Trains.com

Oil Trains & Lag Screws

27527 views
426 replies
1 rating 2 rating 3 rating 4 rating 5 rating
  • Member since
    January 2014
  • 8,221 posts
Posted by Euclid on Tuesday, June 14, 2016 12:13 PM

AnthonyV
Just a few thoughts:

If friction between the plate and tie as a result fo the clamping force provides significant resistance to movement, would not the weight of the car provide a similar clamping force, enhancing the friction?

Once the lags get loose and act like pins, how does that differ from a spike?  Do plates fail with spikes as well?

Question:  Is there any indication which side of the rail the lag failures occur?  Gauge side or field side?  I would think that the screws on the inside would be subjected to shear, bending, and tensile forces where the screws on the ouside would be subjected just shear and bending.

I think the weight of the car weight would add a lot to the clamping force, although it would be of a pulsing nature as each wheel loads and unloads its weight on the plate.  Whereas, the lateral force of a curve that the clamping force is hoped to offset may be more constant, if it happens to be more spread out through the rail compared to the compression of the weight of the wheels.

Broken lags acting as loose pins differ from spikes in that the loose upper part of the broken lag is not as deep and tight as a spike.  But spikes can loosen and cause the same effect as the loose part of the broken lag, so the lag screws are intended to be an improvement over that shortcoming of spikes.  As I understand it, the U.P. Mosier line only uses the lag screws on curves.

I assume that lag screws typically break on the outside rail of curves.  I don't know if they tend to break more on one side of the rail versus the other sides.  I can see your point that screws on the outside of the rail would not provide much, if any, tensile force on the screws.    

 

  • Member since
    December 2005
  • 217 posts
Posted by AnthonyV on Tuesday, June 14, 2016 11:36 AM

Euclid

Overmod,

You are right that there is no cushion between the plate and the tie.  I gather that your comment on the lack of any spring loading to maintain friction of that joint was directed to the TSB.  They had said this:

“If the screws are tight on the rolled plates, the clamping force produces friction between the plate and the tie, providing lateral resistance to gauge-widening forces in the curve. If the screws are loose, they act as pins, providing the sole lateral resistance. When this happens, the flat bottom rolled plates move under load and can break (Photo 8). Failure of the screws is usually due to a combination of shear and bending, as all screws do not receive even loading.”

It sounds like they are saying that clamping force is maintained just by keeping the screws tight, notwithstanding your point that the slightest loosening of the screws would eliminate 100% of the clamping force because there is very little storage of that force.

Regarding the quotes you used in your earlier post that I mentioned, I have no problem with multiple quotes.  Here is what I was referring to: You began with a quote that I said, and attributed it to me.  Then you followed with several other quotes with no attribution identified.  Some of those were said by me and some were said by the TSB of Canada.  In the case of one or more of those that were said by the TSB, you responded as though I had said it.  Here is an example:     

Statement by the TSB that you quoted:

“Even with broken lag screws, there was little indication that the curve was under stress, as track gauge was maintained by the lag screws that did not break. The remaining intact lag screws experienced much higher lateral curving forces.”

Your response to the quote from the TSB:

“These are important points to remember, together.  It is not exactly a contradiction to what you said above, because what you meant was that there was little VISIBLE sign of the curve being under stress, a plurality of screws (whether still clamping or loose but acting as pins) holding the plates in rough gauge no matter how far they were deflected or bending.  Naturally, the fewer the fasteners taking up the particular vectors of load, the more failures were induced, leading to some acceleration of deterioration which -- unsurprisingly -- proceeded to the point of failure.” 

 

 ********************************************

There are things about the quote function of the forum that I do not understand.  With my computer, it leaves a lot of wasted space in the quoted material between multiple quotes from a stack of people when all extracted as one quote.  I find that if I try to close up some of that wasted space by editing it out, the quote structure loses its natural attribution labeling of who actually said the quote.  And of course, once the message is posted, there is no way to edit the quote in a way to re-insert that natural attribution because you cannot re-take quotes during an edit. 

 

 

Just a few thoughts:

If friction between the plate and tie as a result fo the clamping force provides significant resistance to movement, would not the weight of the car provide a similar clamping force, enhancing the friction?

Once the lags get loose and act like pins, how does that differ from a spike?  Do plates fail with spikes as well?

Question:  Is there any indication which side of the rail the lag failures occur?  Gauge side or field side?  I would think that the screws on the inside would be subjected to shear, bending, and tensile forces where the screws on the ouside would be subjected just shear and bending.

  • Member since
    September 2003
  • 21,669 posts
Posted by Overmod on Tuesday, June 14, 2016 10:38 AM

Euclid
It sounds like they [the TSB] are saying that clamping force is maintained just by keeping the screws tight, notwithstanding your point that the slightest loosening of the screws would eliminate 100% of the clamping force because there is very little storage of that force

Yes, and further, I think the TSB establishes (although they do not make the point directly in context) that if one screw loosens, it throws additional force on the others, including in directions the screws and their 'interaction zone' with the holes in the tie are not designed to withstand directly.

I remember a discussion of similar forces in conjunction with the wrought-iron rivets used in Titanic's construction.  While I do not want to encourage any comparison of modern lag-screw fabrication or 'quality' with wrought-iron riveting technology and its demonstrable shortcomings, I do see a number of applicable parallels in the assumptions of the "fabricating systems", and personally I experience a similar level of alarm looking at that situation.

 

Regarding the quotes you used in your earlier post that I mentioned, I have no problem with multiple quotes. Here is what I was referring to: You began with a quote that I said, and attributed it to me. Then you followed with several other quotes with no attribution identified.

That is just laziness or sloppy application of "proper" attribution semantics. In particular, with respect to:

[quote] ... Here is an example: Statement by the TSB that you quoted: “Even with broken lag screws, there was little indication that the curve was under stress, as track gauge was maintained by the lag screws that did not break. The remaining intact lag screws experienced much higher lateral curving forces.” Your response to the quote from the TSB: “These are important points to remember, together. It is not exactly a contradiction to what you said above, because what you meant was that there was little VISIBLE sign of the curve being under stress, a plurality of screws (whether still clamping or loose but acting as pins) holding the plates in rough gauge no matter how far they were deflected or bending. Naturally, the fewer the fasteners taking up the particular vectors of load, the more failures were induced, leading to some acceleration of deterioration which -- unsurprisingly -- proceeded to the point of failure.”]

the only excuse I can provide is that I was commenting on your having provided the TSB comment in your post, rather than your expressing the idea(s) in the TSB comment as your own ... it's a poor excuse.  Made even more so by the work I have to do on quotes inserted by the forum software ...

There are things about the quote function of the forum that I do not understand. With my computer, it leaves a lot of wasted space in the quoted material between multiple quotes from a stack of people when all extracted as one quote. I find that if I try to close up some of that wasted space by editing it out, the quote structure loses its natural attribution labeling of who actually said the quote. And of course, once the message is posted, there is no way to edit the quote in a way to re-insert that natural attribution because you cannot re-take quotes during an edit.

I have to do a relatively enormous amount of editing, especially with nested quotes, to get them to register properly and not contain excessive 'white space'.  I note, for example, that the 'close quote' tag is autospaced by a couple of carriage returns down from the end of the material that was quoted -- hard to explain this as anything but sloppy coding or poor choice of usability -- but with some care I can use delete and backspace to edge out most of the vertical space ... occasionally using the edit-undo feature if I go too far and screw up the weird default paragraph spacing my system renders.  (I also have to go through and re-insert all the little emoticon smileys that aren't properly rendered in quoted text, another telling little coding failure on tech's part, and I usually just delete them if I'm short on time or patience...)

Something that has relatively recently shown up in the circus programming is the use of some kind of tabs-and-backspace-set-quote-level thing when I quote something that already contains nested quotes.  Here even one delete or backspace 'too far' completely deranges how the quotes display, and of course there is no help or guidance for how that formatting works or was intended to work.  So I have to go through manually and insert all the quote and /quote tags properly, checking each time by "posting early" that the actual result renders correctly, and this takes time and -- occasionally, as in a post I made a couple of days ago -- the tags even though seemingly correctly formatted fail to produce actual quotes; they just sit there in undifferentiated text unrecognized.

Now, I will say this: on the system I am using (Firefox 47 on a Lenovo X220 running Windows 10) I get all the quotes in editable form when I respond to a post, so I can go in and clean up all the previous quote syntax if I want (and I try to do that, including correcting spelling errors or trimming some of the actual material -- some people, Juniatha for one, don't think that is ethical.)  So you will have to tell me what system you're using and what the problems with quote editing actually manifest as -- do you have screen captures that show what you're seeing?

I am evidently not smart enough, or gamer enough, to figure out the mystery-meat syntax for easy 'default' formatting of quotes without excess white space but with proper quote level preserved.  I also often find it tedious to insert the "user="" " tags into quotes to get the attributes properly attributized... although of course that's a nifty way to comment on sources, not just identify who originally said a particular thing.

Maybe tech will improve this ... but I think it is a relatively small priority for them, and in any case I would much rather have them concentrate on fixing the issues with improperly-crafted scripting or hangs in the ad-serving mechanisms for forum pages, or hunting down and disciplining advertisers using certain kinds of full-motion ads that are annoying...

  • Member since
    January 2014
  • 8,221 posts
Posted by Euclid on Tuesday, June 14, 2016 9:49 AM

Overmod,

You are right that there is no cushion between the plate and the tie.  I gather that your comment on the lack of any spring loading to maintain friction of that joint was directed to the TSB.  They had said this:

“If the screws are tight on the rolled plates, the clamping force produces friction between the plate and the tie, providing lateral resistance to gauge-widening forces in the curve. If the screws are loose, they act as pins, providing the sole lateral resistance. When this happens, the flat bottom rolled plates move under load and can break (Photo 8). Failure of the screws is usually due to a combination of shear and bending, as all screws do not receive even loading.”

It sounds like they are saying that clamping force is maintained just by keeping the screws tight, notwithstanding your point that the slightest loosening of the screws would eliminate 100% of the clamping force because there is very little storage of that force.

Regarding the quotes you used in your earlier post that I mentioned, I have no problem with multiple quotes.  Here is what I was referring to: You began with a quote that I said, and attributed it to me.  Then you followed with several other quotes with no attribution identified.  Some of those were said by me and some were said by the TSB of Canada.  In the case of one or more of those that were said by the TSB, you responded as though I had said it.  Here is an example:     

Statement by the TSB that you quoted:

“Even with broken lag screws, there was little indication that the curve was under stress, as track gauge was maintained by the lag screws that did not break. The remaining intact lag screws experienced much higher lateral curving forces.”

Your response to the quote from the TSB:

“These are important points to remember, together.  It is not exactly a contradiction to what you said above, because what you meant was that there was little VISIBLE sign of the curve being under stress, a plurality of screws (whether still clamping or loose but acting as pins) holding the plates in rough gauge no matter how far they were deflected or bending.  Naturally, the fewer the fasteners taking up the particular vectors of load, the more failures were induced, leading to some acceleration of deterioration which -- unsurprisingly -- proceeded to the point of failure.” 

 

 ********************************************

There are things about the quote function of the forum that I do not understand.  With my computer, it leaves a lot of wasted space in the quoted material between multiple quotes from a stack of people when all extracted as one quote.  I find that if I try to close up some of that wasted space by editing it out, the quote structure loses its natural attribution labeling of who actually said the quote.  And of course, once the message is posted, there is no way to edit the quote in a way to re-insert that natural attribution because you cannot re-take quotes during an edit. 

 

  • Member since
    November 2005
  • 4,190 posts
Posted by wanswheel on Tuesday, June 14, 2016 1:14 AM

From NTSB statement by Christopher O’Neil

“The NTSB recognizes the impact of this accident, or any rail accident involving hazardous materials, in environmentally sensitive areas and in proximity to residential, commercial and recreational areas, and understands the concerns of those affected…

“Upon notification of this accident we gathered information from the Federal Railroad Administration, the operator (Union Pacific) and local responders to understand the preliminary information related to the accident. Based on that information, the NTSB did not launch a full investigative team. However, because the accident involved the breaching of rail cars, including the CPC-1232 tank cars, we are actively collecting information from the FRA and the UP. Information collected will be assessed to determine whether further safety recommendations are necessary.”

  • Member since
    September 2003
  • 21,669 posts
Posted by Overmod on Monday, June 13, 2016 11:28 PM

I'm using the 'quote' function to cite references, whether in a previous post or an outside reference.  If that's irritating or confusing to anyone I can go back to using inline single or double quote marks, or type in a more specific 'user+"xxx"' tag to identify each reference when it comes up.

Euclid
If I am not mistaken, the plates in the CN wreck had elastomer cushion plates under them.

I think you are mistaken, but it is important to realize where the specific "mistake" is, and why.  Go back and look at a couple of descriptions of the Pandrol wood-tie system (and the UP lag-screw explanation) and see where the 'elastomer cushion plate' is located.

Early spring-clip track systems suffered from metal-to-metal contact and lack of shock attenuation -- there is no place for the reflected energy that makes the rail 'bounce' against the clips to go but into noise and vibration and so forth.  So a compliant 'damping' substance is put under the base of the rail, and sometimes between the clip and the top of the base, to absorb the running shock. 

It would make less than no sense to put the elastomer under the plate itself, where compression would instantly unload the screws.  The idea is to have the plate held down solidly whatever the forces in the rail itself, with the relatively elastic wood of the tie being all the 'elastomer' needed under the plate, and compression of the wood by the volutes of the lag screws providing the clamping force and preload to keep the plate bases firmly on the ties so they cannot slide.

If so, perhaps the compression of those plates [he means 'elastomer plates'] would act as a force accumulator to provide active compression that would maintain the friction to resist lateral movement of the plate on the tie.

That is a good thought in principle, but not the place you want the compliance, and of course far more expensive than putting it in the right place, e.g. between the plate and the underside of the bolt heads to keep the plate clamped down on a flat datum surface (and also in a different way between the rail base and the top of the plate to damp the shock between rail and the structure that elastically holds and aligns it)

[EDIT -- I seem to be getting more and more confusing in describing things so they make no sense to 'normal people'.  The bottom of the 'plates' -- rolled or cast -- needs to rest firmly on the ties, with the screws holding it down very firmly BOTH so that it can't shift laterally in any direction AND so that it can't lift or rock.  It might help to think of the design as re-creating the effect of seats and anchors in concrete ties, where any relative motion is a Bad Thing but very careful damping and shock reduction needs to exist between the rail and the clips and seats that hold it in line and gauge and cant.]

 

  • Member since
    January 2014
  • 8,221 posts
Posted by Euclid on Monday, June 13, 2016 8:36 PM

Electroliner 1935

 Also, has anyone learned about the cause of the bolt failure?

As I understand it, the screws fail from fatique cracking, I suppose from a combination of tension and bending force.  From what I see, this U.P. derailment has the same cause as the CN derailment in 2012, yet U.P. is acting like the cause of the Mosier wreck was unprecedented. 

It also seems as though the cause of the CN wreck was unanticipated because they had no previous experience with breaking lag screws.  I would have thought that the CN wreck would have been recognized throughout the industry for revealing this screw breaking problem and the inability to detect it.  So I am puzzled that this has apparently now caught U.P. off guard.

  • Member since
    January 2014
  • 8,221 posts
Posted by Euclid on Monday, June 13, 2016 8:16 PM

Overmod,

You brought up some points in the tenth post from the top of the preceding page which I was going to respond to, but I am a little confused.  In some of the quotes there that you respond to, you appear to be responding to me.  But those quotes are what I quoted from the TSB report of the CN wreck.

I definitely understand your point about the lag screws not being able to provide clamping much force unless they were to have spring washers or some means to store energy.  It was the TSB that said the plates lost some of their ability to resist lateral movement due to the loss of clamping pressure, which was the result of the screws breaking.  If I am not mistaken, the plates in the CN wreck had elastomer cushion plates under them.  If so, perhaps the compression of those plates would act as a force accumulator to provide active compression that would maintain the friction to resist lateral movement of the plate on the tie.     

  • Member since
    September 2010
  • 2,515 posts
Posted by Electroliner 1935 on Monday, June 13, 2016 7:23 PM

Another question, my experience with most tie plates was that they had ridges that would dig into a tie and help hold it in place on the tie, Earlier, someone spoke about rolled plates. Do these have anything other than the bolts to maintain their position? Also, has anyone learned about the cause of the bolt failure? Is it corrosion, wear, pulled apart (under tension) or shear (from side forces like the plate cutting it). The UP slide indicated that the bolt was in place but the top was separated from the bottom. Thus the rail had NO horizontal restraint nor vertical but the weight of the train. Horizontal force causes the rail to slide to the outside. If all the force is on the bolt, it could overload it and cause it to shear. This is the failure but I don't think they have stated the bolts failure mode or cause. Two crashes indicates that stronger bolts may not be the solution. 

 Are these Pandrol systems better?

I just found one answer in Euclid's post of the Canadian report.

Photo 6. Cast plate (left) and rolled plate (right) 


Photo 7. Bottom of plates (Note: lugs on bottom of cast plate [left] and smooth bottom of rolled plate [right]) 

So the rolled plate cannot provide any horizontal restraint to the structure and ALL the horizontal forces are on the bolts. I think this is a significant factor.

  • Member since
    May 2003
  • From: US
  • 25,292 posts
Posted by BaltACD on Monday, June 13, 2016 7:01 PM

[quote user="zugmann"] 

Euclid
In any case, what I find most interesting is that this pileup appears to have been preceded by a relatively long derailed-dragging event. If so, there was plenty of time to stop the train if the derailment had been detected. 

1500' = 25 60' cars.  Not very long at all, especially if the train was moving at anything above a snail's pace.  And that's if the rail just overturned at one small spot, and not for a few car lengths. 

PS: don't put any type of "expert" tag on me.  Cause I sure as heck ain't one./quote]

30 MPH = 44 feet per second
1500 feet = 34.1 seconds

 

Never too old to have a happy childhood!

              

  • Member since
    January 2002
  • From: Canterlot
  • 9,575 posts
Posted by zugmann on Monday, June 13, 2016 6:18 PM

Euclid
In any case, what I find most interesting is that this pileup appears to have been preceded by a relatively long derailed-dragging event. If so, there was plenty of time to stop the train if the derailment had been detected.

1500' = 25 60' cars.  Not very long at all, especially if the train was moving at anything above a snail's pace.  And that's if the rail just overturned at one small spot, and not for a few car lengths.

 

PS: don't put any type of "expert" tag on me.  Cause I sure as heck ain't one.

It's been fun.  But it isn't much fun anymore.   Signing off for now. 


  

The opinions expressed here represent my own and not those of my employer, any other railroad, company, or person.t fun any

  • Member since
    December 2012
  • 310 posts
Posted by Cotton Belt MP104 on Monday, June 13, 2016 11:59 AM
I have a suggestion: each post/poster can be self-identified as expert= active railroad employee non-expert= “foamer” …….. (I can say that, ‘cause I are one!) Observation: There are some threads on this website, that surely could be used as examples in a class on human relations/psychology/sociology. Fact: As a retired teacher, we were taught, if a student is trying to disrupt a class with some small noise that is not noticeable from where it is coming ……. Try “extinction”. That is to say, do NOT pay attention, don’t react, let it go. Sure it works. Student will not get the attention and tire of the activity. Too it can be so annoying to fellow students that they take care of it w/o a word said. They all look at you know who, guilty person stops, teacher has never stopped teaching. Comments/theory/idea/suggestions get under your skin???? Chill, ignore, you can’t/won’t stoop to a low level slam if that post is ignored. Endmrw0612161153
The ONE the ONLY/ Paragould, Arkansas/ Est. 1883 / formerly called The Crossing/ a portmanteau/ JW Paramore (Cotton Belt RR) Jay Gould (MoPac)/crossed at our town/ None other, NOWHERE in the world
  • Member since
    January 2014
  • 8,221 posts
Posted by Euclid on Monday, June 13, 2016 11:04 AM

Overmod
 

What remains to be seen is whether there is 3/8 of a mile of fully 'widened-out' gauge between the original point of failure (which was the original few broken bolts that allowed the first wheelset to drop) and the point where the bulk of the accordioning is observed -- which is how I thought this accident was being described -- or if the accordioning started 3/8 of a mile 'behind' the point of derailment (as the derailed part "dug in" and the rear part's momentum started running in the slack...)  Until I get that clarified (and it may well be known; it's just obfuscated by lousy news coverage and what passes for press releases) I'll withhold further comment.

 

I don't follow what you are saying in the above quote. 

Incidentally, my comment that Dave replied to above was just an observation of what I assumed would have happened.  It was not intended to prove anything.  It assumes that the gage widened and dropped a wheel inside of its rail.  That is the way it is described by U.P., and it it is also what is shown in their graphic.  I don't recall anyone from U.P. saying that the rail tipped over, so I assumed that it just shifted sideways.  What I am suggesting is that the wheel fell off the rail in the wide gage, and then acted sort of like a can opener and extended the wide gage all the way to the point of pileup. 

In any case, what I find most interesting is that this pileup appears to have been preceded by a relatively long derailed-dragging event.  If so, there was plenty of time to stop the train if the derailment had been detected. 

  • Member since
    September 2003
  • 21,669 posts
Posted by Overmod on Monday, June 13, 2016 12:04 AM

schlimm
However, you seem to be acting as some self-appointed judge of the postings. I'm not aware of any announcement by the forum moderators of your new reponsibilities.

Good point.  Assuredly no one died and left me in charge of some truth police. And it does not matter to the actual topic of discussion even if they had, and I were.

I now see that nobody's going to figure out how the single-bolt tale got started, either; I'm not sure whether quoting Chaos News bot coverage of a media story reporting a press release that has misconceptions in it is sophisticated sarcasm, but it sure is too complex for me to comment any further on.

One note to the latest controversy between Euclid and Dave: in some cases it should be possible for a wheelset that has dropped into a short wide-gauge 'hole' to pop back up -- into guiding, or maybe into derailment 'over the high side'  -- if there is enough lateral restoring force exerted by the rails on the wheelrim faces and outer tread corner.  But it should be noted that each following wheelset will be dropping down into the same 'hole' and tending to lengthen the wide spot or turn over a rail, and the resistance 'felt' by the parts of the train both before and behind that area will likely be increasing.

What remains to be seen is whether there is 3/8 of a mile of fully 'widened-out' gauge between the original point of failure (which was the original few broken bolts that allowed the first wheelset to drop) and the point where the bulk of the accordioning is observed -- which is how I thought this accident was being described -- or if the accordioning started 3/8 of a mile 'behind' the point of derailment (as the derailed part "dug in" and the rear part's momentum started running in the slack...)  Until I get that clarified (and it may well be known; it's just obfuscated by lousy news coverage and what passes for press releases) I'll withhold further comment.

  • Member since
    January 2014
  • 8,221 posts
Posted by Euclid on Sunday, June 12, 2016 10:27 PM

Electroliner 1935

I still wonder what the specified sheer strength of the bolts is and what tests have been done to confirm that the bolts met that spec. There was back in the 80's if I recall a rash of venders supplying bolts that didn't meet specifications. 

 

I think this derailment could lead to a big controversy about strength of materials, engineering, and overall approach.  And if this basically duplicates the CN wreck, it will raise even more questions. 

It does open the door to the development of a lag screw testing machine.  All that is needed is a pull on each screw to make sure that it has not separated due to breakage.   

The beauty of the cut spike is that it always tells you what it is doing.

  • Member since
    September 2010
  • 2,515 posts
Posted by Electroliner 1935 on Sunday, June 12, 2016 10:11 PM

I still wonder what the specified sheer strength of the bolts is and what tests have been done to confirm that the bolts met that spec. There was back in the 80's if I recall a rash of venders supplying bolts that didn't meet specifications. 

  • Member since
    January 2014
  • 8,221 posts
Posted by Euclid on Sunday, June 12, 2016 10:04 PM

dehusman
 
Euclid
I would like to know what happened right after a wheel dropped into the gage.

 

The train derailed.

 
Euclid
Bear in mind that at the start of this 1500 ft. of travel, when the wheel dropped off the rail, the opposite wheel is still running on the rail in the proper position (although tipped out of perfect rail contact).

 

And how would you know that?  Once one rail tips over there is nothing restraining the wheels to stay on the other rail and the "derailed" wheel is now lower than the other wheel so gravity is pulling things towards the side of the track where the rail is turned over.  Plus ther is more drag on the derailed wheel which slows that side of the truck down causing the truck to "steer" towards the turned over rail.

 
Dave,
 
Do you have a refernece that says the rail tipped over?  I recall that was the case with the CN wreck, but I don't recall it said about the Mosier wreck.
  • Member since
    September 2010
  • 2,515 posts
Posted by Electroliner 1935 on Sunday, June 12, 2016 10:02 PM

dehusman
Bingo, the derailed wheel is tearing stuff up on the ties until bends the rail enough that the rail breaks or the wheels dig in and then stuff piles. up.

Exactly!

  • Member since
    September 2003
  • From: Omaha, NE
  • 10,621 posts
Posted by dehusman on Sunday, June 12, 2016 9:43 PM

Euclid
I would like to know what happened right after a wheel dropped into the gage.

The train derailed.

Euclid
Bear in mind that at the start of this 1500 ft. of travel, when the wheel dropped off the rail, the opposite wheel is still running on the rail in the proper position (although tipped out of perfect rail contact).

And how would you know that?  Once one rail tips over there is nothing restraining the wheels to stay on the other rail and the "derailed" wheel is now lower than the other wheel so gravity is pulling things towards the side of the track where the rail is turned over.  Plus ther is more drag on the derailed wheel which slows that side of the truck down causing the truck to "steer" towards the turned over rail.

Euclid
I doubt that the gage had widened over this entire 1500 ft. of distance. You would be considering thousands of screws being broken if that were the case.

Dude, the wheels of the cars are between the rails.  Unless you think that gauge is between 5 and 6 feet, I gaurantee the guage was compromised.

Once the wheel hits the tie it begins tearing things up.  It will shove the rail sideaways or turn the rail over.

Euclid
There would be a big squeeze contest between the derailed wheel shoving against the inside of its rail, and the flange running normally for the wheel still on the rail on the opposite side.

No.  No contest whatsoever.  The wheels win.  Once the rail turns over only one flange might be in contact with the rail.  There is no "squeezing", squeezing requires a force from both sides and there is nothing to provide the force from one side.

Euclid
If that did not happen, then the derailed wheels, as the gage closed to normal, would displace their rail off of its position on the ties. The first derailed wheel would shear off all of the hardware connecting the rail to each tie, one-by-one.

Bingo, the derailed wheel is tearing stuff up on the ties until bends the rail enough that the rail breaks or the wheels dig in and then stuff piles. up.

Dave H. Painted side goes up. My website : wnbranch.com

  • Member since
    January 2014
  • 8,221 posts
Posted by Euclid on Sunday, June 12, 2016 8:33 PM

I would like to know what happened right after a wheel dropped into the gage.  It would seem this derailment happened on the inside of the curve at East Lake (north side) as the train would naturally crowd the outer rail (south side).  For various reasons, I speculate that the derailed wheel ran in line, but on the ties for some considerable distance, such as 1500 ft. for instance.

This running would be between the point where the first wheel derailed to the point where the pileup and fire occurred.  Bear in mind that at the start of this 1500 ft. of travel, when the wheel dropped off the rail, the opposite wheel is still running on the rail in the proper position (although tipped out of perfect rail contact).

I doubt that the gage had widened over this entire 1500 ft. of distance.  You would be considering thousands of screws being broken if that were the case. 

More likely, the gage widened in a rather short distance of say a few hundred feet; say at the curve at East Lake.  Then a wheel fell off of the rail on the inside of its rail where the gage widened. 

It seems to me that the next wheel to follow would also drop off the rail in the same spot.  Each wheel in succession would drop off the rail, so that many cars would be running with wheels on the derailment side all running on the ties, and pressing against the inside of their rail.

But, in a relatively short distance, the widened gage would close back to normal.  Where do those derailed wheels go then? 

There would be a big squeeze contest between the derailed wheel shoving against the inside of its rail, and the flange running normally for the wheel still on the rail on the opposite side. I would guess that the derailed wheels pressing against side of their rail would possibly cause a reaction force that would cause the opposite wheel flange to climb the rail, and derail to the outside of that rail. 

If that did not happen, then the derailed wheels, as the gage closed to normal, would displace their rail off of its position on the ties.  The first derailed wheel would shear off all of the hardware connecting the rail to each tie, one-by-one.   

  • Member since
    January 2014
  • 8,221 posts
Posted by Euclid on Sunday, June 12, 2016 7:18 PM

http://www.cbsnews.com/news/railroad-reveals-cause-of-fiery-oregon-oil-train-derailment/

One broken bolt that missed detection sounds better than say 75 broken bolts.  But I see that the report of ONE broken bolt can either minimize the problem or accentuate it depending on how you look at it.  Quote from the link:

“Mosier Fire Chief Jim Appleton, who fought the blaze after the derailment, said he appreciated Union Pacific's maintenance and safety procedures but the risk of one broken bolt resulting in such a disaster or potentially worse disasters means regulators should not allow shipments of crude oil to travel by train through the area.”

  • Member since
    July 2006
  • 9,610 posts
Posted by schlimm on Sunday, June 12, 2016 6:06 PM

Overmod
Overmod wrote the following post 1 hours ago: schlimm Picky picky. Faulty trackwork is accurate and my post was one week ago. ' Faulty trackwork' is indeed accurate, and you are to be commended for posting the statement a week ago (when Euclid was still off on the UDE 'conspiracy theory').  However, the problem I had was with the end statement in your post earlier today, which I did not quote Sometimes even us non-experts have better hunches than the resident rail experts here. I took that as both a 'dig' at Euclid's presumption to be a 'rail expert' and as a claim that your hunch as to the technical details of the faulty trackwork were 'better' than his.  And here, arrogance is not in order.

I was not making a dig at Euclid; I suggest you try reading more accurately and in context.  I was not dealing with specifics, just a hunch that the crash's cause was trackwork, not brakes.  I never pretended to be an expert anymore than he did/does.  I don't know if you are an actual expert or a railroader or just another non-expert observer.  However, you seem to be acting as some self-appointed judge of the postings.  I'm not aware of any announcement by the forum moderators of your new reponsibilities.

C&NW, CA&E, MILW, CGW and IC fan

  • Member since
    November 2005
  • 4,190 posts
Posted by wanswheel on Sunday, June 12, 2016 5:19 PM

Euclid
schlimm
wanswheel
Euclid

 a failed fastener is referring to ONE fastener. 

Excerpt from KATU Portland, Jun. 11

http://katu.com/news/local/officials-multiple-broken-fasteners-led-to-mosier-oil-train-derailment-explosion

Union Pacific officials told KATU News Friday night they believe multiple loose fasteners on the track led to the crash.

So now we have the UP spokesperson (who must be getting this from UP experts) saying what some of us non-experts have been saying?

Yes, us non-experts can cover a lot of ground quickly with our freedom to speculate.

  • Member since
    September 2003
  • 21,669 posts
Posted by Overmod on Sunday, June 12, 2016 4:33 PM

schlimm
Picky picky. Faulty trackwork is accurate and my post was one week ago.

' Faulty trackwork' is indeed accurate, and you are to be commended for posting the statement a week ago (when Euclid was still off on the UDE 'conspiracy theory').  However, the problem I had was with the end statement in your post earlier today, which I did not quote

Sometimes even us non-experts have better hunches than the resident rail experts here.

I took that as both a 'dig' at Euclid's presumption to be a 'rail expert' and as a claim that your hunch as to the technical details of the faulty trackwork were 'better' than his.  And here, arrogance is not in order.

For better or worse, Euclid did 'discover' both the type of construction that uses "bolts" to secure trackwork and the precedent in Canada for that type of trackwork to fail.  That was corroborated by later objective news reports, one of the exhibits used in the Mosier community meeting, and statements by UP "expert" Jacobs.

On the other hand, none of the 'better hunches' in your actual post a week ago is even close to the actual cause of the derailment -- not sun kinks, not broken ties, certainly not loose spikes as there were no spikes in the faulty trackwork at all. 

It's the last statement that causes the 'pickiness', not the fact that you were correct in a broad sense back a week ago before Euclid figured things out for a while.  I now see he is back on the 'single bolt' hypothesis -- which is based only on improper comprehension by Oregon officials in a series of press releases, as far as I can tell; UP made it clear at the meeting that it was a fastener SYSTEM that caused the problem, and has repeatedly discussed how multiple fasteners had to fail to produce the dropped wheelset... etc.   So perhaps it was a fluke that he figured out something right, against considerable mockery and opposition as he did so.  But in all fairness, you essentially mocked him for getting it right, and implied you were better at understanding the specific faulty trackwork when you were not.  That was the source of the pickiness.

  • Member since
    July 2010
  • From: Louisiana
  • 2,310 posts
Posted by Paul of Covington on Sunday, June 12, 2016 4:32 PM

   This thread is getting more and more entertaining.

_____________ 

  "A stranger's just a friend you ain't met yet." --- Dave Gardner

  • Member since
    July 2006
  • 9,610 posts
Posted by schlimm on Sunday, June 12, 2016 3:54 PM

Overmod
Overmod wrote the following post 4 hours ago: schlimm Euclid was all over the place with his theories. Not to boast, but I posted this on page 1 of this thread: Posted by schlimm on Sunday, June 05, 2016 1:52 PM I would only add that the derailment and release of oil may have had nothing to do with any form of braking if its cause was faulty trackwork: summer? kinks? or loose spikes/broken ties, etc.

But you did not provide any relevant detail of the 'faulty trackwork' itself, and in the event none of the causes you actually speculated were accurate.

Picky picky.  Faulty trackwork is accurate and my post was one week ago.

C&NW, CA&E, MILW, CGW and IC fan

  • Member since
    January 2014
  • 8,221 posts
Posted by Euclid on Sunday, June 12, 2016 3:41 PM

 

tree68
 
Euclid
Incidentally, the reports from U.P. say that one or more lag screws failed and caused the wreck.  So there is still an indirect reference to ONE bolt failing.  I don't know how there can be any doubt that would necessitate the quantification of "one or more." 

 

Because it overs all the bases.  Why are you obsessing over the "one" part?  This was, by all appearances, a lag screw failure.  Not "one" lag screw failure, but a failure of lag screws securing the track in place.

 

My concern over the mention of one or more lag screws failing is that they either know that it was only one; or they know that it was more than one.  They have even said it was several.  So why say "one or more"?  I am sure that they have counted them by now, so why don't they tell us what the number was?  If they need to cover their bases, they can use the word "approximately" ahead of the number.

I don't believe this amounts to a message referring to a lag screw as meaning a class of lag screws, although that misunderstanding has also entered the discussion as seen in the article where experts doubt that one broken lag screw would derail the train.  They based their whole argument on the literal interpretation of "one" as meaning one. 

As clarification of that point, the article that I cited yesterday said this:

 

MOSIER, Ore. – A broken lag bolt on the track caused an oil train to derail near Mosier on June 3, Union Pacific officials said in a town meeting Friday night.

 

  • Member since
    December 2001
  • From: Northern New York
  • 25,022 posts
Posted by tree68 on Sunday, June 12, 2016 3:25 PM

Euclid
Incidentally, the reports from U.P. say that one or more lag screws failed and caused the wreck.  So there is still an indirect reference to ONE bolt failing.  I don't know how there can be any doubt that would necessitate the quantification of "one or more." 

Because it overs all the bases.  Why are you obsessing over the "one" part?  This was, by all appearances, a lag screw failure.  Not "one" lag screw failure, but a failure of lag screws securing the track in place.

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
    September 2003
  • 21,669 posts
Posted by Overmod on Sunday, June 12, 2016 2:26 PM

A couple of comments in line -- note that some of this is my own opinion but I think it is factually based; correct me if otherwise.

Euclid
This U.P. Mosier wreck appears to have an identical cause to that of the CN wreck.

I would not make a blanket statement.  Both wrecks appear to have lag-bolt issues (in a Pandrol-style spring-clip system) as contributing factors.  But as you note (I'll say a bit more in a moment) there appears to be a key difference: there were clear warning signs of deterioration at Fabyan Bridge, but none -- despite the presence of several kinds of reasonably alert maintenance -- at Mosier.

Furthermore, according to Jacobs the 'proximate cause' of the derailment was a gauge widening only great enough to drop one wheelset.  One of the things to emerge from the FRA investigation will be just how many 'lag-screw failures' would be needed to permit lateral excursion and/or roll of one rail relative to the other sufficient to drop the wheelset entirely within the gauge.  I expect that will be more than 'one or two plates' worth'.  I do not really expect the problem to be as widespread as the 74 documented failures that triggered the actual derailment event on CN, but (as noted) I'm beginning to worry that comparatively small initial events may lead to catastrophic failure in ways that other track securement systems might not.

As I understand it, there the only way to inspect for this problem with lag screws is to check for gage widening as a precursor to a catastrophic failure of lag screws.


The test is not for 'gauge widening' per se, it has to be for gauge widening upon application of lateral or overturning force -- remember that the geometry train didn't catch anything 'within the past 30 days', so passive measurement of track geometry is most likely not producing whatever effect caused the Mosier derailment.  The CN test by contrast not only detected a number of gauge widenings before the accident, but typified them (as your citation indicated) as 'near-urgent'.

 

If the screws are tight on the rolled plates, the clamping force produces friction between the plate and the tie, providing lateral resistance to gauge-widening forces in the curve.

Note that this is one of the places the design fails some basic mechanical design -- where you are relying on fastener tension to provide lateral resistance between two smooth faces, some elastic provision like Belleville or lock washers needs to be provided to keep the surfaces clamped.  In this system, the elastic clamping between tieplate and rail base does nothing for tie securement, and anything that creates even slight space between tie face and rolled plate bottom may cause shifting of the plate over against the shank of the screw.  This might both deflect the metal and distort the edge of the hole in the tie, admitting water and contaminants.  Even a small amount of looseness in the lags -- from whatever cause -- reduces the effective clamping force to zero, and at that point the loading on the screws becomes as you note, but there is more to the issue.  The clips are still providing spring action to give elastic shock absorption to the rail, except now the effect is to magnify the shock between the tieplate and the shoulders on the lags, a kind of 'hammer blow' that increases asany of the 'clearances' increase.

 

 

... the flat bottom rolled plates move under load and can break (Photo 8). Failure of the screws is usually due to a combination of shear and bending, as all screws do not receive even loading.

My immediate concern with the 'rolled-plate' version of the system was that, unlike the cast plate, there were no features on the bottom of the tieplate that provided lateral 'keying' against lateral movement.  (I did not think the features present on the cast plates were very substantial or shaped well, but at least they were present.)  Note the picture that shows polish and scratching across the entire base of a plate!  You'd best believe this shows not only substantial lateral movement with contaminants present, but shows the movement persisted for a substantial length of time.  I suspect there was a design assumption along the general lines that having four lags per plate would prohibit all but a 'clearance' amount of lateral shift, and relative tolerance of rolling force, hence no perceived need for keys on the underside of the plate.

 

In this occurrence, a number of lag screws in the vicinity of the POD broke off in the tie. The breaks, which occurred over a considerable period of time, were due to fatigue at the transition point between the shank and the threads, where the cross-sectional area is reduced.

I am not fully convinced that this is the sole reason for breakage, or that there is some structural design defect in that area of a standardized lag screw.  On the other hand, I have seen precisely this kind of problem completely ignored by engineers (in the crankpins of the original Precor elliptical exercise machine).  In both cases, interestingly enough, a stress-originated failure was complicated by multiple stress-raisers forming at points of corrosion originated by 'foreign materials' in a wet environment

 

 

Even with broken lag screws, there was little indication that the curve was under stress, as track gauge was maintained by the lag screws that did not break. The remaining intact lag screws experienced much higher lateral curving forces.

These are important points to remember, together.  It is not exactly a contradiction to what you said above, because what you meant was that there was little VISIBLE sign of the curve being under stress, a plurality of screws (whether still clamping or loose but acting as pins) holding the plates in rough gauge no matter how far they were deflected or bending.  Naturally, the fewer the fasteners taking up the particular vectors of load, the more failures were induced, leading to some acceleration of deterioration which -- unsurprisingly -- proceeded to the point of failure. 

I consider it instructive to note that the reported prompt failure at Fabyan Bridge here was not a lateral gauge failure (as alleged at Mosier so far) but a rolling failure, where the vertical clamping integrity of the lags was lost for a sufficiently long span to allow the railhead to roll out and dump wheelsets into the gauge.  Whether or not the subsequent effects of the wheelsets forcing rails apart laterally was the same in both cases will probably not be determined until the FRA investigation is reasonably far along. 

 

... there is still an indirect reference to ONE bolt failing.  I don't know how there can be any doubt that would necessitate the quantification of "one or more."  In the CN wreck, 74 lag screws failed.

I don't think there is a way to drop a wheelset into the gauge without some combination of plates being loose, screws deflecting laterally, and screws pulling up vertically, perhaps on both sides and not just the more-heavily-loaded rail in a curve.  Again, it will probably be wise to wait for more specific detail from forensics before making more speculation about particular mechanisms.

Remediating loose fasteners may be as complex as trying to test them for looseness or corrosion.  Spikes can be easily pulled and redriven, and their holes plugged easily; tension on lag screws can't be determined properly by backing and retightening them (any more than torque-to-yield bolts can be 'tested' by torquing them backward) and any tensioning can only be done by putting a proper rotational tool on them -- there is no inspection of the shank, threads, material, etc. without physically unscrewing them, and that is likely to damage any tie wood in directions other than optimal for vertical clamping ... repeatedly.

  • Member since
    September 2003
  • 21,669 posts
Posted by Overmod on Sunday, June 12, 2016 1:12 PM

Overmod, please refrain from citing me within gratuitous criticism of other people.

 

Exactly where and how do you allege I have done this?  Tell me where and I will take it out of a secondhand quote.

You were mentioned in my posts only only to establish a timeline, as a matter of historical record .. or are you now asserting that you didn't officially post something at a particular time if you don't like how it's subsequently used to substantiate or illustrate something?

Have I supposedly misquoted anything in one of your provided links, or referred to you in any way that attributes an opinion to you that you have not yourself first expressed?  I try very carefully to avoid that, and I have a long track record of objectivity as an editor and director in so avoiding it.

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