Trains.com

Passenger car construction and safety

3792 views
20 replies
1 rating 2 rating 3 rating 4 rating 5 rating
  • Member since
    March 2008
  • 773 posts
Passenger car construction and safety
Posted by ruderunner on Sunday, February 11, 2018 7:05 AM

This is spurred by the recent accident in Cayce but I thought it may be better in it's own thread since it doesn't pertain to how the accident occured but rather the aftermath.

My concern arose after determining the cafe car actually did break in half, which I find quite surprising considering it was the midpoint of the consist.  Had it been near the front of the train I wouldn't have been all that surprised.  What's disturbing is that cars ahead of it held up remarkably well, even the first car is intact, yes bent but not broken.

So im rather curious as to how the cafe car ended up in such bad shape.  Was it compromised in some way?  Just plain old?  Metal fatigue? Rust?  Improper repairs done in the past?  How does it's age and history compare to the other cars involved?

In a way it's good that it did fail in that it absorbed some of the energy of the impact and probably saved a few more lives and injuries.  

 

Modeling the Cleveland and Pittsburgh during the PennCentral era starting on the Cleveland lakefront and ending in Mingo junction

  • Member since
    April 2003
  • 305,205 posts
Posted by Anonymous on Sunday, February 11, 2018 9:13 AM

In most cases failures don't have just one reason.

One component might be the buff strength test in itself. The test car is stressed with 800,000 lbs along the centerline on the inner raft stops.

If the cafe car was on a curve there was an angle to the cars in front and behind resulting in buff loads not only along the centerline. Additionally there would have been loads at both ends perpendicular to that car, possibly changing the equation considerably.

But here too we better wait for the NTSB's findigs
Regards, Volker

  • Member since
    September 2003
  • 21,669 posts
Posted by Overmod on Sunday, February 11, 2018 11:31 AM

I expect to see the cafe car breakage turn up discussed in a FRA Research Report.  One early hypothesis would be reflected shock from the collision combining with momentum and run-in from the trailing mass, perhaps with momentarily very low effective adhesion, to produce shock buckling in the midpoint of the monocoque shell which then resulted in the two ends of the car rotating outward.

Think of this as shock far unlike that of a typical buff test, which is imposed hydraulically at a slow rate.  This is more like the effect of a hammer strike, perhaps like that of two hammers of different weight meeting at different speeds.  There might be far more force in the brief area where the shock forces combined.

From what I have seen this is not a straight compression failure, which would have bent the car closed in the middle like a beer can instead of what I have seen in the pictures so far, a clean break separation with the open ends relatively undistorted.

Please post or link detailed pictures of the damage as you find them. 

  • Member since
    March 2008
  • 773 posts
Posted by ruderunner on Sunday, February 11, 2018 12:50 PM

I agree that the forces in the accident are different than the compression test, more like a pile driver vs hydraulic press. And I would hope that NTSB report would cover it.

But it just stands out starkly that this car is in much worse shape than the others. And while being on the curve would likely have an effect, I'm surprised that the wreckage is on the inside of the curve. I would have expected it to end up towards the outside.

Overmod, in your description above do you mean to imply that the compression forces at the car ends used the truck pins as the fulcrums to buckle the car towards the inside of the curve? Would the flanges have stayed between the rails in that case?

Another scenario, though probably unlikely, could the cafe car have been stringlined by the braking forces before the run in hit it? This would have pulled the car to the inside of the curve first and allowed the compressive forces a geater angle to bend the car. I can picture it in my mind but having a hard time describing it.

Another thought but almost certainly an impossiblity, did it get pulled apart before running back together?  I'd suspect more damage along the broken ends though.

Modeling the Cleveland and Pittsburgh during the PennCentral era starting on the Cleveland lakefront and ending in Mingo junction

  • Member since
    July 2010
  • From: Louisiana
  • 2,310 posts
Posted by Paul of Covington on Sunday, February 11, 2018 1:21 PM

ruderunner
But it just stands out starkly that this car is in much worse shape than the others. And while being on the curve would likely have an effect, I'm surprised that the wreckage is on the inside of the curve. I would have expected it to end up towards the outside.

   As a non-railroader, I'm almost afraid to post something here, but I was thinking the same thing about the curve being a factor in the car breaking.   But as I see it, it would have broken to the outside of the curve.   The track leading to the siding is basically an "S" curve, and I was thinking that the impact came when the car was on the second part of the "S" curve.   Whether that would affect the type of stress the car saw, I don't know.

_____________ 

  "A stranger's just a friend you ain't met yet." --- Dave Gardner

  • Member since
    September 2003
  • 21,669 posts
Posted by Overmod on Sunday, February 11, 2018 1:37 PM

Always good to see people thinking.

ruderunner

... it just stands out starkly that [the broken cafe car] is in much worse shape than the others. And while being on the curve would likely have an effect, I'm surprised that the wreckage is on the inside of the curve. I would have expected it to end up towards the outside.

As would I.  If this were a pure run-in I'd expect the car to buckle more 'plastically' instead of abruptly breaking, and even if initially broken vertically to have the ends displace outward in falling back, even on what I believe is a comparatively shallow curve at the 'point of incidence'.

Overmod, in your description above do you mean to imply that the compression forces at the car ends used the truck pins as the fulcrums to buckle the car towards the inside of the curve?

I had not considered this before you brought it up.  My hypothesis involves 'nodes' of shock in the carbody, not compression force per se, but it seems reasonable that the truck pivots could act this way for a short period, as if they were a locating fixture for a longitudinal hammer blow.  On the other hand I find it highly significant how far the trucks are from the track, and I am looking forward to reading any 'forensic evidence' of damage they suffered, how they are aligned 'at rest', and what disturbance of the track or the surrounding earth tells how they got there.

Would the flanges have stayed between the rails in that case?

My first impression, very preliminary, is that an accident producing the force to break a car in that manner and project the ends to where they are would have lifted the trucks sufficient to derail them (or eliminate their 'functionality' as pivot points for the kinematic).  I am sure there will be software-based recreations of the forces involved, with part of the analysis being recursive to ensure any hypothesis accounts for the whole of the observed effects.

Another scenario, though probably unlikely, could the cafe car have been stringlined by the braking forces before the run in hit it? This would have pulled the car to the inside of the curve first and allowed the compressive forces a geater angle to bend the car.

I don't think this is likely, but don't let that influence your thinking.  What I know about the actual braking at this point is little more than could be gleaned from the EDR telemetry, which indicates an interesting sequence of events; the brakes were set up partially with the train still under power, with enough time to take out any slack and fully apply brake mechanisms on all the cars, and then a hard emergency application (via the modal 'button' on the assistant engineer's side) was made, effective for three seconds, which produced nontrivial deceleration of the head end (to 50 mph) in that time and corresponding distance.  I do not think that anywhere after the locomotive made contact there would have been any stringline force in the train, only deceleration forces as the momentum of the cars came sequentially (and at that point near-immediately) up against the virtually-immovable obstruction of the forward part of the train.

What may have happened in part is that some of the shock or run-in might have caused momentary loss of wheel adhesion, as likely observed on some wheels of the tied-down CSX train in its 15-foot recoil.  What else may have happened is that the car deflected vertically as well as horizontally.  But these are just suspicions from prior high-speed impact analysis.

Another thought but almost certainly an impossibility, did it get pulled apart before running back together?  I'd suspect more damage along the broken ends though.

The pictures I have seen so far certainly seem consistent with an 'axial tensional separation' of the two halves of the car, but I can't at this point imagine how that could have been produced with the car rotated as it is, where it is.

As a matter of pure and rank speculation, only thrown out for brainstorming: the car might have been initially deformed by run-in or other forces as it passed through the switch, setting it up to be essentially shattered with much less effort at a later stage of the accident.

  • Member since
    March 2008
  • 773 posts
Posted by ruderunner on Sunday, February 11, 2018 2:47 PM

Paul, good point about the S curve through the switch. I can see how that would throw the car towards the inside of the track curve, which is the outside of the switch curve. But, I believe that the consists was already completely through the switch at the time of impact correct?

If so then the S curve wouldn't really explain its final position.

Although, the whiplash through the S curve might have stressed and weakened the carbody. This starts to sound like Overmod last point.

As Volker pointed out, there's likely multiple factors besides just the forces in the accident.  Which is why I started with wondering if the car may have had it's structure compromised in some way, perhaps even a design flaw?

Modeling the Cleveland and Pittsburgh during the PennCentral era starting on the Cleveland lakefront and ending in Mingo junction

  • Member since
    September 2003
  • 21,669 posts
Posted by Overmod on Sunday, February 11, 2018 3:51 PM

ruderunner
I believe that the consist was already completely through the switch at the time of impact, correct?

I have not measured from the available angles to see if the consist would have been wholly clear of the S bend at the point of impact with the cafe in compromised.  On the other hand it is striking to me that aside from the Amfleet car that tried to follow the locomotive 'up' most of the consist, including cars before and behind the separated one, are sitting in line and from what I can tell fully railed and coupled.  It could be checked whether the cafe initially broke in or just after the switch and then came to rest as we see it, but I cannot really see any interpretation where the head end would hit at 50mph with its first trailing cars little displaced longitudinally, but the whole back end would have stopped with no collision at all 'short' of the V formed by the broken car vestibules.

Let me put out the call again for links or images of the damage, and the area around the car.

  • Member since
    September 2014
  • 80 posts
Posted by ROBIN LUETHE on Sunday, February 11, 2018 7:01 PM

Just thinking of the physics involved (not a physicist). The locomotive and first few cars would be impacted most by the collision. The cars at the end would be most impacted by the braking effect. A car in the middle could be the meeting point, momentum from behind, and collision from in front. Just curious. 

  • Member since
    September 2011
  • 6,449 posts
Posted by MidlandMike on Sunday, February 11, 2018 10:32 PM

Did the dinning car have a service door in the center?

  • Member since
    September 2003
  • 21,669 posts
Posted by Overmod on Sunday, February 11, 2018 11:20 PM

MidlandMike
Did the dining car have a service door in the center?

Not a diner, an Amfleet II cafe car

I did not see any evidence of a center door or, indeed, any opening that might weaken the shell in the area of the break.

Perhaps there is something in the framing of the food-service or underfloor equipment that would explain things.

  • Member since
    March 2008
  • 773 posts
Posted by ruderunner on Monday, February 12, 2018 4:50 AM

ROBIN LUETHE

Just thinking of the physics involved (not a physicist). The locomotive and first few cars would be impacted most by the collision. The cars at the end would be most impacted by the braking effect. A car in the middle could be the meeting point, momentum from behind, and collision from in front. Just curious. 

 

 

This is along with my thinking but I would expect the run in of the momentum to have damaged the cars at the front. They would have received a double whammy in the initial impact followed almost instantly by the run in impact.

I'm not sure the braking could have slowed the rear half of the train that fast which is why I suspect my stringline explanation will be false.

Overmod, I haven't had a chance to dig up photos yet specific to this discussion but, I do recall that Balt posted one showing the end of the consists completely in the siding. Maybe not by a couple carlengths though which MAY put the cafe closer to the switch.

Modeling the Cleveland and Pittsburgh during the PennCentral era starting on the Cleveland lakefront and ending in Mingo junction

  • Member since
    September 2003
  • 21,669 posts
Posted by Overmod on Monday, February 12, 2018 7:34 AM

No question all the consist is in the siding; the only real issue is if it would all have been wholly past the S-curve at the switch prior to the 15' recoil and the breaking and displacement of the cafe.  My belief based on the photos is that it would indeed have been, and that all the momentum and reflection effects are the result of substantially inline consist momentum at initial impact.

I have not measured the track deflection caused under the train, which is one reason I am waiting for the first kinematic-modeling 'accident reconstruction' videos.  One of you will likely see these before I do; please post their links when you do.

  • Member since
    March 2008
  • 773 posts
Posted by ruderunner on Monday, February 12, 2018 9:10 AM

i thought someone had posted a simulation video in the other thread, taken from a newscast I believe.  How scientific or accurate it was is unknown.

And yes Balts pic does show lots of track deformation before the switch

Modeling the Cleveland and Pittsburgh during the PennCentral era starting on the Cleveland lakefront and ending in Mingo junction

  • Member since
    October 2012
  • 177 posts
Posted by Jim200 on Monday, February 12, 2018 9:14 AM

In the Occupant Volume Integrity test conducted by the FRA in the link below, we see that the Budd coach was good to 1.0 million pounds and the roof started buckling, but the floor continued to take load to nearly 1.2 million pounds. From the picture we can see the side deformation, which if the load actuator could continue, would result in a severe bending in the middle.

The Amfleet II cars may be different than the Budd, but I think we can say that the tightlock coupler and the end strength and end design of the cars stood up well in the crash. However, there needs to be an analysis of the center buckling and how best to correct that. There may be angular buff forces which overcome the floor strength especially on the outside perimeter member. If the floor is distorted upward or downward, then a sidewall truss with diagonal members may be required under the windows or other changes as determined by the analysis. The next passenger fleet can reflect these findings.

https://www.fra.dot.gov/Elib/Document/120

 

 

  • Member since
    May 2003
  • From: US
  • 25,292 posts
Posted by BaltACD on Monday, February 12, 2018 11:14 AM

How many times have we been regaelled by various posters over the years that US car construction standards are archaic and should be changed to the lighter duty European 'soda can' standards in the interests of saving weight.

Never too old to have a happy childhood!

              

  • Member since
    June 2009
  • From: Dallas, TX
  • 6,952 posts
Posted by CMStPnP on Monday, February 12, 2018 12:15 PM

BaltACD
How many times have we been regaelled by various posters over the years that US car construction standards are archaic and should be changed to the lighter duty European 'soda can' standards in the interests of saving weight.

I have never advocated that and in fact I think it is really dumb if your in a trainset that is sharing ROW with heavy freight trains to start attempting to make the car lighter and less crash worthy.

If your operating on a dedicated ROW for passenger trains only....different story.

  • Member since
    April 2003
  • 305,205 posts
Posted by Anonymous on Tuesday, February 13, 2018 4:01 AM

CMStPnP

 

 
BaltACD
How many times have we been regaelled by various posters over the years that US car construction standards are archaic and should be changed to the lighter duty European 'soda can' standards in the interests of saving weight.

 

I have never advocated that and in fact I think it is really dumb if your in a trainset that is sharing ROW with heavy freight trains to start attempting to make the car lighter and less crash worthy.

If your operating on a dedicated ROW for passenger trains only....different story.

 

There are accident scenarios that can overburden the best crashworthiness design.

Overall the European crashworthiness design is at least as good as the American. Or why do you think FRA adopted it for Tier 2 passenger equipment. The buff loads can always be adopted.

The difference is: The American standard tries to make vehicles as rigid as possible, a design you perhaps wouldn't accept in your car.

The European standard follows the automobile design. There are crumple zones in unoccupied areas that reduce the forces with CEM elements to a level that the occupant zones have to withstand. The remaining buff force is about 337,200 lbs. The CEM elements can be designed to reduce the American 800,000 lbs buff load to the European 337,200 lbs. Among the CEM elements are pushback couplers. The first car takes 75% of the crash energy the remaining cars 25%.

Before you judge you should perhaps read this FRA publication with the results of American simulations comparing the different designs: https://www.fra.dot.gov/Elib/Document/2125

Especially interesting is Table 1 on page 3 summarizing the simulation results.

The EN 15227 was developed by analysing 900 accident reports. The chosen design level covers 80% of these accidents acknowleging that not all accidents can be handled.
Regards, Volker

 

  • Member since
    March 2008
  • 773 posts
Posted by ruderunner on Tuesday, February 13, 2018 8:06 PM

being in the automotive industry I can attest to the improvements that crumple zones add to a motor vehicle. I'm open to wether or not such design is helpful in connected cars. But I'd prefer not to go there as I feel that's beyond the scope of this thread.

Modeling the Cleveland and Pittsburgh during the PennCentral era starting on the Cleveland lakefront and ending in Mingo junction

  • Member since
    April 2003
  • 305,205 posts
Posted by Anonymous on Wednesday, February 14, 2018 4:33 AM

The Amtrak 91 crash shows, I think, that crumble zone work in connected cars. In this case the crumble zone was the top of the luckily unoccupied CSX locomotive.

I don't want to envision what might have happened to the passengers if the P42 had not climbed the CSX locomotive.

The broken car is another story. Here I would estimate that it had some undetected structural problems.
Regards, Volker

  • Member since
    September 2010
  • 2,515 posts
Posted by Electroliner 1935 on Wednesday, February 14, 2018 9:53 PM

BaltACD

How many times have we been regaelled by various posters over the years that US car construction standards are archaic and should be changed to the lighter duty European 'soda can' standards in the interests of saving weight. 

Yet we put 3000 lb lightweight automobiles on the highways with 90,000 lb trucks. I survived a single car accident on I-65 when I was a passenger dozing on a sunny afternoon and my wife was driving. She fell asleep doing 65 mph and hit the guardrail. It totaled the car. Fortunately. traffic was light and no other vehicles were around. Air bags, seat belts and crumple zone designs worked.

 But if a loaded semi hits you, byby.

 

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