Sounds like a ‘Sun Kink’ is entering the possible cause list.
https://news.yahoo.com/investigators-seek-cause-deadly-montana-050015962.html
A sun kink seems most probable to me. The whole derailment has the look of what happens with a sun kink. The stored up stress in the rails is suddenly released by a passing train, thus causing the track to develop kinks, which become larger as the train passes over them. The track develops into a pattern of waves as reversing track curves in a serpentine pattern. The engines get past the danger zone because kinks are just beginning as the engines enter the danger zone. Indeed it is the locomotive that often starts track misalignment that begins to form the lengthening serpentine pattern of worsening track alignment.
The rest of the train follows the misalignment, but as the waves get larger, cars begin to derail. The sun kink waves produce the fish tail effect with the train. So the cause for derailment is not just one abrupt event with train or track. Instead, it is the development of a serpentine pattern of track that can be several hundred feet long. Then that entire pattern can start to derail cars as the kinks expand.
The most violent part of the derailment is at the end of the train if it reaches the pattern of kinks. The kinks are largest at the end of their pattern, and so the train fishtailing is the most violent at the end. That might explain why the last three cars apparently toppled over on their sides as they left the track, and still had enough toppling momentum to roll another 180 degrees to their opposite sides.
My problem with a sun kink is:
It was only 84F. Not exactly a record temperature. Two months earlier, temperatures were 10 degrees higher. Shouldn’t the derailment have happened then?
The derailment point, if it’s actually ahead of the switch, is on a curve. A curve provides a place for rail expansion–the curve bows outwards. The big problem with thermal rail expansion is on straight track, where there’s no chance for sideways relief.
It’s not that hot, and you have a curve to relieve the stress. I’m not goin’ for it, at the moment.
Here is a quote from BaltACD’s link:
“Did the switch play some role? It might have been that the front of the train hit the switch and it started fish-tailing and that flipped the back part of the train,” Clarke said.
That’s a theory I’ve been tossing out, though I’m backing down a bit on it, since I can now see the distances involved.
Sure is interesting.
Ed
Sun kinks are not confined to occurring only with the highest possible temperatures. It was 85, as I understand. Their occurence is also influenced by recently previous train traffic that can stretch or compress rail by train braking. Look at tha video I added.
RE: sun kinks…
It’s not just the ambient temperature at any given moment, you must also consider how rapidly the temperature has changed in the past day or so.
We had a derailment from a heat kink during the middle of winter a couple years ago. There was lots of snow and the ground was frozen, but a Chinook had blown in and the temperature had risen around 30 degrees in less than a day.
A heavy westbound loaded grain train pounded and destabilized a pre-existing soft spot defect in the track, but made it over without derailing, they didn’t notice anything unusual compared to previous trips so didn’t report it. The next train was an eastbound empty grain train, which hit the kink at nearly 50 mph. Their lead unit stayed on but the second unit derailed, fell over and took the train into the ditch.
A soft spot is not a heat kink.
A heat kink is based on the rails attaining a temperature “dramatically” higher than the ground they are sitting on, and the consequent differential expansion, and lateral displacement of the rails.
A “pre-existing soft spot defect” doesn’t fit that.
I am NOT saying this event cannot be a heat kink. I AM saying that it’s not been demonstrated that it’s likely.
Ed
Yes, it was. Is that statement supposed to convince me that there was a sun kink?
I looked. I read.
I didn’t see much train braking. It looked to be happily trundling along. So, unless you can demonstrate that train braking, I think we can cross that off.
I read suggestions that there was a heat kink. I read suggestions that the recently relayed trackage was crap and that the train was too heavy. It looked to me like the stuff hit the fan when those two or three huge containers that were in the middle of a long string of empty flats arrived.
I’ll go with the “huge containers” combined with the “relayed trackage” for $50, Johnny.
Since you are asserting above that the highest possible temperatures are not necessary, the references to temperature in the comments are irrelevant.
I see nothing above to convince me that it HAD to be a sun kink.
Ed
For the incident that I was describing, if you saw the footage from the lead locomotive’s camera and discussed the incident with the crew and company officials (as I did), you would also describe it as a heat kink. They all did.
The rails expanded and buckled due to the rapid rise in temperature, expanding faster than the ground (as you noted), and happened to buckle in that particular location (in the centre of a long straight section) due to the soft spot. That’s no different than any other heat kink, it occurred at the weakest spot.
I stand by my description.
I also recall another derailment which was caused by a combination of incomplete track work (tie changeouts) in a curve and the surface thawing and softening due to another dramatic, rapid rise in temperature (this one was in spring).
The Joplin wreck may have nothing to do with any of this.
If you discount the baggage car, half of the transition car, the diner and the lounge, the Builder had 6.5 cars for people to travel in. It had 141 passengers. So there were less than 22 people per “travel car.” Doesn’t seem like much. How many total people on the train would represent full capacity?
I’m not an MOW whiz, but I am aware that the track bed is an important factor when it comes to sun kinks. If the ballast can’t hold the track from lateral movement, the expansion of the rails may force them into such a kink.
I would opine that a pre-existing soft spot might well be a candidate for such a railbed failure.
In the past, some folks on here complained that the Talgo trains and other European designs aren’t passenger-safe compared to Superliners. Hmm.
As I mentioned, a sun kink can be gradually developed over several hundred feet of track as a train passes over it. It will develop by growing longer in the direction opposite of the train direction. So as the train passes, the cars toward the rear of the train will be encountering the worst track misalignment.
A sun kink has three components:
-
The compressing of rail lengthwise which makes the rail shorter and stores the compression like a spring.
-
The relative ability of the track bed to prevent the rail from buckling to dissipate the increasing rail compression.
-
The failure of the track bed to hold the rail against the rising rail compression; thus that failure causing the rail to buckle.
A lot more.
A Superliner coach has 78 seats upstairs and a variable number downstairs. Two fully loaded coaches would have exceeded the number of persons on that particular train. Four coaches would put you at a number greater than 312 (assuming at least one of them had lower level seats). I don’t recall the exact number, but a sleeper has a capacity of about half of a coach. Three and a half sleepers, call that 140 people. You’re already over 450 and you haven’t accounted for the crew yet.
Of course, this represents a scenario that I don’t believe is possible in the first place. I’m not sure there’s any route or even route segment of an Amtrak train that 100 percent of seats are occupied at a given moment. Looking over past accidents, accidents off the NEC seem to have 100 to 200 passengers at a the time and on the NEC have 200 to 300 passengers.
Braking would be initiated the moment the train came apart, unless the engineer thought something was amiss before that. The event recorder will tell.
I have been on both the Coast Starlight and Pacific Surfliners when there were standees due to lack of seats. To this day I don’t know how that happend on the all reserved Starlight, but the extras were hanging out in the lounge, standing in isles, etc.
The person in the interview several posts back said the entire Portland sleeper was sold out, and so he opted for the dorm-sleeper up front. The Portland sleeper was the last car on the train.
Ed
This is a battle Amtrak has been fighting for years. While the train may not have been full in that segment, there are likely some segments where it is near capacity - various city pairs, etc.
Critics would take that 141 passenger count and complain about underuse, ignoring the other variables.
The NTSB might beg to disagree, for instance, in RAR0302 discussing a 4-fatality Amtrak derailment in 2002 from a heat kink on a curve in Florida.
Well…
It wasn’t exactly a “heat kink”. Ambient temperature was in the low 80’s. In Florida. It’s kinda difficult to blame the problem on the high heat.
I did read a lot of it, and it is fascinating.
The failure(s) was in the track and roadbed. It was not done properly. They went ON and ON about all the details. I loved it! I did not know it was that complicated. It certainly demonstrates how important it is to have good track crews. The crews seemed to be good people, but with inadequate training and inadequate management and inadequate qulity control. And likely inadequate appreciation for their efforts.
Thanks much for finding this and presenting it! A great read.
Ed
From the report Chuck linked:
13.Had the two-way end-of-train device been activated when the Auto Train’s air brakes were put in emergency, the severity of the injuries resulting from the derailment might have been lessened, because the continued forward momentum of the majority of the train’s cars into the stopped passenger cars would have been reduced.
I don’t get it. It says the engineer waited about 15 seconds (after going into emergency) before pinging the EOT. It was a forty-car train, and I would have thought that the entire train would have been at full braking within about five seconds after going into emergency in the cab. Not so?