Oil Train

[BaltACD]

The accordion process stops when the stresses on the coupling are not sufficient to break the coupling between cars in two.  The entire derailment procees is an exercise in depleting the kinetic energy in the moving train to zero.  It cannot be depleted to zero, instantly.

[Euclid]

tree68

 

Euclid

...the zigzag jackknifing of the accordion process can suddenly cease,...

 

 

And this would happen how?

 

The derailment process is generally chaotic, but patterns can form amid the random action.  One classic pattern is the jackknifing, zigzag that forms an accordion pattern.  For as long as the forming accordion pattern continues, it can protect the pileup from a direct impact.

But the continuation can suspend at any point if random factors intervene and restore the chaos.  Generally, that could be caused by a parting at the joint of the first incoming car on the rails in the string of trailing cars and the next car ahead that is about to jackknife.  It could be influenced by the extent to which the zigzagging cars dig into the ground and resist.  It could also be influenced by cut banks along the roadbed, or by the presence of fills leaving the ends of the zigzagged cars unsupported.  It could also be influenced by a jackknifing car failing to fully fold up against the car ahead of it, thus changing the deflection angle for the next car coming in.

Generally, my only point is that the accordion process could continue until the train stops, or it can end at any point before the train stops.  It also might never form at all during the derailment.  Instead, the cars might pile up in a completely chaotic, random pattern.  That way, every incoming car might have the opportunity to strike other cars that are relatively anchored by the random pileup.

But again, the point is not so much that these collision impacts can occur.  Instead, it is that the real world force potential is so vastly greater than just one moving car striking a stationary car; the simplistic model of collision which appears to be assumed in the crash testing of tank cars.    

[MidlandMike]

Euclid

...

The couplers and draft gear of the incoming string of cars will be able to rigidly transfer the momentum forward because those cars are running on the rails of undamaged track.  So the incoming string of cars would act just like a colliding train as shown in the crash test video.         

 

The only way the incoming cars could ridgidly transfer all their momentum, is if all those cars were welded together as a single car and were structurally ridgid so that they themselves would not crumple.  The couplers/draft gear do not accomplish that.  Additionally all the cars are identically constructed, and the following cars are no stronger than the cars already derailed.

[tree68]

Euclid

...the zigzag jackknifing of the accordion process can suddenly cease,...

And this would happen how?

[Euclid]

Midland Mike,

Well, the process that I am referring to does include the accordion-like bunching of cars that you mention.  That is where each car entering the growing heap is forced to nestle into the heap, and become backed up by the mass of the other cars ahead of it in the heap.  That accordion heap of cars is directly within the derailment process. 

But the process also includes the string of cars that are rolling in and feeding cars into that accordion heap. As the heap grows larger, the zigzag jackknifing of the accordion process can suddenly cease, and therefore allow the incoming string of cars to punch directly into the mid-section of a tank car crosswise in the heap.  Maybe the next car in that string will bypass the car that has struck the mid-section of the crosswise car.  And then that bypassing car also strikes the mid-section side of the car struck by the previous car.  So, all of a sudden this process turns into a series of direct collisions between the incoming cars and the cars that are trapped in the heap with no way to deflect.

The couplers and draft gear of the incoming string of cars will be able to rigidly transfer the momentum forward because those cars are running on the rails of undamaged track.  So the incoming string of cars would act just like a colliding train as shown in the crash test video.         

[MidlandMike]

I presume several following cars give more force than a single car, however, couplers/draft gear do not ridgidly transfer the momentum of an entire trainload of following cars, and in fact fail, and often give way to an accordian-like bunching of cars.

[Euclid]

MAXIMUM FORCE POTENTIAL

 

I have mentioned the possibility of tank cars being trapped in a pileup and then being squeeze-burst from the impact from an oncoming string of cars still running on the rails.  As hard to believe as that may be, the point is only secondary to a larger point.  The larger point is that the force needed to squeeze-burst a tank car is available, and will be practically applied depending on the random dynamics of the derailment and where the derailment occurs in the train.  This maximum force potential comes from the cumulative kinetic energy of several cars running on the rails and coupled together. 

If the conditions do produce this maximum force, it will apply to all types of impact damage including punctures, crushing, crumple, and tears, although these types of damage can be caused by much less force than the maximum force potential that I am referring to.

I have seen a few references to tank car crash testing that seems to ignore this maximum force potential.  Instead, the tests show a rolling mass representing a single tank car, and it collides with an actual tank car from various angles.  Apparently, this type of testing result is what the newest puncture and tear resistance standards are based on.  Where are the tests that show what happens in the maximum force potential, say when 20 cars strike a tank car broadside when the car is trapped in a pile?    

This maximum force potential comes not just from one car striking another as seen in the crash tests.  Instead, it comes from the oncoming cars still on the rails behind the derailment.  The overall dynamics of this force depends on where the derailment occurs in the train.  If it occurs only a few cars up from the hind end, there may be little or no force from the trailing cars.  If it occurs near the head end, there will be maximum potential for the force from the trailing cars; however, there will be relatively little potential for the pileup to grow in its resistance to the force from the trailing cars.   

I believe that this maximum force potential vastly exceeds ability of thicker tank walls and stronger head shields to resist it.  That is why the car builders have said that it is impossible to make cars strong enough to resist breaching in a high speed or high energy derailment.   

This maximum force potential is basically the same force involved when a train runs into another train.  Consider a tank car that is pinned in position and hit by say 2,000 tons of loaded tank cars.  I would expect that pinned tank car to disintegrate.  It may be purely a squeeze-burst, or it may be combined with puncturing, tearing, or cracking of the bending and crumpling wall.  But in general, it seems like the force potential would be way out of proportion to what could be resisted by stronger head shields and tank walls. 

Here is a crash test video that gives a good feeling for the kind of force that would be involved in in a derailment where individual cars are struck by the incoming string of trailing cars.  The music is a little weird, but the video is quite graphic with a variety of impacts.  Notice how rubbery the equipment looks as it is impacted:   

https://www.youtube.com/watch?v=Q9gUFR8P0jQ

[MidlandMike]

Euclid

Midland Mike,

Can you explain how tank cars are typically loaded for unit trains regarding how much of the tank is filled with liquid, and how much is gas space?  I recall that it was brought up in the sloshing theory of oil train problems. 

Someone posted this information in a thread here on the general forum. The post may be in this thread, but I can’t easily find it. Can you confirm whether or not the information about the weight and empty space inside of the tank car is true?  This is quoted from the post:

“Bakken crude oil is a stratified multi constituent liquid. Its weight is such that something like 28,000 gallons are the weight limit for a 30,000 gallon tank car. Visualize the 2000 gallons as about 36 drums of 55 gallon capacity. That's quite a bit of empty space inside a tank car. It is about 269 cubic feet.”

 

I retired before crude-by-rail was widespread, so I can only tell you what I know.  Crude oil comes in a wide range of densities, so it's not like you could design a perfect sized tank car for CBR.  Therefore, you would expect that they would use the next biggest volume sized tank car to contain the max weight load.  At approx 7 lb/gal, 200,000 pounds would be in the 28 to 29,000 gallon range, so there would be some leftover volume in a 30 to 33,000 gal tank car.  Vapors evolving off the crude, plus whatever fumes left over from the previous load, would fill the headspace.    One problem I have with the quote you cited is the ambiguous use of the word "stratified".  It would seem out of context when talking about CBR, as the Canadian report on the Lac Megantic wreck found that the crude in the uninvolved tank cars was not stratified.

[Euclid]

Midland Mike,

Can you explain how tank cars are typically loaded for unit trains regarding how much of the tank is filled with liquid, and how much is gas space?  I recall that it was brought up in the sloshing theory of oil train problems. 

Someone posted this information in a thread here on the general forum. The post may be in this thread, but I can’t easily find it. Can you confirm whether or not the information about the weight and empty space inside of the tank car is true?  This is quoted from the post:

“Bakken crude oil is a stratified multi constituent liquid. Its weight is such that something like 28,000 gallons are the weight limit for a 30,000 gallon tank car. Visualize the 2000 gallons as about 36 drums of 55 gallon capacity. That's quite a bit of empty space inside a tank car. It is about 269 cubic feet.”

[MidlandMike]

schlimm

...  

However, with oil price instability (generally lower) and thus declining Bakken production, this may all be a moot point.  ...

 

North Dakota production is up over 10% from this time last year when the oil prices started to drop.

www.dmr.nd.gov/oilgas/stats/historicaloilprodstats.pdf

 

[MidlandMike]

Euclid

...

But I assume that the majority of the tank volume is occupied by non-compressible oil.  ...

Crude oil is a multi-phase fluid with a dissolved gas fraction.

[schlimm]

dehusman

 

 

schlimm

car-breach spill => pool fire => thermal failure => energetic fireball eruptions

 

 

Exactly.

And this is why I say all this uniformed "problem solving" isn't helpful, because while all the discussions have been focused on brakes and derailment detectors and curious failure modes, the "real' solution may be just a layer of insulation over the tank cars to thermally shield them from the heat of the burning product long enough for the flames to die down or lt the responders get fire or temperature suppression efforts underway.

 

Any mitigation is helpful.  Short-term, fairly inexpensive insulation helps reduce the chances of an accident with car-breach spill and fire from turning into a fireball disaster.  

 

However, with oil price instability (generally lower) and thus declining Bakken production, this may all be a moot point.   And with an end of subsidies for ethanol used as a gasoline additive also likely, that business may decline, same as coal hauling to utility plants.  The rails would be well-advised to start looking for other more reliable long-term sources of traffic.

[Euclid]

 

I can see five basic stages of an oil train wreck where preventive action could be taken:

1)    Derailing of train.

2)    Piling up of cars.

3)    Breaching of cars.

4)    Ignition of oil.

5)    Fire causing more breaching of cars.

 

Preventative means and objectives:

#1 can be furthered by more inspection and defect detection.

#2 is has not been considered until the advent of the ECP mandate.

#3 is not possible in high speed derailments.

#4 is not possible.

#5 buys time to evacuate area.

[dehusman]

schlimm

car-breach spill => pool fire => thermal failure => energetic fireball eruptions

Exactly.

And this is why I say all this uniformed "problem solving" isn't helpful, because while all the discussions have been focused on brakes and derailment detectors and curious failure modes, the "real' solution may be just a layer of insulation over the tank cars to thermally shield them from the heat of the burning product long enough for the flames to die down or lt the responders get fire or temperature suppression efforts underway.

[schlimm]

wanswheel

It seems ‘burst’ = ‘thermally fail.’

car-breach spill => pool fire => thermal failure => energetic fireball eruptions

The public will draw its own conclusions.

[BaltACD]

[wanswheel]

It seems ‘burst’ = ‘thermally fail.’

Excerpt from NTSB safety recommendation, April 3

http://www.ntsb.gov/safety/safety-recs/recletters/R-15-014-017.pdf

On February 16, 2015, at 1:15 p.m. eastern standard time, an eastbound CSX crude oil unit train derailed 27 loaded tank cars in Mount Carbon, Fayette County, West Virginia. The train consisted of 2 locomotives followed by a buffer car, 109 tank cars, and a single trailing end buffer car. The train was transporting about 3.1 million gallons of Bakken crude oil, UN1267, Class 3, Packing Group I, from Manitou, North Dakota, destined for the Plains Marketing Terminal in Yorktown, Virginia. Event recorder data indicated that the train was being operated at 33 mph at the time of the accident, below the 50 mph maximum authorized speed. At the time of the accident, CSX had a temporary 40 mph speed restriction on the territory due to cold weather. The weather was 15° F with eight inches of recent snow.

About 300 people were evacuated from within a one-half mile radius of the scene. About 378,000 gallons of crude oil was released in this accident from tank car punctures, damaged valves and fittings, and tank car shell thermal failures. Crude oil was discharged into the Kanawha River and contaminated soils in the area of the derailment, and fueled a postaccident fire. Much of the crude oil was consumed in the postaccident fire. Emergency responders allowed the fire to burn itself out and it was extinguished by 8:00 p.m. on February 17, 2015, more than 30 hours after the derailment.

 

All of the tank cars involved in this accident were compliant with the CPC-1232 industry standard. None of the tank cars had thermal protection. During the derailment sequence, two tank cars were punctured and released more than 50,000 gallons of crude oil. Of the 27 tank cars that derailed, 19 cars became involved in a pileup and postaccident pool fire. The pool fire caused thermal tank shell failures on 13 tank cars that otherwise survived the accident. Only one car at the edge of the pool fire survived without release. The eight tank cars on either side of the pool fire were not significantly damaged and did not release product.

 

Emergency responders reported that the first thermal failure occurred about 25 minutes after the accident. By about 65 minutes after the accident, at least four thermal failures with energetic fireball eruptions had occurred. The 13th and last thermal failure occurred more than 10 hours after the accident.

 

The NTSB has also collected information relative to three additional recent accidents in which CPC-1232 tank cars derailed and breached due to postaccident thermal failures.

 

On February 14, 2015, a Canadian National (CN) crude oil unit train with 100 tank cars derailed 29 cars in a remote area near Gogama, Ontario, while traveling at 38 mph. All tank cars were less than 4 years old and were compliant with the industry CPC-1232 standard. Investigators found that 19 of the cars were breached and released more than 264,000 gallons of crude oil. Tank car inspections revealed at least five tank cars sustained postaccident thermal failures.

 

On March 5, 2015, a BNSF crude oil unit train traveling at 23 mph with 103 DOT-111 tank cars derailed 21 CPC-1232 tank cars in a rural area south of Galena, Illinois. A postaccident pool fire that began with product released from damaged valves and fittings on some tank cars resulted in five tank car thermal failures.

 

On March 7, 2015, a CN crude oil unit train with 94 DOT-111 tank cars derailed 39 CPC-1232 cars while traveling at 43 mph at the west end of a CN rail bridge that traversed the Macaming River near Gogama, Ontario, about 23 miles from the above-mentioned February 14, 2015, accident location. Five tank cars came to rest in the river and the remaining cars piled up on the west side of the bridge where tank cars were breached, released product, and ignited a large pool fire that destroyed the rail bridge. At least five of the derailed tank cars sustained postaccident thermal failures.

 

Preliminary investigation results indicate that a total of 28 CPC-1232 compliant tank cars thermally failed in these four accidents. One person was injured, and thousands of gallons of crude oil were released, causing significant property damage and environmental pollution. In these four accidents, investigators have found consistent evidence of shell bulging from overpressure, shell material thinning, and tank shell tears in the vapor space of bare steel CPC-1232 tank cars exposed to pool fire conditions. The NTSB concludes that the thermal performance and pressure relief capacity of bare steel tank cars that conform to current federal and industry requirements is insufficient to prevent tank failures from pool fire thermal exposure and the resulting overpressurization.

[Euclid]

tree68

 

dehusman

And when are derailed cars up against a truly "immovable object"? 

 

 

Exactly.

 

The derailed cars can certainly be up against an object that is relatively immovable; enough so to provide enough resistance that they can be crushed by the incoming cars.  Not every derailed car will be in that predicament, but when they pile up, they can certainly resist being moved. 

The perfect illustration is in the aerial views of the Lac Megantic wreck.  The fourth to the last car is crosswise to the track and pushed tightly against the side of the car ahead of it.  It is the last of a whole series of cars packed side to side in the classic accordion pattern.

The last three cars are in line with the track.  The third to last car, and second to last, are side by side since one bypassed the other.  Both of them have perfectly “T-boned” the fourth to the last car in its mid-section, inflicting crush damage to the tank.

The last car is in line with the track directly behind the two bypassed cars.  But this last car was not the last car during the wreck.  Several more cars (approx. 6-9) had been cut and pulled away from the wreck to save them from the fire.

So the fourth to last car (as photographed after the hind end had been pulled away) had turned crosswise and posed a great deal of resistance.  It was struck at right angles to its mid-section by the following car which was leading about 9-12 cars all rolling forward on the rails.  Notice that the two striking cars did not ride up and over the cylinder of the struck car.  Instead, they dove right into it.   

[tree68]

dehusman

And when are derailed cars up against a truly "immovable object"? 

Exactly.

[Wizlish]

Euclid

What I am referring to is tank cars going into an accordion pattern where they line up side to side, crosswise to the track. Then suddenly the formation of this zigzag pattern is disrupted, and the next car running in line with the track collides with the mid-section of a car that is crosswise to the track. Say that colliding car, in line with the track, is leading twenty more loaded cars all running on the rails. That is the way the sledge hammer hits the pop can.

About the only way that's going to happen is if there is a break-in-two somewhere in the train, and the 'accordioning' takes place on the forward section.  Then the rear section slides against the side of the accordioned car.  The general appearance, perhaps, being what happens with the 'back end' of the train in the famous tornado video.

What will happen here is that the end of the tank car will first buckle the side of the car being 'hit', then penetrate it.  It may proceed to go into the cars accordioned up behind that one, too.  It is possible this will cause 'slosh' against the ends of the tank being hit -- but this will be accelerated mass, not hydrostatic shock.  I'd think there would surely be some FRA tests of side-on or "T-bone" collisions between tanks and accordioned cars.

There are huge forces governing why consists 'accordion', and I suggest you work through the physics to see why T-bone collisions at high momentum aren't that likely.

 

[dehusman]

tree68

Perhaps a practical experiment would be in order. Stack several full soda cans on top of each other (end to end, not on their sides), then hit the top can with a large hammer, swinging downward. See if any of them burst. That would be roughly analogous to a train of tank cars hitting an immovable object.

And when are derailed cars up against a truly "immovable object"?  Derailed cars are norally not constrained by an artificial barrier.  None of the cars in any of the oil train derailments were against an "immovable barrier".  They also can move up, left and right.  About the only "immovable" barrier, is down, into the ground.  If you hit them on an end they will rotate.  If you hit a tank car above the center line the car will go above it. If you hit a car that is at an angle, the moving car will be deflected along side the stationary car.

[Euclid]

I want to stop and clarify something.  I am not talking about tank cars being crushed lengthwise as though the train suddenly stops and the slack runs in hard.  I am not referring to telescoping the incoming cars as they impact the heap.  From the last two posts, I get the impression that that is what is being inferred.  

What I am referring to is tank cars going into an accordion pattern where they line up side to side, crosswise to the track.  Then suddenly the formation of this zigzag pattern is disrupted, and the next car running in line with the track collides with the mid-section of a car that is crosswise to the track.  Say that colliding car, in line with the track, is leading twenty more loaded cars all running on the rails.  That is the way the sledge hammer hits the pop can.

It also would depend on the struck car being sufficiently backed up with enough other cars in the accordion pattern to provide sufficient resistance to the impacting force.

[Norm48327]

"Perhaps a practical experiment would be in order.  Stack several full soda cans on top of each other (end to end, not on their sides), then hit the top can with a large hammer, swinging downward.  See if any of them burst."

Shades of Gallagher's 'Sledge-o-matic'. Ought to be fun to watch.

Let Bucky try it.

[cx500]

Indeed, sometimes the internal pressure may burst the tank.  In every case that I am aware of there was another factor that was the root cause, namely fire.  An intense external fire concentrating on an area of the tankcar above the liquid line will eventually raise the temperature of the shell sufficiently that the steel will weaken.  And of course that same heat is increasing the pressure.

Otherwise, the tankcar shell and shape is such that the forces you envision will not develop.  Any punctures will be due to point loads, such as knuckles or rails.  To get some idea, take the empty cardboard tube from the inside of a toilet paper roll, stand it up on end, and press down as hard as you can with your hand. Even though it is made of thin cardboard you will find it has surprising strength.  And that is the direction in which the strongest compressive forces will act on tank cars in a derailment.

You may not "imagine NTSB reports getting into that kind of detail".  Perhaps you should start looking at some, or similar ones on the Canadian TSB site.  Where appropriate they most certainly can and do. Speculation may be fun.  But idle speculation without bothering to do research is just plain lazy ignorance.

[tree68]

Per 49CFR179.201-1, the design burst pressure for a DOT111A tank car is 500 PSI.  If the car is full to the point of overflowing, that kind of pressure might be possible during an incident, if the circumstances were right.  Given the normal air space left in the car, not so much.  

I don't do hydraulics, so I can't guess how much the air space would have to be reduced to produce that kind of pressure.

But I really agree that it's a moot point.  By the time you reach that kind of compression, that half inch of steel will have suffered mechanical damage that will have breached the tank.  

Perhaps a practical experiment would be in order.  Stack several full soda cans on top of each other (end to end, not on their sides), then hit the top can with a large hammer, swinging downward.  See if any of them burst.  That would be roughly analogous to a train of tank cars hitting an immovable object.

[Euclid]

Dave,

Sure, I agree that if it never happens, there will be no way to find documentation that it does happen.  However, I am not convinced that there is no documentation just because you have never seen any.  I would check for documentation, but I have no idea where to find it.  If I found documentation of the minute details of oil train derailment damage, I would review it all.  But until I see it, I would not even assume that it exists.  I cannot imagine ICC or NTSB reports getting into that kind of detail.

I agree with your point that the fireball explosions seen on news coverage are from heating by fire after the derailment occurs, and are not related to the squeeze bursting that I am referring to.      

I don’t understand this part of your comment:  “You will find lots of examples of cars being punctured because if you apply enough pressure to "burst" the shell you will most likely exceeded the strength of the shell and will puncture or tear it before you burst it.”

If there is enough pressure to burst the tank, why wouldn’t it just burst?  I consider tears and punctures to be damage that results from contacting obstacles outside of the tank.  I consider bursting to be an opening of a crack in the wall due to internal pressure.  However, this rise in pressure is the result of external pressure that collapses the tank.  Maybe that is what you are referring to as leading to a rupture before the tank actually bursts.  Indeed, I expect that these modes do often combine such that the collapse could weaken an area by bending, and then the pressure rise would open that weakened area.

But my largest point in all of this is that the potential force available to produce this damage is far greater than what tank cars can withstand without breaching.  That force can drive punctures, tears, buckling, and cracking; although these modes do not need such a high force.  They can also happen when a single tank car collides with something or is struck by another single tank car.  It would likely be only the squeeze-burst mode that requires the maximum force potential of a line of cars feeding into a large pile.  If that occurs, I assume that no amount of practical strengthening can prevent any of the breach modes.  That seems to be what the car builders have recently stated.       

[dehusman]

Euclid

You raise good points. All I am saying is that I believe this happens often. If it can be proven otherwise, so be it. If I could procure examples, I would, but how can I do that?

If it happens as often as you say it does, then there should be dozens of examples in NTSB reports of derailments involving tank cars where the cars were collapsed by being squeezed (and it should happen with any tank car carrying a liquid, not just hazmat).  Since it happens so often and is such a common mode of failure (if you are correct) then you should be able to cite numerous examples in the NTSB and ICC accident reports.

On the other hand, since it doesn't happen that way, there really isn't any way I can find documentation of something that didn't happen.

You will find lots of examples of cars being subjected to heating and rupturing from over pressurization due to the heating, that's what was happening in the in the oil train derailments.  All those explosions you see on video, those are the cars cooking off.  The actual derailment took maybe a minute, it took a lot longer to get camera there to start recording the accident, so all those explosions on video occurred loooooong after the collision.

You will find lots of examples of cars being punctured because if you apply enough pressure to "burst" the shell you will most likely exceeded the strength of the shell and will puncture or tear it before you burst it.

[Wizlish]

Euclid

As you mention, the relief valve would open before the tank bursting pressure would be reached. However the open valve may not be sufficient to vent the pressure if the increase is too rapid, as it would be in a collision.

Actually, that latter point is the one I was making in an earlier context -- it's valuable to remember the difference between flow and pressure relief.  There is a reason that drop plugs on Super-Power-size steam boilers are pointless for their advertised 'safety relief' purposes -- by Nathan's own calculations (if you do them rather than just read about what they are) you would need many of them, not just the typical number installed, to have an actual effect on dropping boiler pressure -- and with that many actually venting, you'd have at least the same danger to the crew in the cab as you'd have with busted elements.  But I digress...

 

I don’t understand what you mean when you say this: “Likewise, you would need to reduce the actual contained volume of the tank car something like 90%, without cracking or breaching the tank in any way in the process, before you would develop any significant hydrostatic pressure against the tank.” Why would the volume reduction have to be as high as 90%?

Because I thought it had been established, perhaps in this very thread, that the oil 'weighs out' long before it 'cubes out', and that there is supposed to be surprisingly little actual liquid in these loads.  That also goes a long way toward establishing why so many of them go off like bombs or provide amazing near-critical-mixture flame shows.

I have no idea how much air space is in a load of crude oil. But I assume that the majority of the tank volume is occupied by non-compressible oil.

That's a wrong assumption.  I would PM a couple of the people who work in the petroleum industry and find out exactly what volume is 'loaded' for the different crudes, and perhaps whether any gas-blanket overpressure was applied to the cars to ensure some of the volatiles stayed in the liquid phase. 

This is a fundamental consideration for oil-train safety!  (Lest you think I'm picking... I didn't figure this out until a poster mentioned it here, and yes, I'm kicking myself for not having thought about it.)

 

One thing I'm wondering is if there are good "statistical" studies of the actual failure modes of tank cars in severe collisions.  In order to get 'smooth' compression of a tank-car structure to the point where either gas or hydrostatic pressure become feasible causes of rupture, very specific force needs to be applied 'just right', and progressively 'right', to cause the cylindrical side of the car to bend inward past 'flatness' to a U shape, following as you noted what would be required to squeeze toothpaste 'under pressure' from a part-empty tube.  But most of the dynamics in derailments, specifically including jackknife stack-ups, aren't going to produce this (which is one reason why I want to see photographic evidence of the conditions under which it would happen).   I think it is much more likely that the cars would rend, or buckle, or twist, or be penetrated by rails or coupler knuckles, in other words be cracked or torn open, long before pressure popped them open.

Now, I do think that various forces will expel oil from 'breaks' much more strongly during a wreck.  But I want to point out, specifically, that this is not 'hydrostatic bursting'.  In other words, the damage that lets the oil out is the reason for the spills and possible aerosolization/carburetion , not the oil causing the damage.

[Euclid]

 

Wizlish,

You raise good points.  All I am saying is that I believe this happens often.  If it can be proven otherwise, so be it.  If I could procure examples, I would, but how can I do that?  But in the meantime, I don’t see why it would be considered to be an extraordinary claim.  Again, I don’t want to get sidetracked by my use of the term “burst.”  I simply mean opening a large crack and releasing oil through it.  It is not to imply an explosion that, in itself, adds to the catastrophe.  It is just a point as to how the oil gets spilled.     

As you mention, the relief valve would open before the tank bursting pressure would be reached.  However the open valve may not be sufficient to vent the pressure if the increase is too rapid, as it would be in a collision. 

I don’t understand what you mean when you say this:

 “Likewise, you would need to reduce the actual contained volume of the tank car something like 90%, without cracking or breaching the tank in any way in the process, before you would develop any significant hydrostatic pressure against the tank.”

Why would the volume reduction have to be as high as 90%?  I have no idea how much air space is in a load of crude oil.  But I assume that the majority of the tank volume is occupied by non-compressible oil.  If, for example, 100% of the tank were filled with oil, and if there were no relief valve; then it would seem that the slightest reduction of volume would burst the tank.  The only possible relief to prevent bursting would be the tensile elasticity of the tank wall.

So with an oil load probably filling at least 85% of the tank volume, I would think that volume reduction could easily achieve a bursting pressure.  And, as I mentioned, I expect this crushing of the vessel also is often accompanied by buckling fracture, punctures, and tearing—all being driven by the force of multiple tank cars still on the rails and running into the pileup. That force is the essence of my point. 

[Wizlish]

Euclid

Just to clarify: I referred to a tank car caught in a pileup, and subjected external squeeze that collapses its tank, thus raising its pressure high enough to cause it to “burst.” ...  The “bursting” will be similar to squeezing toothpaste out of the tube, as you say, in that it will be mostly hydraulic. Although, there will be compressed air as well to the extent that the tank fill includes air space.

I think that part of the problem is that, given the actual amount of oil that constitutes a 'full' load in a typical HHFT car, you would have to have a very special kind of accident,probably an unlikely accident, to develop this method of failure.

First, long before you get enough air or gas pressure developed in a partly empty tank to 'burst' it, you will have to compress its structure far beyond the point it will have broken in some other way, probably by kinking or puncture.  Are you reasoning by analogy with those 'vacuum' incidents where low vapor pressure collapses the tank (so it looks like an empty toothpaste tube)?

I do think it is possible for 'gas' pressure inside the tank to build up faster during an accident than safety valves, rupture disks, or valves would relieve.  On the other hand, the actual pressure that the tank structure would withstand would be comparatively high, and almost surely the tank would fail due to the forces doing the 'compression' (and in doing so, relieve the gas pressure) long before the steel vessel would burst.

Likewise, you would need to reduce the actual contained volume of the tank car something like 90%, without cracking or breaching the tank in any way in the process, before you would develop any significant hydrostatic pressure against the tank   As you may know, the hydrostatic pressure that then develops will be equally (and fairly effectively!) distributed equally to all interior points of the now-collapsed shell.  That's not likely to 'burst' anything; more likely, it would tend to open up a folded part of the tank (necessarily folded, to get the drastic reduction in interior volume necessary to develop the hydrostatic pressure).  Conversely (via the same sort of mechanism in a hydraulic ram) even fairly small cracks or breaks in the shell will 'flow' enough oil to relieve the hydrostatic pressure below what would be necessary for structural 'bursting'.   (Of course, this might easily produce a well-carbureted spray at the break points, and that would pose a potential danger even for non-volatile crude, but that's not the failure you're describing...)

What you need to find is pictures of tank cars that have been 'crushed flat' but that don't show damage caused by the 'crushing flat' (like, for example, tank heads popped loose by physical crushing).  I have not seen such a picture.  Note that since you are the one making extraordinary claims, it is 'on you' to provide physical demonstration of the prospective failure mode; if it happens as often as you think it does, there should be no difficulty in procuring examples.