Passenger brakes on the passenger cars. Lousy tread brakes and probably no graduated release on a great many top-heavy car-carrying freight cars. We have those here who can tell what braking a train like that under emergency conditions,
Electroliner 1935 adkrr64 Lithonia Operator 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? I don't understand that either. My understanding is that a brake application propagates through the train at about the speed of sound. And each car has a vent valve to help speed the propagation of an emergency application through the train. So the last car should have been emergency long before 15 seconds, even on a relatively long consist like the Auto-Train. We don't use EOT on our RR, so maybe there is something there I am not aware of. Facts are Facts. According to the event recorder information, the engineer initiated a service brake application at 5:07:57 p.m. He initiated an emergency brake application at the head end of the train at 5:08:01 p.m. By 5:08:02 p.m., the trainline pressure at the head of the train was 98 psi, the rear-end trainline pressure was 108 psi, and the train speed was 55 mph. At 5:08:04 p.m., the trainline pressure at the head of the train was zero and the rear-end trainline pressure was 108 psi. At 5:08:05 p.m., the pressure at the rear of the train dropped 2 psi to 106 psi and remained at 106 psi for 7 seconds. At 5:08:11 p.m., the EOT went into emergency. The trainline pressure at the rear end dropped to zero at 5:08:12 p.m. At 5:08:15, the engineer activated the EOT device. (See appendix E for more detailed event recorder information. And after reading the Autotrain report, I looks very much like a similar case except the track failure may have occurred after the engines pased over it. Has any loco video been released?
adkrr64 Lithonia Operator 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? I don't understand that either. My understanding is that a brake application propagates through the train at about the speed of sound. And each car has a vent valve to help speed the propagation of an emergency application through the train. So the last car should have been emergency long before 15 seconds, even on a relatively long consist like the Auto-Train. We don't use EOT on our RR, so maybe there is something there I am not aware of.
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?
I don't understand that either. My understanding is that a brake application propagates through the train at about the speed of sound. And each car has a vent valve to help speed the propagation of an emergency application through the train. So the last car should have been emergency long before 15 seconds, even on a relatively long consist like the Auto-Train. We don't use EOT on our RR, so maybe there is something there I am not aware of.
Facts are Facts.
According to the event recorder information, the engineer initiated a service brake application at 5:07:57 p.m. He initiated an emergency brake application at the head end of the train at 5:08:01 p.m. By 5:08:02 p.m., the trainline pressure at the head of the train was 98 psi, the rear-end trainline pressure was 108 psi, and the train speed was 55 mph. At 5:08:04 p.m., the trainline pressure at the head of the train was zero and the rear-end trainline pressure was 108 psi. At 5:08:05 p.m., the pressure at the rear of the train dropped 2 psi to 106 psi and remained at 106 psi for 7 seconds. At 5:08:11 p.m., the EOT went into emergency. The trainline pressure at the rear end dropped to zero at 5:08:12 p.m. At 5:08:15, the engineer activated the EOT device. (See appendix E for more detailed event recorder information.
And after reading the Autotrain report, I looks very much like a similar case except the track failure may have occurred after the engines pased over it. Has any loco video been released?
So one has to wonder: why not simply wire it so that when you go into emergency it pings the EOT? How hard would that be? When would you not want the EOT to dump simultaneously?
I never would have dreamed that it would take ten seconds for a forty-car train to fully go into emergency. But given that, with both ends dumping simultaneously, I guess it would have taken five seconds to get the job done. Right?
Then you have to wonder: if the train stopped twice as suddenly, what kind of accidents would have been caused by that in itself?
Still in training.
7j43k, I got this site up on Google Earth and it seems the original curve has been modified to a less sharper curve from the original construction; but does not appear to be recent. Visually my eye says the train had exited the curve before derailing. But of course there may be a different analysis by the NTSB inspection.
adkrr64Lithonia Operator 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? I don't understand that either. My understanding is that a brake application propagates through the train at about the speed of sound. And each car has a vent valve to help speed the propagation of an emergency application through the train. So the last car should have been emergency long before 15 seconds, even on a relatively long consist like the Auto-Train. We don't use EOT on our RR, so maybe there is something there I am not aware of.
With that small of a train, you probably couldn't even get your hand from the automatic brake handle to the EOT dump switch before the rear end was at 0.
Now hitting it is a good reflex (just in case something got crimped off), but I think the NTSB was really reaching on that 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
Lithonia OperatorI 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?
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.
ChuckCobleigh 7j43k 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. 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.
7j43k 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.
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
7j43kThe 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.
Lithonia Operator 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?
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?
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.
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...
NittanyLion Lithonia Operator 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 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.
Lithonia Operator 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
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.
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.
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.
EuclidIt would interesting to know how braking was executed during this range. How much of the train was on the ground before braking was initiated?
Braking would be initiated the moment the train came apart, unless the engineer thought something was amiss before that. The event recorder will tell.
Lithonia OperatorSo 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
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.
It looks like the locomotive and first cars remaining on the rails stopped about 500 feet or so after passing over the switch. I suspect the actual sun kink range with the three components in this case is possibly over 1000 ft. long and extends approximately from the switch location, back in the direction opposite the train direction.
During the passage of the train through the sun kink range, the actual first two components I have listed above were in existence, but showed no visible signs to the engineer. But the locomotive itself began initiating component #3 as it ran over the range of #1 and #2.
Therefore each car of the train encountered worse track alignment at each point in the entire sunk kink range. So even though the two engines and 2 or 3 head end cars passed entirely through the range, the track they encountered was at a constant degree of varied misalignment that was insufficient to derail the that equipment.
But behind that head end equipment, the track alignment was bad enough to cause derailment directly; or to cause the equipment to swerve enough to add force to the growing track misalignment.
By the time the last cars progressed into the range, the track was heavily damaged by the serpentine kinks sufficiently to flip the last three cars off over in a 360-degree rollover.
So, I speculate that this whole process began when the train entered the sun kink range 1000-2000 feet short of the switch. It ended with the head end still on the rails approx. 500 ft. past the switch.
It would interesting to know how braking was executed during this range. How much of the train was on the ground before braking was initiated?
In the past, some folks on here complained that the Talgo trains and other European designs aren't passenger-safe compared to Superliners. Hmm.
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.
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.
Greetings from Alberta
-an Articulate Malcontent
Euclid Sun kinks are not confined to occurring only with the highest possible temperatures. It was 85, as I understand.
Sun kinks are not confined to occurring only with the highest possible temperatures. It was 85, as I understand.
Yes, it was. Is that statement supposed to convince me that there was a sun kink?
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.
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.
SD70Dude RE: sun kinks..... We had a derailment from a heat kink during the middle of winter a couple years ago. A heavy westbound loaded grain train pounded and destabilized a pre-existing soft spot defect in the track...
RE: sun kinks.....
We had a derailment from a heat kink during the middle of winter a couple years ago.
A heavy westbound loaded grain train pounded and destabilized a pre-existing soft spot defect in the track...
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.
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.
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.
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.
BaltACD Sounds like a 'Sun Kink' is entering the possible cause list. https://news.yahoo.com/investigators-seek-cause-deadly-montana-050015962.html
Sounds like a 'Sun Kink' is entering the possible cause list.
https://news.yahoo.com/investigators-seek-cause-deadly-montana-050015962.html
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.
Here is an example of what I am describing. Look at the video at 2:58. This train is only moving maybe 30 mph. Imagine that Amtrak train doing this at 77 mph.
https://www.youtube.com/watch?v=skfalqhzpkU
Never too old to have a happy childhood!
Electroliner 1935 Is it conceivable that the three cars rolled completely over falling onto their right side and rolling orver onto their left side? They look far enough from the ROW to have done so. Wow.
Is it conceivable that the three cars rolled completely over falling onto their right side and rolling orver onto their left side? They look far enough from the ROW to have done so. Wow.
Not really. Early photos show them on their "right" side, later on their "left". All three would have had to have been near-tippy to have continued the roll minutes or hours after they came to rest. In the early photos, you can see people being rescued--the cars were pretty stable. I can't see why they'd then suddenly all roll over another 180 degrees.
Looks to me like BNSF rolled them farther over. Which is weird.
7j43kA very knowledgeable passenger. And thoughtful. And likable. Good interview!
Being asked a lot of questions he shouldn't have been asked, IMO.
Our community is FREE to join. To participate you must either login or register for an account.