Oil Train

[Murphy Siding]

zugmann

 

 

schlimm

Based on the conclusion of the article above, safer practice for oil unit trains would be slower, which is what the FRA ordered. 

 

 

 

 

 

But remember this:  the refineries need so much oil a day.  If we slow down the trains, are we not going to need more trains to keep the supply constant?  If a refinery needs a loaded train a day, it doesn't matter if they are moving at 60mph or 30mph to the refinery, they will still need a loaded train a day.  

So now you have (approx) 2x the number of loaded trains on the main, albeit slowly crawling to the refinery. 

 

Now the question for those with a better undersatanding of probablility and statistics than me (just about anyone) - is that safer?   Is it better to have more slower trains, or fewer faster trains?

 

I do now know.

 

 

   I'll say C) More slower trains, but make them longer.  The article also suggests that cars further back in the train are less likely to derail.  Just add some more more further back there.

[Euclid]

schlimm

The number cars derailed (and where in the train) when all movement is finished.

 

Okay, I can then see why the probability drops toward the end of the train.  There would be fewer cars available to provide the combined kinetic energy to shove cars into the derailment.

[Randy Stahl]

It's true that a 10 mph derailment is not as bad as a 40 mph derailment for obvious reasons. I'm not saying that oil trains should plod along at slow speeds but as Shlimm pointed out, that is exactly what the FRA is saying and doing and with real facts to back it up, certainly more facts than the woman advocating ECP ! 

 

 

[schlimm]

zugmann

Now the question for those with a better undersatanding of probablility and statistics than me (just about anyone) - is that safer?   Is it better to have more slower trains, or fewer faster trains?

Speed is a larger contributor than length.  So I think you'd have to do a joint probability comparison.  I think that is what the FRA did.

[schlimm]

The number cars derailed (and where in the train) when all movement is finished.

[Euclid]

When then refer to the probability of cars to derail, are they referring only to first cars to derail; or are they referring to any cars involved in a mass derailment?  I can see a big difference.

[zugmann]

Randy Stahl

What Shlimm said sounds right, I don't have stats right in front of me but having been onsite for a couple hundred derailments it has the right feel.

I'm not saying it isn't right - just how useful is the stat?  

How do you use that number?

[zugmann]

schlimm

Based on the conclusion of the article above, safer practice for oil unit trains would be slower, which is what the FRA ordered. 

 

 

But remember this:  the refineries need so much oil a day.  If we slow down the trains, are we not going to need more trains to keep the supply constant?  If a refinery needs a loaded train a day, it doesn't matter if they are moving at 60mph or 30mph to the refinery, they will still need a loaded train a day.  

So now you have (approx) 2x the number of loaded trains on the main, albeit slowly crawling to the refinery. 

 

Now the question for those with a better undersatanding of probablility and statistics than me (just about anyone) - is that safer?   Is it better to have more slower trains, or fewer faster trains?

 

I do now know.

 

[Randy Stahl]

schlimm

 

zugmann

I'm nto saying the statistic is wrong, but the way it is presented may be misleading.

 

 

It was pretty clear, but check the article linked above.  The article is fairly short.

 

 

What Shlimm said sounds right, I don't have stats right in front of me but having been onsite for a couple hundred derailments it has the right feel.

 

[schlimm]

Based on the conclusion of the article above, safer practice for oil unit trains would be slower, which is what the FRA ordered. 

[Euclid]

Dave,

I understand your point about item #5.  The only time it would apply is if the brake hoses parted prior to the cars separating because once they separate, jackknifing would begin.  So it would be too late for an ECP caused subdued slack run-in to help.  And I can see your point that most of the time, the hoses will not part until the cars separate.    

[schlimm]

zugmann

I'm nto saying the statistic is wrong, but the way it is presented may be misleading.

It was pretty clear, but check the article linked above.  The article is fairly short.

[zugmann]

schlimm

 

"Cars positioned near the front or rear of a train have the lowest probability of being derailed in a derailment. As train length is decreased or train speed is increased, the conditional probability of derailment increases for all cars within the train consist." 

 

I'm wondering how they figured that one out.

If you have a 5 car train at any speed and hit a broken rail, there's a good chance you're going to have all 5 cars derailed.  

 

While a 50 car train at the same speed, the rear cars are going to stop long before they hit the point of the broken rail.

 

 I'm nto saying the statistic is wrong, but the way it is presented may be misleading.

[schlimm]

 

"Cars positioned near the front or rear of a train have the lowest probability of being derailed in a derailment. As train length is decreased or train speed is increased, the conditional probability of derailment increases for all cars within the train consist." 

[tree68]

Here's your derailment dynamics right here:

https://www.youtube.com/watch?v=a-smEEHYdGQ

This is the well known video of the train hit by a tornado.  What's important for this discussion is at the end - where the trailing part of the train, whose brake line is clearly disconnected (which means the trailing part of the train is in emergency) piles into the locomotives.  Note that the car that hits the locomotive is also derailed and being pushed by the consist behind it.  Note, too,  the accordian pile-up that follows.

 

[Randy Stahl]

The biggest problem are the double shelf couplers. I've seen many minor derailments turned into major derailments because the cars would not come apart and pop the air hoses apart, double shelf couplers insure that most (if not all) of the train will derail. If the idea is to get the train into emergency and the brakes applied this is a problem. I'd rather have the cars come apart and leave at least some of the cars on the track.

 

Now I understand why they are there but I don't think it was well thought out.

[dehusman]

Euclid

Item # 5: Reduction of jackknifing and pileup due to a delay in its onset, because of the elimination of severe slack run-in caused by the PCP sequential application (such run-in depending on where the derailment is located in the train).   

 

Only if the run in is causing the jacknifing.  Based on my experience and all the reports I've read, that has not been the case.

 

Here's a thought.

 

The derailment is detected when the train goes in emergency (regardless of brake system.) 

What causes the emergency brake application?  Normally the brake application happens when the air hoses between the cars become uncoupled. 

What has to happen to get the air hoses to uncouple? Normally either the air hoses are dragging on the ground or the cars are at such an angle to each other that the hoses are pulled apart.

If the cars are at an angle to each other great enough to cause the air hoses to come apart, the baby is pretty much born, they are in the process of jacknifing already. 

Since air hoses are generally high enough to clear the top of rail when the car is on the rail, they air hoses are high enough to clear the top of the ties when derailed.  The only way they come uncoupled is to catch on something on the top of the tie.  If that is a switch or a road crossing then the impact may knock the wheels out from under the car, cause the derailed car to go off the ends of the ties or cause the car to be skewed to one side or the other.  Pretty much in each of those cases, the jacknife is in progress.

It is only in the very small group of derailments where the derailed car remains upright, in line and on the ties there the slack might be a contributing cause to the pileup.

 

ECP doesn't stop slack, it reduces it.  ECP doesn't apply the brakes any harder, it just applies them at near the same time.  In a mixed freight train there will still be slack because different weight cars will stop at different rates.  In a unit train the stopping will be be much more uniform because the cars are typically a near uniform weight, and may have similar brake systems with similar brake shoes.

 

You are over emphasizing the role slack plays in the train dynamics of the derailment and the jacknifing.

 

 

[schlimm]

jeffhergert

I have no statistics

 

"In this paper, we have examined the probabilities of derailment for freight trains and freight cars as affected by train length, train speed, and positioning of cars within the consist. "

 http://railtec.illinois.edu/CEE/pdf/Conference%20Proceedings/2005/Anderson%20and%20Barkan%202005.pdf

and another:

http://www.afpm.org/uploadedFiles/Content/Policy_Positions/Agency_Comments/Documents/AFPM%20CBR%20Comments%20and%20Exhibits.pdf

[BroadwayLion]

Euclid

Regarding my item #3: The reduction in jackknifing would result from delaying its onset to the extent that it is caused by more braking ahead of the derailment than behind it, due to sequential application of PCP brakes. Any delay in the onset of jackknifing gives more time for the braking and the derailed cars to dissipate kinetic energy; so once jackknifing begins, its duration will be shorter, and the pileup will include fewer cars.

In experience of LION (for whatever that is worth) trains do not derail because they jackknife, they jackknife because they derailed. Most deralments (that I have read about) are probably laid at the doorstep of MOW (collisions notwithstanding).

Something in the middle derails, and you have a jackknifed accordion before any emergency brakes (to say nothing of service brakes) can begin to apply. Thus the LION's *SUGGESTION* of Guard rails (inside the gauge) and guard timbers (outside of the gauge). You tell me this is not practicle, and that it will not work anyway. OK, LION can accept that, scratch that Idea. Does anybody have any other ideas?

PS. It works on railroad of LION, but then what is railroad of LION except a toy train up in the attic.

 

ROAR

[zugmann]

I wonder if the ECP brake movement kind of went out the window due to more AC engines with extended range dynamics being added to the roster.  Those things are amazing.  

If you have older DC engines and are going down a grade, even with full dynamics applied, there are many times you have to squeeze 10# or so on.  The newer AC engines?  Half the time you have to back off the dynamics so you don't stop.  

Granted I don't work in any severre grade territory, but we have a couple hills that have been known to stall trains out.

[Euclid]

Dave,

I understand your points.  My item #4 just means that the trains stops quicker so the hind end cars have less time to feed cars into the derailment.

Regarding my item #3: The reduction in jackknifing would result from delaying its onset to the extent that it is caused by more braking ahead of the derailment than behind it, due to sequential application of PCP brakes.  Any delay in the onset of jackknifing gives more time for the braking and the derailed cars to dissipate kinetic energy; so once jackknifing begins, its duration will be shorter, and the pileup will include fewer cars.     

[BroadwayLion]

jeffhergert

I don't want to even think about cars with special couplers that won't interface with normal couplers. Set one of those babies out on line, and you are either going to have to have the capacity to install a temporary compatible normal coupler or hope a train with the same equipment comes along soon after it's repaired. Some places that might not be a problem. Other places, it might be. Some routes might only see a unit train once a week, or the destination may only receive a train that uses a particular route weekly. (An originating point may send trains to a few different receivers. A receiving point may get trains from different originating points.)

Of the transit type couplers espoused by the LION, they couple and uncouple in the field just like any other coupler. It is a coupler after all! You use a key in stead of a bar to open it. No big deal, any CR can do it. Each car does have a "mini=knuckle" that fits into the coupler to allow connection to other equipment. --[It *is* Railroad equipment, designed and used by railroaders == they know how to make things work.]--

The ones that I have seen on the LIRR or MNCR are rather small things, so you will not go fast or far or pull a heavy load, but you can get a car back to the nearest yard if you tack it on to the back of any train.

You can see a pair of such adapter knuckles hanging to the left of the coupler on this LIRR service/protect locomotive.

ROAR

[jeffhergert]

I have no statistics, but I think most cars that derail because of severe slack action, whether from train handling or an emergency application, will be an empty car.  Even then, it's going to be an empty that's in a train consisting of mixed loaded and empty cars.

As to slack, I just don't think there's that much in an oil train.  I think going to draw bar connected cars isn't going to gain you that much.  It could cause more headache than it's worth if you make the blocks of drawbar connected cars to big.  Not from needing slack to start (I've only had to take slack 3 or 4 times since I went into engine service in 2004. All of those times were from having an unplanned stop on the side of a heavy [for us "flat landers" anyway] grade with a heavy train.), but from what to do with the blocks/train when a detector sniffs out a defect that needs to be set out.  Plant rationalization has reduced a lot of set out points to just a few cars, assuming MOW doesn't have them locked out with machinery.  Even if you can set out the block, all the other loads in good cars are delayed.  Not only from sitting waiting for the car men to fix the defect, but also now those cars are most likely going to continue to their destination using the manifest (loose car) system.  Instead of one train to their destination, they may now see a few yards and trains to get their.

I don't want to even think about cars with special couplers that won't interface with normal couplers.  Set one of those babies out on line, and you are either going to have to have the capacity to install a temporary compatible normal coupler or hope a train with the same equipment comes along soon after it's repaired.  Some places that might not be a problem.  Other places, it might be.  Some routes might only see a unit train once a week, or the destination may only receive a train that uses a particular route weekly.  (An originating point may send trains to a few different receivers.  A receiving point may get trains from different originating points.)

Jeff

 

[Euclid]

I understand that once the emergency application is made, there will be no difference between the effect of ECP versus PCP brakes.  Whether there is enough difference prior to an emergency application to justify ECP brakes is the pertinent question.  No single solution is going to be the answer.  So, as the article says, PHSMA is seeking a multi-faceted solution.  In my opinion, ECP brakes will offer a vast improvement compared to the improvements in the 1232 tank car that have been implemented. 

In my list of ECP brake advantages above, I speculate that item #3 will be the most advantageous.  I will also add to the list item #5, which I think would be the second most advantageous factor. 

Item # 5: Reduction of jackknifing and pileup due to a delay in its onset, because of the elimination of severe slack run-in caused by the PCP sequential application (such run-in depending on where the derailment is located in the train).   

[Buslist]

dehusman

A 100 car train is roughly 6000 feet long.  The emergency application propogates at about 900 ft/sec.  Since the emergency propogates in both directions from a train separation, it effectively moves at about 1800 ft/sec.  That takes about 3-4 sec to propogate through the entire train.  Since the brakes are setting while that is happening, there is some brake application during that period.  Assuming a linear propogation, that is effectively 1/2 braking force for 3-4 seconds, then full braking after that.

Assuming 40 mph, that's about 1 car per second (59 ft/sec).  Assuming that the braking effort while setting has no effect on the speed of the train, the train will move about 4 car lengths during that 4 sec.

The millions to buy ECP is buying you about 4 sec of increasing the brake force from half to full effort and about 3-4 cars, 200-250 ft of distance before the brakes begin applying. 

Once the brakes are fully applied, the braking rates between the conventional and ECP trains will be nearly identical.

 

 

Oh come on who cares about facts?

If you ever listened to Car Talk on NPR the Click and Clack brothers had an interesting question. When someone posts an erroneous opinion in a blog and someone else posts an equally errouneous response, do we know more or less as a result of the posts?

[dehusman]

Euclid

1)    Fewer derailments due to fewer undesired emergency applications.

I don't see that since there is still all the parts of the conventional brake system in place, most of the UDE's are caused by failures in the air hoses and train lines, and theose will remain.  It actually adds one more route of failure, the control system for the ECP system, so the opportunity for failure is greater with the ECP system than without it.

2)    Fewer derailments due to better slack control, which results from brakes applying simultaneously on all cars.

 

I agree for cases when the train air is being used, there will be reduced slack (but not eliminated since gravity trumps ECP), in cases where dynamic braking is being used, the performance and slack would be the same since ECP has no interaction with ECP (unless it is some form of blended braking).

 

3)    Reduction of jackknifing and pileup due to simultaneous brake application ahead of and behind the derailment.

Don't understand this one since if the train derails, the emergency applicationis automaticaly propogated in both directions, front and rear simultaneously, with conventional brakes.  How that prevents jacknifing, I don't understand.  The jackniking is caused by the rear of the train running into the derailed cars, not necessarily the head end cars.  ECP will not prevent the derailed cars from stopping.  It will incrementally start the brake application quicker in the rear portion, but the brakeing curve will be the same (sme brake valve, same brake pressure, same brake shoes, same wheels, same friction on the same rails).

 

4)    Reduction of jackknifing and pileup due to reduction of the stopping time, which is due to the simultaneous application. 

The jacknifing is due to the derailed cars stopping more quickly than the rear of the train (higher coefficient of friction shoving a car across the dirt than on the rail).  A very common arrangement after a derailment is the head end, on the rails, one or two derailed cars attached to the head end, a gap in the train, the pile of cars, then the rear of the trains.  The pile was caused by the rear running into the derailed cars, not the head end.

 

 

 

[/quote]

[dehusman]

A 100 car train is roughly 6000 feet long.  The emergency application propogates at about 900 ft/sec.  Since the emergency propogates in both directions from a train separation, it effectively moves at about 1800 ft/sec.  That takes about 3-4 sec to propogate through the entire train.  Since the brakes are setting while that is happening, there is some brake application during that period.  Assuming a linear propogation, that is effectively 1/2 braking force for 3-4 seconds, then full braking after that.

Assuming 40 mph, that's about 1 car per second (59 ft/sec).  Assuming that the braking effort while setting has no effect on the speed of the train, the train will move about 4 car lengths during that 4 sec.

The millions to buy ECP is buying you about 4 sec of increasing the brake force from half to full effort and about 3-4 cars, 200-250 ft of distance before the brakes begin applying. 

Once the brakes are fully applied, the braking rates between the conventional and ECP trains will be nearly identical.

[edblysard]

 Once the train is in emergency braking, no matter what type of

brake system put it there, the remaining actions/reactions will depend totally on the weight, speed and mass of the train.

EPC brakes are great at helping control braking during normal train movement, but so far, most carriers have decided that the small increase in performance do not justify the expense of the system.

In emergency situations, they do little to mitigate the damage....once the train is in emergency, and the brakes are on completely you are at the mercy of Newton's third law.

Once the friction overcomes the forward momentum, the train will begin to stop.

EPC will do little to aid that.

I guess the Westinghouse system, working at the speed of sound, isn't fast enough?

[Euclid]

ECP brakes offer the following advantages over PCP brakes:

 

1)    Fewer derailments due to fewer undesired emergency applications.

 

2)    Fewer derailments due to better slack control, which results from brakes applying simultaneously on all cars.

 

3)    Reduction of jackknifing and pileup due to simultaneous brake application ahead of and behind the derailment.

 

4)    Reduction of jackknifing and pileup due to reduction of the stopping time, which is due to the simultaneous application.

 

5)    Reduction of jackknifing and pileup due to a delay in its onset, because of the elimination of severe slack run-in (such run-in depending on where the derailment is located in the train).   

 

 

I agree that these advantages might only reduce the number of derailed cars, and it only takes one to start a fire.  But the fewer cars in the heap, the smaller the fire will be.  This is because fire can spread easily in the heaped cars stacked like cordwood in a fireplace, but not as easily to cars strung out in a line. 

[Norm48327]

"The big cause for these incidents when they happen is you get the pileup, all the cars run into each other, and crash one upon the other upon the other. These new ECP brakes will have control of each tank car so you won't have that kind of pileup."

Her lack of knowledge of the laws of physics is obvious.