Trains.com

East Palestine wreck

8316 views
142 replies
1 rating 2 rating 3 rating 4 rating 5 rating
  • Member since
    March 2003
  • From: Central Iowa
  • 6,542 posts
Posted by jeffhergert on Monday, February 27, 2023 3:01 PM

BaltACD

 

 
tree68
 
7j43k
I am beginning to think that there should be one conductor in a caboose, 

The cost of the added man and car notwithstanding, this goes back to one of the reasons cabeese disappeared.  And it would actually be worse now.

Slack.  

Slack going out can take a caboose from zero to something in an instant.

Slack running in will do the opposite.

Even if the train is on the move, the sudden changes in speed can be dangerous. 

If you have a 200 car train (upwards of 13,000 feet) and there's just six inches of slack between each car, that's a lot of slop.

These days, I think you'd have to put in a NASCAR-style seat/harness for the poor soul stuck in said caboose. 

 

Back in the day of cabooses - Conductors had Engineers they liked working with, they also had Engineers that the DETESTED working with.

 

 

Still true today.  The same from the engineer's perspective but for different reasons.

Jeff

  • Member since
    July 2010
  • 232 posts
Posted by ns145 on Monday, February 27, 2023 3:23 PM

A little too much in cab "slack action", eh?

  • Member since
    May 2004
  • 7,461 posts
Posted by 7j43k on Monday, February 27, 2023 3:32 PM

I don't think I agree that train length affects the intensity of the "snap" in the caboose.  It's all in the acceleration of the car that's going to be pulling the caboose.  Or deceleration, for that matter.

The faster the acceleration, the faster the velocity when the slack is taken up.  The faster the preceding car is going when that happens, the greater the "snap" in the caboose.

IF a locomotive can deliver acceleration to a train at a constant input, the resulting snap at the end of the train will always be the same, no matter the train length.

What likely happens with a long train is that the engineer needs to run out the slack to get it moving.  For a short train, he doesn't.  As the train length grows, the more necessary it is.

Ed

 

  • Member since
    December 2017
  • From: I've been everywhere, man
  • 4,138 posts
Posted by SD70Dude on Monday, February 27, 2023 3:32 PM

When switching some engineers forget that the cars also have brakes, not just the engine.  And some conductors don't understand that it's ok to tell the engineer to stop before the drawbars actually hit.  

Greetings from Alberta

-an Articulate Malcontent

  • Member since
    December 2001
  • From: Northern New York
  • 23,977 posts
Posted by tree68 on Monday, February 27, 2023 4:01 PM

SD70Dude
And some conductors don't understand that it's ok to tell the engineer to stop before the drawbars actually hit. 

That gets to be an art, in a way.  Knowing how a given engineer likes to come in and judging when you should tell them to stop so the momentum is enough to make the pin drop, but not so much as to move the car you're coupling to.

When we're doing some of our runarounds, we are working with occupied cars, oftimes with passengers watching what we're doing from the vestibule.

Had one engineer proud as punch at how perfectly he'd hit the coupler.  Had to have him back away and do it again - just a tad harder...

Then there was the engineer who liked to use every bit of that 4 MPH when making a hitch...

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
    December 2017
  • 81 posts
Posted by PennsyBoomer on Monday, February 27, 2023 6:38 PM
PennsyBoomer
Regardless, I think one upshot of this event will be that shipments such as vinyl chloride do not belong in high tonnage 150 car trains anymore.

 

From Overmod:

"I have seen a number of discussions (but have no citations) that restricting all PIH and hazmat to specific key trains only (and this potentially goes with re-imposing key train speed limits) is an invitation to activism or outright terrorism in the current climate"

From Tree68:

What are the options?  

You don't want to put all the hazmat in one train

Oil and ethanol travel in trains approx. 105 cars, and thus you already have concentrated consists of hazmat, even if somewhat more benign than some other obnoxious commodities. And a case might be made that it would be more practicable to protect concentrated shipments rather than random blocks of four or six or twenty cars, if sabotage is the premise. We'll have to see how the aftermath of this event plays and what political pressure evolves as to how restrictive a response may be necessary and in what form. But any change whatsoever is going to cost, whether it occasions more detectors, more inspections or different handling.    

 

 

  • Member since
    July 2006
  • 1,388 posts
Posted by NKP guy on Monday, February 27, 2023 6:38 PM

   I recall witnessing a particularly violent coupling at Albany of the new locomotives onto the westbound Lake Shore Limited one evening years ago.

BAM!

"Beginner's Night," the car knocker said outloud, to no one in particular.

  • Member since
    May 2003
  • From: US
  • 23,582 posts
Posted by BaltACD on Monday, February 27, 2023 7:58 PM

Train Handling is not One Size Fits All in any physical actions - braking or accelerating.

Never too old to have a happy childhood!

              

  • Member since
    March 2003
  • From: Central Iowa
  • 6,542 posts
Posted by jeffhergert on Monday, February 27, 2023 11:14 PM

7j43k

I don't think I agree that train length affects the intensity of the "snap" in the caboose.  It's all in the acceleration of the car that's going to be pulling the caboose.  Or deceleration, for that matter.

The faster the acceleration, the faster the velocity when the slack is taken up.  The faster the preceding car is going when that happens, the greater the "snap" in the caboose.

IF a locomotive can deliver acceleration to a train at a constant input, the resulting snap at the end of the train will always be the same, no matter the train length.

What likely happens with a long train is that the engineer needs to run out the slack to get it moving.  For a short train, he doesn't.  As the train length grows, the more necessary it is.

Ed

 

 

Only under certain circumstances does one need slack in the train to start it. That was mostly a steam engine practice. Diesels put out their highest traction effort at slow speeds. They normally can start most trains stretched out.

The adjustment of slack is what causes the snap. And can break knuckles and pull out drawbar. The snap is when one part of the train is moving faster than the other and then the parts try to equalize at one speed. Either causing a run in or out.

The longer the train, the more chance for slack to be moving in different directions at the same time. DP helps, but doesn't entirely eliminate slack. Especially with long travel car cushioning drawbar.

It's usually the big ugly manifests with lots of those drawbars that have the problems. When talking about limiting train length, those are the ones that need to be limited, DP or not.

Jeff

 

  • Member since
    September 2003
  • 19,826 posts
Posted by Overmod on Tuesday, February 28, 2023 9:18 AM

In the "Let's get Euclid chasing his tail" category:

I have long thought that there should be a device that restricts travel and/or compliance in hydraulic cushion underframes when the car is moving above a certain speed.  This could be done even more easily if you have a 'smart car' installation that controls braking based on car weight, monitors individual journals and wheel condition, etc.  Benefits if you can get power off an ECP 220V line...

Sic em!

  • Member since
    July 2014
  • 494 posts
Posted by Fred M Cain on Tuesday, February 28, 2023 11:01 AM

SD60MAC9500

Bearings go bad unfortunately this bearing burnt off in place that just happened to be East Palestine.. As well I know people are giving NS the 10th degree on bearing/wheelset inspection. However the covered hopper that initiated the derailment came off of UP. Did UP bother to do a proper inspection? More than likely this bearing was already in a poor conditon on the UP.. All this will come up in the cars MD-11 history. 32N originates in Madison, IL off the TRRA and carries quite a bit of UP interchange traffic from the Texas/Gulf to the East Coast..

You know the more I hear and read about this tragedy, the more I'm beginning to wonder if perhaps NS was even really negligent.  Could it be, could it just be that it was simply a "freak accident"?

I know that in today's world, people tend to strongly believe that "there is no such thing as an accident".  No, we simply MUST find someone to blame.

Trial lawyers and liberal politicians clearly smell blood.  They are sharpening their knives. Some of the lawyers will end up making a small fortune on the litigation of all this.

I don't think that much of Pete Buttigieg and normally tend to disagree sharply with him but when he says that 1,000 derailments a year is far too many, that's kinda hard to disagree with.

What exactly is the issue here?  Could it be that there is something wrong or basically flawed with the technology of modern American freight car equipment?  Are some cars too heavy for the technology used in the roller bearing journals?  Are further advances in metalurgy necessary to help address things?

I have no answers but it's something to think about.

  • Member since
    February 2007
  • From: Christiana, TN
  • 2,112 posts
Posted by CSX Robert on Tuesday, February 28, 2023 11:51 AM

Fred M Cain
You know the more I hear and read about this tragedy, the more I'm beginning to wonder if perhaps NS was even really negligent.  Could it be, could it just be that it was simply a "freak accident"?

So far, that seems to be the case.  If you read the NTSB preliminary report, there is nothing to indicate anything anybody did or did not do caused this.

  • Member since
    December 2001
  • From: Northern New York
  • 23,977 posts
Posted by tree68 on Tuesday, February 28, 2023 12:05 PM

Fred M Cain
You know the more I hear and read about this tragedy, the more I'm beginning to wonder if perhaps NS was even really negligent.  Could it be, could it just be that it was simply a "freak accident"?

In the colloquial - "stuff happens."

History is rife with "undiagnosed flaws."  One disintegrating roller, finally giving up the ghost after how many thousands of miles, would certainly wreak havoc with a bearing.  Given a 36" wheel, that's 1760 revolutions per mile, at 39 MPH, that's about 1,100 RPM.  

Coulda, shoulda, woulda facing off against "did."

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
    July 2010
  • 232 posts
Posted by ns145 on Tuesday, February 28, 2023 12:53 PM

tree68

 

 
Fred M Cain
You know the more I hear and read about this tragedy, the more I'm beginning to wonder if perhaps NS was even really negligent.  Could it be, could it just be that it was simply a "freak accident"?

 

In the colloquial - "stuff happens."

History is rife with "undiagnosed flaws."  One disintegrating roller, finally giving up the ghost after how many thousands of miles, would certainly wreak havoc with a bearing.  Given a 36" wheel, that's 1760 revolutions per mile, at 39 MPH, that's about 1,100 RPM.  

Coulda, shoulda, woulda facing off against "did."

 

The big game changer here are those lineside security cameras that clearly showed something was wrong with that covered hopper 20 miles before the derailment.  To the general public that's totally unacceptable.

I was looking for more info on roller bearing failures and found this enlightening Canadian Transport Safety Board report on a 2013 CP derailment involving a burnt off journal: https://www.tsb.gc.ca/eng/rapports-reports/rail/2013/r13t0122/r13t0122.html

The report illustrates all of the roller bearing parts and explains how wear and tear, especially spalling on the raceways or rollers, can cause a bearing to fail.  They weren't able to determine what went wrong with the failed bearing because it was too badly damaged, but when they examined the intact bearing on the other side of the axle they found that it was condemnable.  They also make the following statement about 2/3's of the way thru the report: "It is recognized in the railway industry that wheelset roller bearings can fail catastrophically in as few as 10 to 15 miles on a train travelling at track speed".  If that is the case, then it is very hard to justify 20 mile detector spacings - even more so on a rail line with a traffic density well in excess of 100 million gross tons.

Update: page 062 of NS' 2016 Pittsburgh Division track chart book shows the traffic density thru East Palestine, OH was 129.8 million gross tons: http://www.multimodalways.org/docs/railroads/companies/NS/NS%20Track%20Charts/NS%20Pgh%20Division%20Track%20Chart%202016.pdf

 

 

  • Member since
    January 2014
  • 7,863 posts
Posted by Euclid on Tuesday, February 28, 2023 5:45 PM
Brake performance played no part in this disaster.  It also seems likely that the detectors performed fine.  All three reported the overheat condition.  The first two reported the temperature and it was too low to stop the train and address the issue.  The final detector also reported the problem along with a command to stop immediately and address the problem.  But almost at the same time as that last report, the train derailed.  Time had run out to address the problem and avoid the derailment. 
 
So the detectors did their job of detecting the train condition.  But the problem was how the detectors were set to rank the danger of the problem they detected. This setting is a judgement call subject to several variables including bearing temperature, ambient temperature, the normal bearing heat production, and the range of features and their timing in a typical bearing failure.  This would require an ongoing collection and continuous review of statistics that would be consulted with every new bearing overheat condition found by detectors.
 
In conflict with setting the detection for ranking how quickly a bearing failure will derail the train, is the need to avoid overestimating the derailment risk; and causing unnecessary delay to trains by stopping them for inspection to verify and evaluate the actual risk that may be developing as a bearing is warming while running. 
 
The reason this derailment happened is because the detectors were set too lenient in ranking the risk of this bearing failure occurring within a certain timeframe.  But this judgement of leniency is in a direct tradeoff with the detector spacing along the track.  In this wreck, the spacing between the Salem and East Palestine detectors was a little too long.  If it was just 19 miles instead of 20, there would have been no derailment.  All they needed was the distance it would have taken to stop the train. 
 
So I conclude that either the detector settings should be made more restrictive or the detector spacing should be reduced.
  • Member since
    December 2001
  • From: Northern New York
  • 23,977 posts
Posted by tree68 on Tuesday, February 28, 2023 5:57 PM

Euclid
In this wreck, the spacing between the Salem and East Palestine detectors was a little too long.  If it was just 19 miles instead of 20, there would have been no derailment.  All they needed was the distance it would have taken to stop the train. 

Which, at track speed for a train of that size, would have been a mile.  I suspect that even if there had been another detector 10 miles out, they ran a high risk of a derailment, unless that intermediate detector tripped an alarm.

Note the Canadian study/report, cited elsewhere, which put the failure distance at a little as ten miles.

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
    January 2014
  • 7,863 posts
Posted by Euclid on Tuesday, February 28, 2023 6:13 PM

tree68

 

 
Euclid
In this wreck, the spacing between the Salem and East Palestine detectors was a little too long.  If it was just 19 miles instead of 20, there would have been no derailment.  All they needed was the distance it would have taken to stop the train. 

 

Which, at track speed for a train of that size, would have been a mile.  I suspect that even if there had been another detector 10 miles out, they ran a high risk of a derailment, unless that intermediate detector tripped an alarm.

Note the Canadian study/report, cited elsewhere, which put the failure distance at a little as ten miles.

 

I don't know what the spacing should be.  As I mentioned, that depends on the detector settings for  the maximum allowable temperature before sending a message to stop and inspect the train.  Lower that threshold, and you can increase the spacing, and vice versa.  

So what I have mentioned about the 19 versus 20 miles is just an example.  We know the last detector, 20 miles from Salem, did call for the train to stop, and the derailment happened to occur at that same moment.  So the 20 mile interval was not quite short enough to prevent the derailment. 

But there are still other pieces to the picture.  With other bearing failures the failure interval from detector #2 might have only been 5 miles.  But overall, the detector sampling resolution was not high enough to prevent this derailment.  And raising that resolution will cost money.  So there is the ultimate tradeoff.    

  • Member since
    February 2007
  • From: Christiana, TN
  • 2,112 posts
Posted by CSX Robert on Tuesday, February 28, 2023 9:00 PM

Euclid
Brake performance played no part in this disaster.  It also seems likely that the detectors performed fine.  All three reported the overheat condition.

To be clear, the first detector did not report an overheat condition and the I don't believe the second did either.  Wheel bearings do not operate at ambient temperature, the 38° over ambient is well within normal operating parameters - definitely not an overheat.  The second one, while at the high end, I believe is still in normal range (there's variety of what's consider 'normal' from what I've read).  It definitely heated up significantly, but one thing the preliminary report did not have was the train speed at each detector.  If the train had slowed significantly before the first detector and the speed back up to track speed before the second, then the tenperature rise could be expected.  Also, a comparison of other bearing temperatures could help.  If they all rose significantly between the two detectors. that one bearing may appear normal, but if if the others didn't rise, or if this one rose significantly more, then maybe that should have been a sign of an issue.

  • Member since
    March 2003
  • From: Central Iowa
  • 6,542 posts
Posted by jeffhergert on Tuesday, February 28, 2023 10:24 PM

The first two didn't didn't trigger an alarm to the crew.  They recorded the condition.  I don't know, but would guess NS has some kind of "bearing desk" that monitors such things.

It's possible to fool hot box detectors.  Our instructions are to avoid using air brakes close or while passing over them when possible.  They either could have a false activation or mask one that's hot.  Detectors also compare the temperature of the entire train.  It would be nice to have information as to temps on the rest of the train.  If there were something that may have appeared to account for an elevated temp between the first and second detector.

Jeff   

  • Member since
    January 2014
  • 7,863 posts
Posted by Euclid on Wednesday, March 1, 2023 7:15 AM

CSX Robert

 

 
Euclid
Brake performance played no part in this disaster.  It also seems likely that the detectors performed fine.  All three reported the overheat condition.

 

To be clear, the first detector did not report an overheat condition and the I don't believe the second did either.  Wheel bearings do not operate at ambient temperature, the 38° over ambient is well within normal operating parameters - definitely not an overheat.  The second one, while at the high end, I believe is still in normal range (there's variety of what's consider 'normal' from what I've read).  It definitely heated up significantly, but one thing the preliminary report did not have was the train speed at each detector.  If the train had slowed significantly before the first detector and the speed back up to track speed before the second, then the tenperature rise could be expected.  Also, a comparison of other bearing temperatures could help.  If they all rose significantly between the two detectors. that one bearing may appear normal, but if if the others didn't rise, or if this one rose significantly more, then maybe that should have been a sign of an issue.

 

Yes, I should not have used the term “overheat condition” regarding the first two detectors. My point was that all three detectors worked properly.  The first two each reported a temperature rise above ambient, and it was a rising trend even though not rising to the extent of requiring warning the crew. 
 
As I understand, the third detector found the temperature high enough to warn the crew and did so very shortly before, after, or during the start of the derailment.   So, all 3 detectors worked as they were intended to, according to their settings.  But my overall point was the settings having a threshold that was too high to prevent the disaster.  That suggests either changing the settings to lower the threshold or increase the frequency of detectors.    
  • Member since
    December 2001
  • From: Northern New York
  • 23,977 posts
Posted by tree68 on Wednesday, March 1, 2023 9:40 AM

Euclid
But my overall point was the settings having a threshold that was too high to prevent the disaster.  That suggests either changing the settings to lower the threshold or increase the frequency of detectors.  

I would opine that the temperature trigger level of the detectors, etc, is set based on experience.  It's not like these are new to the railroads.  That the system can detect trends is an indication of where that technology has gotten.

The danger of setting the thresholds lower is a likely increase in false alarms which will require a stop for inspection.  With a 13,000 foot train on a busy mainline, this would be a problem.  Most folks could cover that distance in under an hour, on a good surface.  That's one way, so we're talking up to a two hour delay, even if nothing is found.  

Adding acoustic sensors to existing detectors may be one of the better solutions.  Given the sparks/fire seen in that vicinity, the Salem detector may have sensed something amiss, which probably would have led to an earlier stop.  The sites already exist, including the necessary power and communications.  This would simply be adding another sensor to the suite.

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
    December 2007
  • From: Georgia USA SW of Atlanta
  • 11,300 posts
Posted by blue streak 1 on Wednesday, March 1, 2023 1:36 PM

This route appears to be rather busy.  It now has 2 main tracks at the accident scene.  The site appears that at one time a third track was between the present 2 main tracks.  If the 2 MTs are at  approximately HBD mile posts locations 79, 69, and 49 that means that 2 complete on track monitors at each location.  Why did NS not have HBDs at approximately MP 59 ?

As well are there detectors at MPs ~ 89 and 39 and onward ?  Just trying for  information of NS's thoughts on spacing. 

Find it interesting that CP had more derailments due to defects than CN which had a centralized monitoring system. Granted the total numbers are too small for statistical comparsion. 

  • Member since
    August 2022
  • From: New England (Cape Cod)
  • 120 posts
Posted by DonRicardo on Monday, March 6, 2023 12:27 PM

How fast does a typical 150 car frieghter go? Would reducing speeds, while that would increase times and costs, help reduce these derailments? There was just another 20 car derailment, and I understand that since Jan 1st, ther ehave been at least a dozen.

Is sabotage a possibility?

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