It's hardly just Pandrol plates. many new tie plates manufactured after about 1995, including plain old cut spike plates are flat bottomed (no ridge created during rolling, unless the railroad specifies) . Largely done to avoid the problems with rail cant and adverse tie abrasion. Adz-Pads / abrasion pads are typically placed under the tie plate for the same reason, to avoid adzing a trough in the top of the tie.
Electroliner 1935 Good one Zug. Does anyone know whether these bolts are installed with a torque limiting wrench? It strikes me that they could be overstressed on installation if too much (Value to be determined) torque was applied and the point of failure would be at the top of the threads. This could increase corosion at the stress point. And the fact that the plates have no ridges would increase the horizontal forces impressed on the bolt causing the failures.
Good one Zug.
Does anyone know whether these bolts are installed with a torque limiting wrench? It strikes me that they could be overstressed on installation if too much (Value to be determined) torque was applied and the point of failure would be at the top of the threads. This could increase corosion at the stress point. And the fact that the plates have no ridges would increase the horizontal forces impressed on the bolt causing the failures.
While one would not want to overtighten these it must be remembered that it is a steel screw in WOOD tie. Excess tension would just strip the wood threads. Tension, or tightness, if that makes more sense to you, is not an issue. The tie plates are moving laterally , not up and down. This ain't rocket surgery. The railroads used tie plates with the embossed bottom for at least 50 years and suddenly Pandrol has "improved" them by being flat and slick?
UC Bearcat Guy: Almost all are installed with pneumatic production machines with torque limiters or later on out of face with a hand drill with the torque limiter. (really no difference hydraulic/ electric/pneumatic for what has to handle that big a screw - and with a production gang and that machinery required...)... adzpads or something similar is almost standard with those types of pandrol plates or McKay/DE-Clip plates (as if hairpin spikes weren't already enough of a headache, this comes along)
Lewis has plants in Minneapolis and La Junta, and the stuff is machine cut. The next fun and games may well be to see if the necking problem is tied to the US made coach screws or the foreign made (largely Chinese) screws looking at the Q/C marks on the three quarters of the drive head. UPRR is probably dragging out purchase orders and supporting required test results from the independent Q/C certified labs looking for a pattern.
Can I point out that Norm just screwed the whole thread?
Sorry.. I couldn't resist any longer.
Norm48327
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
http://www.pandrol.com/wp-content/uploads/Track_Report_11.pdf
Apparently Pandrol has better salesmen than website people. The link describes a "cast swaged shoulder". There is no such animal. Then it describes the process where heated metal is struck by a die. That is not casting, where molten metal is poured into a mold to freeze nor is it swaging which is a cold forming process. That describes forging or hot heading. Further research shows the bottom of one of these plates and it is dead smooth, slick as a babies butt. That allows the plate to move on the tie. Plates that I am familiar with had a diamond pattern embossed on the bottom to grip the tie when weight is applied. I think pandrol and anyone who bought this song and dance has screwed up badly.
Per their web site it is hot headed. This is like forging but done in one strike. The heated material is held in a die that forms the threads and shank and struck by a die that forms the head, all in one operation. Very common for cheap low strength fasteners.
Found the above image at Lewis Bolt and Nut Co. They've been making fasteners for the railroads since 1927. Looking at the threads, they appear to be cut rather than rolled (rolled threads are stronger than cut threads) and where the threads stop at the shank looks like a weak spot.
Comments from professional bolt makers welcome.
Norm
wanswheelHow often are lag screws replaced? UP rep said Mosier track was lag-screwed in 1999. If the one or more that broke was/were original, maximum life is 17 years?
I think the premise 'at the time' was that the lags would outlast head-hardened rail ground at something like the 'magic wear rate', and be pulled (and cleaned and probably replaced without individual NDT inspection) en masse when the rail was turned or replaced. While I might not bet large sums of money that this assumption was not revisited when bainitic rail steels, etc. were being examined for adoption, I would not at all be surprised to find that it had not been.
The issue in any case is not that the lags would not screw out or pull 'proud' in service; it's that once they are not clamping, they experience stresses I don't think were modeled or analyzed correctly. Fabyan Bridge is sort of an extreme case in that many, many lags were known to be simultaneously loose, and the overall track geometry determined by a significant measure was verging on a need for critical intervention. So it's not surprising that what turned out to be a grossly unsafe condition progressed to catastrophe without obvious further warning -- what's new at Mosier was that similar catastrophic progression could result with good-faith inspection twice a week and conscientious attention to geometry and metallurgical rail integrity. It is only 'icing on the cake' that much of the "good faith inspection" as conducted would not detect precisely the most critical type of lag failure...
I have not seen any information that indicates that breakage of lag screws was ever anticipated. It was known that they would unscrew. In the TSB report of the CN wreck at Fabyan Bridge, it said that partially unscrewed lag screws were easy to see because their heads were standing higher than normal. So inspectors turned those screws back down tight. But the broken lag screws gave no visual indication, so nothing was done about them during the tightening of loose screws.
The lag screws at the Fabyan Bridge site were installed in 1997 and the wreck happened 1/21/2012. So that is probably around 14 years of service. The report mentions that fatigue cracks began “many years” prior to the wreck.
How often are lag screws replaced? UP rep said Mosier track was lag-screwed in 1999. If the one or more that broke was/were original, maximum life is 17 years?
Technology only moves forward with disasters attributed to the current technology. Creeping technology.
One failure is a fluke. Two failures indicates that the 1st wasn't a fluke and it is time to hit the design boards again. This is the path that humankind has taken to get to today and it is the path that will be followed into the future.
Improved technology always finds the weakest link in existing technology and then that gets improved and we find the next weakest link.
Never too old to have a happy childhood!
I am wondering if this breaking lag screw problem is just a matter of growing pains in an evolution from cut spikes to screw spikes. Or is the problem perhaps now a thing of the past due to better fastening systems making the lag screw obsolete?
I do not know the answer to that. But whatever the answer, I am surprised that the problem was able to sneak up on the CN in 2012 and cause the disaster at Fabyan Trestle. Then after being thoroughly revealed to the industry in the subsequent TSB instigation, this lag screw problem still managed to sneak up on the U.P. in the Mosier wreck. Finally, after all this warning, the U.P. is checking all their lag screws.
The fundamental advantage of screws is that they hold better than nails. In construction, most of the loading is static, but rail fastenings are subject to dynamic loading under the passing trains. Apparently this dynamic loading and the resultant fatigue cracking has been underestimated as a colossal engineering error, only to be revealed by train wrecks. It would be interesting to learn how that miscalculation happened.
We have now learned that the holding power of the screw in a tie exceeds the screw’s ability to resist the dynamic loading over time. Can this deficiency be eliminated simply by making the screw stronger? Could the problem be solved by making the screw larger, such as 1.25” major diameter?
What about the threads themselves. The deep “V” threads of lag screws and their transition to the unthreaded shank create stress risers that that spawns the birth of fatigue cracks that grow over time and use. Could the geometry of the threads and shank be made more streamlined to reduce stress risers sufficiently?
Perhaps a better approach would be to use a rounded and shallower thread like a bottle thread, and use more of them at a finer pitch. That way, it leaves more material for a stronger shank in the threaded area. What if the hole in the tie were tapped with the proper threads to achieve a more precise relationship between the internal threads of the tie and external threads of the screw?
The goal should be to never have a screw break under any circumstance.
A sufficiently improved track screw will cost more, and some might say it will not be cost effective, but how cost effective are conventional screws that cause train wrecks?
In order to remove the burden of train wrecks from conventional screws, it requires costly regular inspection. Maybe that money should be spent on better screws.
AnthonyV Overmod I don't think anyone has implemented sprung-clipped track with spikes, at least not successfully www.pandrol.com/wp-content/uploads/Track_Report_11.pdf
Overmod I don't think anyone has implemented sprung-clipped track with spikes, at least not successfully
I don't think anyone has implemented sprung-clipped track with spikes, at least not successfully
www.pandrol.com/wp-content/uploads/Track_Report_11.pdf
I appear to be wrong, repeatedly. Note the discussion of how the plates are fabricated from standard AREMA.
I noted also that some of the plates in R12E0008 have rectangular holes in them, which I understood to be used for a purpose like temporary spiking to position while the lags were driven. It might be interesting to find a discussion of why those holes are provided where they are.
In the process of searching for rail clip deflection data, I found this Pandrol document that discusses various projects, including a CSX project that uses clips and spikes (Page 22).
I found it interesting that the rail is held in place with both clips and spikes.
It's not terribly difficult to differentiate between a recent break in a bolt and one that's been there for a while.
The question would be whether enough "forensic" information was preserved in the effort to re-open the line.
Inspection of the bolts beyond the reaches of the damage from the wreck will provide a lot of information as well - or maybe it already has.
Larry 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...
wanswheel[Jason Rea, chief engineer for the western region of Union Pacific Railroad said] The railroad doesn’t know how many were sheared before the derailment, but some were sheared after a wheel was derailed.
I would say that some broke over a long period of time leading up to the derailment. Then one or more broke as the final triggering event for the cause of the derailment. And finally one or more broke after the derailment began.
As I mentioned earlier, dragging the derailed car or cars for 3/10 of a mile between the initial point of derailment and the final resting place could have broken several screws on each tie.
Euclid Norm48327 What is the (supposed) advantage of the lag screws? https://en.wikipedia.org/wiki/Rail_fastening_system “The screw spike has a higher cost to manufacture than the rail spike, but has the advantage of greater fixing power - approximately twice that of a rail spike.”
Norm48327 What is the (supposed) advantage of the lag screws?
What is the (supposed) advantage of the lag screws?
https://en.wikipedia.org/wiki/Rail_fastening_system
“The screw spike has a higher cost to manufacture than the rail spike, but has the advantage of greater fixing power - approximately twice that of a rail spike.”
MC posting under two screen names now?
Overmod AnthonyV My question was in reference to the comment noted above. Have you gotten a good enough answer yet?
AnthonyV My question was in reference to the comment noted above.
Have you gotten a good enough answer yet?
What are you talking about? I was responding to Tree68's comment that we were not talking about loose lag screws, when in fact there was discussion of this.
AnthonyVMy question was in reference to the comment noted above.
tree68 AnthonyV Once the lags get loose and act like pins, how does that differ from a spike? It's been my impression here that the problem is not loose lags - it's broken lags. A lag bolt without a head would have near zero holding power on the track structure. A loose spike may not be able to prevent a rail from rolling over, but would likely still help hold the track in gauge. If the lags are failing due to corrosion, the damage to the metal may reach into the tie. If that is the case, a broken lag may have no presence above the top of the tie. And, as there is apparently no current method for inspecting the lags, a section of track may well be almost literally hanging on by a thread so that when a few bolts give way and begin the process of allowing the rail to move, other bolts may follow in rapid succession. Then there comes the maintenance angle. If a spike is coming out, I can drive it back in, or pull it out, drive in a plug, then redrive a spike into the tie. Problem solved. I would opine that dealing with broken lags may well require complete replacement of the tie, unless the tie place can be moved over to where new lags can be driven.
AnthonyV Once the lags get loose and act like pins, how does that differ from a spike?
It's been my impression here that the problem is not loose lags - it's broken lags.
A lag bolt without a head would have near zero holding power on the track structure.
A loose spike may not be able to prevent a rail from rolling over, but would likely still help hold the track in gauge.
If the lags are failing due to corrosion, the damage to the metal may reach into the tie. If that is the case, a broken lag may have no presence above the top of the tie. And, as there is apparently no current method for inspecting the lags, a section of track may well be almost literally hanging on by a thread so that when a few bolts give way and begin the process of allowing the rail to move, other bolts may follow in rapid succession.
Then there comes the maintenance angle. If a spike is coming out, I can drive it back in, or pull it out, drive in a plug, then redrive a spike into the tie. Problem solved.
I would opine that dealing with broken lags may well require complete replacement of the tie, unless the tie place can be moved over to where new lags can be driven.
Euclid 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.
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.
My question was in reference to the comment noted above.
So; a question for Mud Chicken.
Why the lag screws. I have photos of a class one near me doing a curve rail I believe is six degrees or more and they used rolled tie plates and Pandrol clips. That was in '06 and they just redid it this year. What is the (supposed) advantage of the lag screws?
Excerpt from The Dalles Chronicle, Jun. 14
Jason Rea, chief engineer for the western region of Union Pacific Railroad, described at a community meeting Friday in Mosier what had caused the June 3 derailment of 16 oil cars.
The so-called lag screws, which are threaded, are used on curves instead of a straight track spike. And while the lag screws had been severed about two and a half inches below the head of the screw, the top of the screw did not dislodge. Had they dislodged, it would have been detected by a visible inspection, Rea said.
Rather, the sheared screw or screws remained in place.
“I don’t know of any that it has ever happened to,” Rea told the Chronicle after the meeting. “I’ve never experienced this kind of derailment.” He said he’s seen multiple derailments in his many years with the railroad.
The lag screws were implemented in 1999, he said.
Each rail tie has eight spikes or screws in it. The spikes or screws – four on each end — hold in plates that secure the rail to the tie.
The railroad doesn’t know how many were sheared before the derailment, but some were sheared after a wheel was derailed.
The wheel derailed about 3/10 of a mile east of where the crash actually occurred. Technically, the derailment is where the wheel leaves the rail, and the crash site is called the point of rest.
When the derailment site was inspected, Rea said some of the lags didn’t pop off, but some did.
Overmod The solution I developed involves special tools that are applied to one or more spindles of a machine that drills holes to plug ties in the field.
Overmod,
That sounds very interesting. Did you invent that and develop it as a product that is being sold to the industry? It sounds well thought out and useful.
tree68I would opine that dealing with broken lags may well require complete replacement of the tie, unless the tie place can be moved over to where new lags can be driven.
The solution I developed involves special tools that are applied to one or more spindles of a machine that drills holes to plug ties in the field.
If necessary, a tool is inserted to gauge any deflection or binding of the broken lag stem in the hole, and straighten it as vertical as possible.
A tool that is basically a tube saw with relief for 'saw chips' is then inserted into the hole, and cuts out an annular recess with the broken stem in the center, self-aligning over the metal piece as it goes. Air can be blown in to clear if desired. When drilled to appropriate depth, the stem piece is cross-fractured and lifted out of the hole (probably with a magnet rather than a gripping tool).
A second tool can be inserted to provide 'undercuts' or waves in the wall of the hole, so that the material of the subsequent plug will expand and key into the tie wood more positively.
The plug itself is made out of stronger 'engineered wood' than an ordinary spike plug, and can of course be made as a tube with preboring of appropriate size to expand it into the above-mentioned keying when a new lag screw is inserted. It will probably be driven a bit more than flush with the seating surface so as not to produce a false clamping surface, but I welcome informed opinion that would prescribe otherwise. I expect the normal sort of preservative and then adhesive, fiber-swelling material (cf. 'Chair-Loc'), "wood hardener", etc. that is used with plugging to be applicable here. Grain and binder is such that the 'joint' between the bore and the inserted lag is reasonably watertight even upon some lateral deflection of the bolt shank with age.
Of course ordinary treated-wood plugs of the appropriate 'oversize' could be driven in, and then drilled in the usual fashion for installing the plate fixation system in the first place,
Based on the experience of the 2012 CN derailment, here is what the TSB says to do in order to inspect for this problem:
When conducting inspections of curves, turnouts, and bridge decks with screw spikes, please be governed by the following:
AnthonyVOnce the lags get loose and act like pins, how does that differ from a spike?
These are reasonable thoughts.
AnthonyVIf 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?
Yes, and it is possible to calculate the length of rail near or over the plate where this clamping force becomes substantial.
Part of the problem is that you should remember that the spring clips always provide a 'perfect' elastic connetion in tension between the plate and the rail base, just as they do when concrete ties are used. But only the inertia of the plate itself -- which is not very much -- actually resists any forces other than compression that the rail exerts on the tie.
And there are vertical accelerations in tension, both due to force transmitted from the train's motion to the rail as a beam, and to the acceleration of the rail as the train's weight comes off the increased loading. If you watch even well-maintained CWR track it's not unusual to see the whole structure move up and down repeatedly as the weight comes on and off. There is an upward acceleration on the lags each time the rail rises back up and the springs carry the plate along... although, of course, the load on the rail never comes anywhere near zero (which would be an open invitation to a different sort of catastrophe altogether!)
Perhaps a better way to make the point I think you're intending is that it is precisely at the times of highest increased wheel load that you'd be getting lateral flange force trying to displace the plate sideways against friction, bending lag shanks, etc. So yes, I'd expect this to help against a 'shift under the train'
AnthonyV Question: Is there any indication which side of the rail the lag failures occur? Gauge side or field side?
Once the lags get loose and act like pins, how does that differ from a spike? Do plates fail with spikes as well?[/quote]
I don't think anyone has implemented sprung-clipped track with spikes, at least not successfully -- for reasons you can easily imagine if four substantial lag screws are specified for necessary plate retainment! You'd want to go back and look at the history of systems that used lag-screw type primary retention (I think France being an example of a country that used that method of fixation) to see what the maintenance intervals, failure modes, etc. were as compared to driven spikes.
That noted, part of the issue is how little a 'lifted' spike still deflects laterally, compared to a bolt with a circular shank and perhaps sharp edges on its threaded 'retaining' surfaces, and the degree to which the rail itself would move in spiked track (relative to the hard contact of its seat in the tieplate) rather than lever the whole tieplate itself with a longer lever arm.
Some of this was covered in the TSB report Euclid linked, but I do not think that by the time the accident occurred and the TSB actually investigated the conditions it was possible to determine where the 'action' initially started. Once a lag is loose, gauge or field side, it is likely to aid in loosening others - it certainly won't help relieve any forces that would tend to cause a derailment, as far as I can see...
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.
Assuming we are talking about the 'loaded' outside rail in a curve, I would tend to agree, except there may be some (relatively slight) tension on the field side if the actual rotation of the plate takes place around its outer edge, rather than the 'fulcrum' of a pinned connection due to lateral contact with the plate shifting outward. Remember that some of the force involved may be the result of energy 'stored' in the spring action of the clips, or transmitted via shock for a railhead impact that accelerates through spring compliance and then transfers momentum when the clip action 'bottoms' more or less abruptly.
Our community is FREE to join. To participate you must either login or register for an account.