How Hot Is Too Hot?

[tree68]

Perry Babin

Why aren't expansion joints installed in rails to compensate for expansion/contraction?

The rails have a tendency to sag at the joints, even with a tie under the joint and proper tamping.  This can lead to harmonic rocking (for us it's about 17 MPH), which if severe enough, can actually throw cars off the tracks.

Back in the day, section crews were out daily watching their ~10 miles of track and making adjustments.  That's why NYC999 could run 112 MPH on jointed rail.  

That level of attention no longer exists.

[BaltACD]

tree68

 

Perry Babin

Why aren't expansion joints installed in rails to compensate for expansion/contraction? 

The rails have a tendency to sag at the joints, even with a tie under the joint and proper tamping.  This can lead to harmonic rocking (for us it's about 17 MPH), which if severe enough, can actually throw cars off the tracks.

Back in the day, section crews were out daily watching their ~10 miles of track and making adjustments.  That's why NYC999 could run 112 MPH on jointed rail.  

That level of attention no longer exists.

Back in the days of jointed rail, B&O had a restriction on Hi-Cube (4700 cuft+) covered hoppers - trains handling them on 6 degree or greater curves had to operate slower than 12 MPH or greater than 25 MPH through such territory.  The harmonic rocking of the rail joints coupled with the truck center distances would rock cars off the track if operated in the restricted speed range.

[azrail]

Texas has always had hot summers. The same with the deserts of Calif/Arizona, where they always exceed 110 during the hot season. Yet we hear of no heatrelated rail damage/sun kinks in those areas.

[CSSHEGEWISCH]

I would assume that the neutral temperature mentioned in earlier posts is higher in the southwestern deserts than elsewhere.  This would probably lessen but not eliminate sun kinks and related damage.

[BaltACD]

CSSHEGEWISCH

I would assume that the neutral temperature mentioned in earlier posts is higher in the southwestern deserts than elsewhere.  This would probably lessen but not eliminate sun kinks and related damage.

Note the rail expansion chart I posted in page 1 of this thread.

[Euclid]

Perry Babin

Why aren't expansion joints installed in rails to compensate for expansion/contraction?

 

 

They do use expansion joints.  Here are some images: 

 

https://www.google.com/search?newwindow=1&rlz=1C1GGRV_enUS748US748&sxsrf=APwXEdfIs6pCi-G9SHwKgDrvIdUvxObZtg:1687705573035&q=railroad+track+expansion+joints&tbm=isch&sa=X&ved=2ahUKEwj856mZ2d7_AhWLDzQIHSPjCs8Q0pQJegQICBAB&biw=1920&bih=933&dpr=1

 

Prior to welded rail, they relied on gaps in the rail at the joints.  References indicate that included an option to lubricate the angle bars, and I suppose limiting their bolt torque, but I have never seen the practice in use. 

 

I have seen what appeared to be relatively large expansion joints on the Rat Hole Division of the Southern Ry.  They were large, mostly enclosed housings maybe 8ft. long, and spaced at distances of at least a mile or much more.  I have never seen any photos of them, but at the time, I concluded they were expansion joints.  They might have been directed to use with welded rail, but maybe the practice never gained popularity.   The point is that welded rail is much longer than jointed rail, so an expansion joint for welded rail must allow for a much greater amount of movement.  Whereas stick rail is shorter, so its thermal movement has more frequent expansion joints, so the movement for which they compensate is much shorter. 

 

I understand that the current, preferred method of accommodating expansion movement is not to allow it to occur.  Everything is tight and the thermal movement is absorbed by the rail as it yields internally, so it cannot change the length of the rail.  The necessary resistance to prevent the rail expansion and contraction is obtained by better securing the track against buckling or stringlining that would otherwise be produced by the rail elongation and contraction. 

[Perry Babin]

Where this is only rarely a problem, I didn't think it would require many expansion joints. 

I found one of the joints that was shown on your google search (the one on concrete sleepers) that looked like it would be about as mainenance free as any other railroad joint (switches, diamonds...). 

I've seen anchors installed on rails, tightly against both sides of the ties, to help prevent movement. 

[Overmod]

The anchors keep the rail from moving longitudinally in what is known as 'creep' -- repetitive braking combined with shock reducing the spike clamping and tieplate friction so the rail can move cumulatively in the direction of braking.

You'll see people claiming that the clamping action of Pandrol holds the rail against expansion and contraction due to temperature, and in the next breath claim that rebound spring tension in that design eliminates spike pulling.  Be interesting to see what professionals like diningcar and MC think of elastic fixation.

HSR has very carefully designed expansion joints that preserve the very precise track geometry needed for very high speed.  As far as I know these have been part of LGV design since the '70s.

[Euclid]

Perry Babin

Where this is only rarely a problem, I didn't think it would require many expansion joints. 

I found one of the joints that was shown on your google search (the one on concrete sleepers) that looked like it would be about as mainenance free as any other railroad joint (switches, diamonds...). 

I've seen anchors installed on rails, tightly against both sides of the ties, to help prevent movement. 

 

There are two basic approaches to accommodating rail expansion.  Only some of those images show expansion joints.  They look like two switch points that are closed against the two stock rails; which just stub-end after passing the points. The points are not intended to open or close like a switch.  Instead, they allow slippage between the points and the stock rails.  That slippage accommodates the expansion/contraction.  There several variations of that concept shown as well. 

 

The rail anchors you mention are part of the system that controls expansion by physically constraining the rails to prevent expansion or contraction.  To constrain the rails, the rail anchors grab the rail base and also bear against the sides of the ties.  So in order for the rails to get longer, they have to move all of the ties sideways to the line of track.  Then the ballast that is tamped in between the ties makes the ties even harder to move in line with the track.  All of the track support factors work together to prevent the rails from changing length due to expansion or contraction.  So with this approach there is no actual expansion joint because expansion is prohibited. 

[jeffhergert]

azrail

Texas has always had hot summers. The same with the deserts of Calif/Arizona, where they always exceed 110 during the hot season. Yet we hear of no heatrelated rail damage/sun kinks in those areas.

 

Unless one causes a problem, like a derailment, you probably won't hear of them.  Track defects that disrupt traffic happen all the time but they don't make the evening news.  Many don't even make it to or through the railfan grape vine.  Most are found and fixed without mishap.  In most cases, the worst thing is delaying trains.  If it causes a major delay Amtrak might make note of it.

Jeff

 

[dpeltier]

Perry Babin

Where this is only rarely a problem, I didn't think it would require many expansion joints. 

On what basis do you think that?

If you cut a well-anchored rail in cold weather, you will see the cut end of the rail snap back as the tension is relieved and the rail contracts. If you use match marks to measure exactly how far back this contraction ends, you will see that the answer is "not very far at all", because the anchoring system develops the tensile forces pretty effectively over a short distance. Similarly, an expansion joint only changes the stress state of the rail in a very localized vicinity. Several hundred feet away, it makes no difference whatsoever to any of the forces that can cause thermal misalignments (or, in the winter, broken rails).

I found one of the joints that was shown on your google search (the one on concrete sleepers) that looked like it would be about as mainenance free as any other railroad joint (switches, diamonds...). 

Joints are not maintenance-free at all. And switch points, which are structurally quite similar to these joints, are far more maintenance-intensive than regular rail joints.

Also, in that photo, there is another set of expansion joints maybe 500' feet away, which gets back to my first point. Since I don't imagine the TGV actually has a dozen sets of expansion joints per kilometer, it's probably safe to assume that these are needed for some special purpose, probably related to the fact that the tracks are on a bridge.

In North America, you will see expansion joints almost exclusively at movable bridges, where the rail has to have free ends in order for the bridge to move, and therefore can't be completely restrained.

Dan

[Perry Babin]

dpeltier

 

On what basis do you think that?

 

 

Ignorance. That's why I'm here (and asking stupid questions), to learn. 

[dpeltier]

Overmod

You'll see people claiming that the clamping action of Pandrol holds the rail against expansion and contraction due to temperature, and in the next breath claim that rebound spring tension in that design eliminates spike pulling.  Be interesting to see what professionals like diningcar and MC think of elastic fixation.

Elastic fasteners in concrete ties or that attach to a timber tie using screw spikes or other similar hold-down fasteners are very effective at preventing relative motion between the rail and the tie in both the longitudinal and vertical directions. They also prevent the rail from canting outward relative to the tie. Rail anchors are not needed in this case.

In a concrete tie, degradation of the concrete under the rail seat can cause a loss of clamping force and allow all three kinds of relative motion to occur. Degradation of a wood tie under a Pandrol plate leads to loss of clamping force between the plate and the tie, making them act more like Pandrol plates fastened to the tie with spikes. (It is possible, however, that degradation of the rail seat with screw spikes may lead to some funny stresses in the plate itself - occasionally I will hear stories about excessive plate cracking in Pandrol plates, and this seems as likely an explanation as anything else I can think of.)

Pandrol plates that are attached to the rail with cut spikes instead of screw spikes are still quite effective at transferring longitudinal loads to the tie and remove the need for rail anchors, but the spikes can see some pretty high shear stresses in the process. See https://railtec.illinois.edu/article/investigation-into-the-effect-of-lateral-and-longitudinal-loads-on-railroad-spike-stress-magnitude-and-location-using-finite-element-analysis/ for a discussion of how this may contribute to broken spikes and wide-gauge problems. Pandrols with spikes are not as effective at transferring vertical (uplift) forces as Pandrols with screw spikes. As far as rail cant goes, Pandrols with spikes probably act similar to curve blocks: by causing the plate to rotate along with the base of the rail, they reduce the amount of cant, but they don't distribute the resulting couple moment forces over the plate / tie interface as effectively as screw spikes.

Dan

[tree68]

It's not unusual for contracting rail to cause a "pull-apart," where the bolts in a joint are literally sheared off.  It can even happen in jointed rail.   That's when you see "fire snakes" used to heat the rail so replacement bolts can be inserted into the existing bolt holes in the joint.

Broken rails in colder climes are often the result of a heretofor unrecognized flaw in the rail and the contraction of the rail from the cold.

Unless the gap is huge, trains may be allowed over the broken point very slowly, and under observation, until the problem is resolved.

[Euclid]

I don’t get the point of elastic fasteners connecting rail to ties, thus allowing rail to get longer while remaining securely connected to the ties.  Obviously heat expanding rail can produce a very significant amount of rail elongation.  It can buckle the track enough to move it completely off of the roadbed. 

 

So the rail can move in relation to the fixed ties.  How can the rail move in relation to its continuous fixed length?

[Overmod]

Euclid

I don’t get the point of elastic fasteners connecting rail to ties, thus allowing rail to get longer while remaining securely connected to the ties.

Go back and read more carefully what Dan Peltier posted.

One point he makes is that the clamping force provided by all the clips (and the friction between the base of the rail and the pad produced by the clamping force) is sufficient to keep the rail from moving longitudinally.  Presumably the expansion in the rail length between two adjacent points of fixation is controlled by those two points, so the aggregate length of a stretch of rail is held.

This bears out why we often see 'sun kink' as just that, a single "squirm".  

Where there is extended bending (as I recall one example was on UP after they tried using asphalt as a 'moisture barrier' under the ballast, which didn't go at all well when heat permeated to that level...) what you have is longitudinal shock allowing the longitudinal expansion of the rail to push ties preferentially sideways.  Again, if the rail were free to move, the result would be a long bow in one direction or the other, come to think of it like the rail on a rail train going around a curve.

[Euclid]

Okay, I guess the elasticity has nothing to do with providing elasticity to allow the rail to move.  Instead it is a means to store an active compression force in the clamping of the fasteners to rail and ties.  In other words, the fasteners are spring loaded to maintain their clamping pressure.   The end objective is to prevent the rails from moving longitudinally.