major factor in its longevity.  Compare that to various interstate highway bridge replacements in the past 20 years or so that I'm aware of that portions of the new bridge fell down in less than 5 years!  Minimal overload design factors and/or steel quality issues caused those failures.

Fortunately, the railroad bridges of yore were designed to handle the steam locomotives of the day, and then some.  How many diesel locomotives of today come anywhere near the weight of a fully loaded UP Big Boy, or a pair of N&W Y6Bs double headed?  In my opinon, bridges that could handle those locomotives will still be around 100 years from now with reasonable maintenance.

 

 

 

 

  • Member since
    May, 2003
  • From: US
  • 14,861 posts
Posted by BaltACD on Wednesday, July 11, 2018 8:42 PM

bratkinson

One of the anecdotal stories the engineers told me was about a customer that had one of their cranes in their factory for over 50 years and wanted new one just like the old one for their new factory.  Of course, there was nothing common between the cranes as steel quality/strengths had been improved, and there was nothing 'common' between dimensions, weights, etc between the old and new cranes.  The new factory opened up with the new crane, and about 6 months later, the engineers were called to find out why the new crane had a very noticible 'sag' in the traveling boom.  The engineers triple and quadruple checked their designs for the crane along with the weight/speed/height requirements from the customer and everything was more than sufficient for the 150% load carrying factor designed into the crane.

So, they travelled to the new factory to see what was going on.  It turned out that like the old factory, they used the crane to pick up still red hot 5-10 ton metal castings and give it a couple of 'shakes' by releasing the hoist brake and then slamming it on before the casting hit the ground to shake loose any left over slag on the casting...just like they always did at the old factory for over 50 years!  

The difference was that in the 1920s when the original crane was built, metalurgy technology was still very limited as was their knowledge of tensile strengths of steel, etc.  Everything was done by slide rule calculations back then, too.  So, back then, stuff was designed for a 200% overload factor, but as they didn't know the limitations of the steel being used, it turned out that the cranes of that era really had close to a 300-400% overload capability!

In effect, the 'shake' technique created a very momentary apparent weight of well over 2-3 times of the casting being picked up.  It's comparable to a head on car accident where the 'crush' stresses far exceed the design limits of the car and it crumbles.  The result of using 'new' technology, lighter steel could be used, etc, and still have a 50% overload safety margin.  Unfortunately, the 'old' technology had roughly a 300% overload safety margin, which was being 'entered' every day.  They didn't tell me who picked up the tab for a new crane...

The same is true with railroad and highway bridges.  When they were built in the late 1800s/early 1900s, the engineers unknowingly were building them with exceedingly large overload handling factors.  The Smithfield Street Bridge in Pittsburgh PA comes to mind.  Built in 1883 and widened twice (per Wikipedia), it still has 4 lanes of traffic plus light rail street cars crossing it every day.  The 'built in' overload factor is a major factor in its longevity.  Compare that to various interstate highway bridge replacements in the past 20 years or so that I'm aware of that portions of the new bridge fell down in less than 5 years!  Minimal overload design factors and/or steel quality issues caused those failures.

Fortunately, the railroad bridges of yore were designed to handle the steam locomotives of the day, and then some.  How many diesel locomotives of today come anywhere near the weight of a fully loaded UP Big Boy, or a pair of N&W Y6Bs double headed?  In my opinon, bridges that could handle those locomotives will still be around 100 years from now with reasonable maintenance.

Part and parcel to engineering to price without understanding the reality of how the product is actually used.

Bridges constructed today will not have the longevity of those constructed 100+ years ago.  100 years ago bridges were constructed to withstand the recipricating impacts of steam engines - that coupled with the limited understanding of exact material strengths meant that 100 years ago bridges were constructed much stronger, for a margin of safety, than they are being constructed today.

         

Never too old to have a happy childhood!

  • Member since
    October, 2006
  • From: Allentown, PA
  • 9,409 posts
Posted by Paul_D_North_Jr on Sunday, July 15, 2018 10:08 AM

Tensile strength is the most important design component, but not the only one by far.  It always enlightening to young engineers when they see that a structure they designed is OK for strength but they forgot to check for deflection (sag) . . . Oops - Sign 

Railroad bridge design these days still uses the Cooper E-series loading system - a pair of 2-8-0s with tenders of various weights, followed by a continuous load of X thousand lbs. per linear foot.  Plus, an alternative load of like 4 ea. 100,000 lb. closely spaced axles - think of a heavy-load flatcar, or the ends of 2 ballasted C-C locomotives close together (I know, that's 6 axles, not 4, but it's the same principle - still 400K lbs. along about 18 ft. of track, if I remember right).  That, and a more sophisticated - and very complicated - approach to fatigue management.  So the greater 'robustness' is designed in these days, instead of happening as an unintentional by-product of lack of detailed knowledge.  Plus, earthquake resistance is now designed in as well.  

Notably, almost all railroad bridges are simple-span designs, not continuous.  That won't mean much to those are are not structural engineers, but it kind of guarantees that some corners can't be cut to save some steel or cost somewhere.  

- PDN.  

"This Fascinating Railroad Business" (title of 1943 book by Robert Selph Henry of the AAR)
  • Member since
    December, 2001
  • From: Denver / La Junta
  • 9,323 posts
Posted by mudchicken on Sunday, July 15, 2018 11:06 AM

Back to the OP, simplifying things: It's not the length - it's the number of cycles compounded by the weight.

As PDN noted as well, it's darn complicated. And "foundation" is more than rails and bridges. Dozens and dozens of things have an effect on track modulus.

Mudchicken Nothing is worth taking the risk of losing a life over. Come home tonight in the same condition that you left home this morning in. Safety begins with ME.... cinscocom-west
  • Member since
    December, 2001
  • From: Northern New York
  • 18,696 posts
Posted by tree68 on Sunday, July 15, 2018 12:20 PM

Thinking about this - the idea that the length of the trains is a factor (as opposed to cycles and weight) would seem to imply that the track can "recover" between trains.

Reading the posts of our resident experts leads me to believe that's not the case.

If you bend a paper clip a bunch of times, but stop before it breaks, lay it down for a while, then start bending it again, it'll break at the same number of flexes as if you hadn't stopped.

LarryWhistling
Resident Microferroequinologist (at least at my house) 
Everyone goes home; Safety begins with you
My Opinion. Standard Disclaimers Apply. No Expiration Date
Come ride the rails with me!
There's one thing about humility - the moment you think you've got it, you've lost it...

  • Member since
    June, 2007
  • 189 posts
Posted by RDG467 on Sunday, July 15, 2018 1:48 PM

Blue Streak, the TTCI in Pueblo uses the Facility for Accerated Service Testing (FAST) loops to evaluate track components for fatigue. www.aar.com. 

  • Member since
    December, 2005
  • From: Cardiff, CA
  • 2,925 posts
Posted by erikem on Sunday, July 15, 2018 2:09 PM

tree68

If you bend a paper clip a bunch of times, but stop before it breaks, lay it down for a while, then start bending it again, it'll break at the same number of flexes as if you hadn't stopped.

The number of times that you can bend a paper clip can vary a huge amount between bending within the elastic limit and bending while exceeding the elastic limit. The elastic limit is the point where bending beyond results in a permanent change in shape. Other than the contact surface of the rail, I don't see anything in the track structure that will approach the elastic limit.

Keep in mind that "fatigue" usually applies to a wear-out mechanism for cyclic loading below the elastic limit. The theory is that such loading will generate micro-cracks that eventually lead to some cracks growing to the point that the piece fails.

As for the original question on length of trains causing problems, the most likely source of problems would be tractive forces from longer trains, with the worst case being an emergency braking application.

  • Member since
    April, 2003
  • 302,135 posts
Posted by Anonymous on Sunday, July 15, 2018 3:06 PM

erikem
Keep in mind that "fatigue" usually applies to a wear-out mechanism for cyclic loading below the elastic limit. The theory is that such loading will generate micro-cracks that eventually lead to some cracks growing to the point that the piece fails.

The dimensioning for fatigue aims on preventing this scenario by limiting the allowable stresses. How much depends on e.g. number of load cycles, tension or compression or alternating.

When ties loose their footing and don't support the rails anymore the fatigue is accelerated. The stresses will most likely stay within the elastic limits but the higher maximal stresses will reduce the number of allowed load cycles.
Regards, Volker

  • Member since
    December, 2005
  • From: Cardiff, CA
  • 2,925 posts
Posted by erikem on Sunday, July 15, 2018 10:16 PM

That's pretty much what my understanding is on fatigue, i.e. there is a cyclic stress limit, that if not exceeded, will allow for unlimited lifetime. This stress limit is below the yield stress for the material and how much below depends on the material.

From what I understand, the issue of fatigue first came up with railroad axles.

With regards to paperclips, being them to the point of causing a permanent change in shape will result in work hardening (won't go into the gory details about dislocations) which generally makes the material more brittle and the increasing amount of work hardening from continual reshaping eventually leads to a brittle fracture.

Anyway, my first exposure to elastic limits and work hardening was from my E45 class at Cal on engineering materials, with Professor Parker being one the lecturers for the course. The course did not cover much material on fatigue, though did get into stress concentrations.

 - Erik

  • Member since
    May, 2003
  • From: US
  • 14,861 posts
Posted by BaltACD on Sunday, July 15, 2018 10:35 PM

Railroads keep statistics about the number of ton miles that are operated over each track.  MofW uses these figures as one of the factors to be used when it becomes necessary replace rail and/or ties on a particular line segment.  Records are kept identifying what kind of rail was installed, when it was installed and any other pertinent information.  Records are also kept on when track last had a surfacing gang go over it and records for tie replacement.  Data is kept on virtually all aspects of tracks.

         

Never too old to have a happy childhood!

  • Member since
    December, 2005
  • From: Cardiff, CA
  • 2,925 posts
Posted by erikem on Monday, July 16, 2018 10:43 AM

Which makes sense as rail wear is proportional to the gross tonnage. Improvements in rail steel and maintennance have more han doubled the gross tonnage that a rail can support before replacing, but those improvements were based on the meticulous records of traffic. I suspect the use of wheel impact detectors is also an important contributor to extension of rail life.

On a related note, Timken was boasting of getting a 10 to 20X improvement in bearing life from improvements in steel quality (fewer inclusions). The 4th edition of "The Railroad, What it is, what it does" had a comment that the inclusions in rail were a common starting point for fatigue cracks, reducingthe number of inclusions in rail steel is one of the reasons that rail is lasting longer.

 - Erik

  • Member since
    July, 2008
  • 1,106 posts
Posted by rdamon on Monday, July 16, 2018 12:50 PM

Reduction of field welds have also been cited as well.

https://www.up.com/aboutup/community/inside_track/long-rail-3-27-2015.htm

  • Member since
    May, 2003
  • From: US
  • 14,861 posts
Posted by BaltACD on Monday, July 16, 2018 3:50 PM

rdamon
Reduction of field welds have also been cited as well.

https://www.up.com/aboutup/community/inside_track/long-rail-3-27-2015.htm

I might be mistaken, but so far it seems that UP is the only one dealing in 'long rail'.

         

Never too old to have a happy childhood!

  • Member since
    April, 2003
  • 302,135 posts
Posted by Anonymous on Monday, July 16, 2018 4:45 PM

I think you are right. From my point of view it is more a cost saving measure.

IIRC (it was more than 15 years ago) flash-butt welded rails have the same fatigue behavier as solid rail. Thermite welded rail has a higher risk of breaking at the weld because of the weld material differing from the rail material.
Regards, Volker

  • Member since
    December, 2001
  • From: Denver / La Junta
  • 9,323 posts
Posted by mudchicken on Monday, July 16, 2018 9:59 PM

Flash-butt welds are stronger than the surrounding rail.

Field welds have a greater chance of failure due to the less controlled environment, the human element and metalurgy of the field charge and molds.

Regardless, both are a major step up and above joints, bolts and angle bars in economic return and effectiveness. Also true is if you fail to maintain any of the above, you are on an accellerated road to failure. Shortlines and branchlines prove this out time and again. Many a shortline failed and abandoned because they had their priorities screwed-up.

Mudchicken Nothing is worth taking the risk of losing a life over. Come home tonight in the same condition that you left home this morning in. Safety begins with ME.... cinscocom-west
  • Member since
    March, 2011
  • 16 posts
Posted by dpeltier on Monday, July 16, 2018 10:05 PM

rdamon

Reduction of field welds have also been cited as well.

https://www.up.com/aboutup/community/inside_track/long-rail-3-27-2015.htm

 

 

The long rail sticks described in the article reduce the number of plant welds, not the number of field welds.

Anyhiw, most production-gang rail relays (which account for the vast majority of new rail consumption) result from rail wear, rather than from fatigue-related issues. In addition to better metallurgy and improved steel quality, rail life has improved due to better rail lubrication, more and better rail grinding programs, and newer rail sections that have more steel in the head of the rail (which allows for more grinding cycles and / or wear before the head gets too thin to live with).

On the other hand, while fatigue-related  defects in the rail don't account for a large footage of rail replacement, they do have a major impact on operations: service interruptions, slow orders, short maintenance windows to replace the section of rail containing the defect, and occasionally a derailment. Pretty much all rail defects are fatigue related, but there are different kinds of fatigue - or at least, there are different kinds of stresses that cause fatigue. Heavy cars or poorly supported ties cause the rail to bend like a bridge beam - but these stresses are usually pretty low compared to the contact stresses at the head of the rail, which is where the most common defects begin. Those contact stresses are affected by car weight but even more so by things like the shape of the wheel and the rail (both of which change as the steel wears), lubrication, and the car's suspension and steering.

Improvements that tend to reduce defects include cleaner steel for reduced crack initiation, but also better lubrication, wheel impact load detectors to remove the highest-impact wheels from circulation, and - again, critically - better grinding programs. Rail grinding does two things: it scrubs away those tiny surface cracks before they grow, and it keeps the rail in a shape that provides a favorable contact point with the wheel, which reduces contact stresses.

It is probably also true that, in some cases, there's a trade-off between reducing rail wear and reducing rail defects. Harder steels wear more slowly, but they are also more brittle, meaning cracks propagate faster, and in some cases they can result in higher contact stresses. TTCI has noted this latter phenomenon at locations such as switch points, where the normal shape of the rail doesn't present a very good match for a typical wheel profile. Softer rail just wears into a more conformal shape before the contact fatigue gets out of hand; harder rail does not.

Dan

  • Member since
    May, 2003
  • From: US
  • 14,861 posts
Posted by BaltACD on Tuesday, July 17, 2018 8:49 AM

         

Never too old to have a happy childhood!

  • Member since
    January, 2002
  • From: west coast
  • 135 posts
Posted by steve14 on Tuesday, July 17, 2018 10:38 AM

Bridges behave differently than track since they are more sensitive to fatigue issues, as Paul notes above. Cycles are important. A 5000 ton train of short wheel base cement cars is worse than a 5000 ton train of auto racks. More wheels can  mean more cycles for the bridge. A short span bridge sees more cycles than a long span. The long span can see a whole car (4 axles for example) as a single cycle where the short span sees each truck as a cycle.

Just read an interesting book that has a lot to do with the fatigue question, among other things, back before this was well anough understood. "Beautiful Railway Bridge of the Silvery Tay" by Peter R Lewis is a well written investigation into the Tay Bridge collapse in 1879 near Dundee Scotland. About 2000+ feet (10 spans) of the 2.75 mile long bridge fell during a heavy strom taking a whole train and about 75 people with it.

This was just at the transition from the use of cast iron and wrought iron to steel in railway bridge construction. There were several other key factors leading to the collapse but fatigue was one of the bigger ones. 

This led to the change in design which was used for the Firth of Forth bridge.

  • Member since
    December, 2007
  • From: Georgia USA SW of Atlanta
  • 8,888 posts
Posted by blue streak 1 on Tuesday, July 17, 2018 6:22 PM

BaltACD

The pictures certainly showed so many failures that we would not have known.  My original post was mainly looking for rail breaks that occurr under long trains where there was a sag in the track structure

 

  • Member since
    May, 2003
  • From: US
  • 14,861 posts
Posted by BaltACD on Tuesday, July 17, 2018 9:59 PM

blue streak 1
 
BaltACD

FRA Rail Inspection Guide

https://www.fra.dot.gov/Elib/Document/15669 

The pictures certainly showed so many failures that we would not have known.  My original post was mainly looking for rail breaks that occurr under long trains where there was a sag in the track structure

What rail fans acknowledge as a broken rail and the rails as they actually break are not the same.  But all kinds of breaks can derail trains.

A 'simple' stright through tangent track break of the ball, web and base of the rail can usually have a train 'walked' over it with a qualified individual observing the location of the break as the train passes over it, on my carrier both trackmen and signal maintainers were given training for making decisions on these kinds of breaks.  Other kinds of breaks and/or other locations of the broken rail will have their risks assessed by a qualified track inspector before a final decision is made to 'walk' the train over the broken rail.

Walking trains over broken rails is a operational expedient to keep traffic fluid until the necessary manpower and materials can be secured and moved into place to make some form of a longer term or permanent repair.  In most cases, welding in a 20 foot or longer 'plug rail' is the permanent repair.  Drilling the rail and applying joint bars is a short term repair, as on the territory I last worked that required at 10 MPH speed restriction at the location until such time as a permanent welded in repair is made.

The 'scary' aspect of broken rails is that in most cases broken rails are discovered after a train has operated through a territory and left a track occupancy light on the Dispatchers model board.  When this happens, on my former carrier, both Signals and MofW are notified to inspect the situation.  One, the other or both go to the scene and find the broken rail.  Since the track occupancy light was NOT ON prior to the train entering the track segment it is only reasonable to believe that the rail broke under the train as it pass.  Only by the grace of the almighty was there no derailment.

         

Never too old to have a happy childhood!

Join our Community!

Our community is FREE to join. To participate you must either login or register for an account.

Search the Community

Newsletter Sign-Up

By signing up you may also receive occasional reader surveys and special offers from Trains magazine.Please view our privacy policy