Oil Train

[MidlandMike]

Wizlish

...

I can easily see a political consensus building around giving railroads the authority to refuse 'automatic' carriage of HHFT or PIH traffic, and demand either a surcharge or additional 'named insured' coverage on specific policies to handle any such traffic.  ...

 

Hunter Harrison says he does not think CP should have to handle oil trains.  His RR in the Bakken region crosses a number of BNSF lines, and he could turn over the oil trains to BNSF, who would gladly take them.  However, CP keeps the long haul, and turns over the trains to eastern roads.  I don't know of any RRs that actually want to avoid these money makers.

[wanswheel]

http://slideplayer.com/slide/236563/

[BaltACD]

wanswheel

http://slideplayer.com/slide/236563/

 

The slide show doesn't explain the nomenclature for all the lines in the graphs displayed.

However, the slide show does show pictures of some of the hardware required.  I must say what is shown of the electrical trainline (that is handling 230 volts) has all the appearances of being toy like when thought of in light of day in, day out railroading.

Of course the slide show is presenting everything from the viewpoint of a vendor that is trying to sell their wares.

[Wizlish]

I am working up some notes and comment on the WABCO slide presentation from seven years ago, but wanted to note something else first.  There's a pretty good presentation on load sensing devices here:

http://slideplayer.com/slide/1592622/

I think this may contribute materially to the discussions we were having about these earlier in the thread, and in other threads that touched on empty/load proportioning of maximum brake effort.

 

[Euclid]

Wizlish, 

 

Thanks for posting your thoughts at the end of the previous page.  You are correct in your assertion that I have made a wrong assumption about the effect of the 6 seconds.  It would not mean that the ECP train stops 6 seconds sooner that the Westinghouse train.  There would be some difference, but it would be less than 6 seconds. 

 

For my term, propagation time, I was referring to just the time for the signal to pass through the train, to the point where the valves would start actuation.  The actuation of the valves from closed to open, and the flow of air would be encompassed in my term, transfer time.  The movement of the brake cylinder piston, levers, rods, brake beams, etc. would be outside of both terms.  

[Euclid]

Wanswheel,

 

Thanks for posting that Wabtec comparison of ECP brakes to conventional air brakes.  That is exactly what I have been looking for.  It shows that in the emergency application, ECP stops 14% quicker than a train with conventional air brakes and no EOT or DP enhancement.  So that would be the best ECP advantage for an emergency application.

 

Then it shows that in an emergency application, ECP stops 7% quicker than a train with some degree of EOT enhancement.  In their chart, it is the maximum enhancement.  I do not know if more enhancement could further reduce the difference in stopping time. 

 

Overall, however, it seems to back up the position of the AAR that the 7% advantage of ECP is relatively small. And likewise, it seems to diminsh the claims by Secretary of Transportation, Anthony Foxx wherein he suggested that the extra stopping power of ECP could help prevent collisions and mitigate pileup effect after a derailment begins.   

 

In regard to avoiding a collision, the total stopping distance of 2924 feet for best performing Westinghouse braked train would be reduced to 2728 feet for the ECP train.  Certainly the 196-ft advantage would help, but as a practical matter, it seems minimal.   

[Wizlish]

Euclid

For my term, propagation time, I was referring to just the time for the signal to pass through the train, to the point where the valves would start actuation.

It looks to me that WABTEC is using the term very similarly at one point in their slide presentation.  You might want to check this, and then post on how their understanding of the term is the same or (perhaps slightly but significantly?) different from your proposed use.  I am seeing something that may be highly interesting beginning to emerge from WABTEC's discussion, involving the use of sonic signals in the ECP 'main reservoir pressure only' trainline, that will explain the emergency-brake 'paradox', but I need to do more checking.

The actuation of the valves from closed to open, and the flow of air would be encompassed in my term, transfer time.

Scientifically: even if this appears a short time, call it out as a defined term (I'd propose 'actuation time' and define it between common points which I would naively propose as being the time the valve receives the activation signal and some proportion -- possibly full commanded port opening -- that the valve has responded.  If you define all the sequential 'factors' in the brake application this way, we will have both better clarity and better accounting of the specific delays or accelerations provided at each stage of the 'comparable' alternatives.  This immediately becomes valuable looking at a 'drink of water from a firehose' set of multiple alternatives like the choice bresented in the WABTEC graphics wanswheel presented from the slideshow.

The movement of the brake cylinder piston, levers, rods, brake beams, etc. would be outside of both terms.

What would you call it?  My "working" term was setup time.  There is likely to be a current term in railroading, but I don't know what it might be; a definition like 'taking up slack' or 'eliminating lost motion' doesn't get the right sense.  Whatever we call it, it's a definable factor and we should give it a specific name and a specific place in analysis even if it's presently 'common' to all kinds of foundation-braked car to which the Feds want to apply ECP in 2021.

[Wizlish]

Meanwhile, I want comments from the experts and the rails regarding a sort of 'elephant in the room' comment that no one so far has seemed to say anything about.

In the WSJ story on the mandated HHFT changes, Tom Simpson of the RSI said this:

"Speaking for the RSI Committee on Tank Cars ... We think that is an unnecessary expenditure because there are existing braking systems that work and can provide the same service at no additional cost to my industry.”

I want to see a technical discussion of the specifics of such systems, their cost, and whether they can provide the actual advantages that ECP offers to hazardous-material transportation ... preferably in direct comparison to both current and proposed methods of ECP implementation.  (Euclid, this SPECIFICALLY concerns your designs.)

[schlimm]

I am still wondering if service brake is safer (lower probability of derailment) than emergency braking?   ECP stops almost as fast on service (3225') as non-ECP emergency (2924-3166).  And the margin of ECP service compared to non-ECP service is large to huge (3748-6270).

[Wizlish]

schlimm

I am still wondering if service brake is safer (lower probability of derailment) than emergency braking?   ECP stops almost as fast on service (3225') as non-ECP emergency (2924-3166).

That is one of the precise points I was trying to establish. 

I think that the sequence of events in a emergency brake stop pose a number of significant hazards to train dynamics compared to a service application, particularly a service application that has proper load-proportioned braking effort per car, the capability of graduated and proportional release, and at least the option of proportional application or release on a car-by-car basis.

Now, whether the lawyers or Federal administrators will accept that 'safety' might involve a longer distance or time to come to a complete stop, but assuring a safe stop even in the presence of broken equipment or derailment, is another story.  I'd be prepared to bet on a sure thing that plaintiff's-bar organizations would line up to show that the Shortest Possible Crash Stop shoulda-woulda-coulda saved their client (in a collision) while equally cheerfully showing that Those Big Fireballs Wouldna Erupted if you hadn't slammed the brakes so hard in a derailment.

I'd be interested to read Dave Husman's response to this reasonably close distance, as I know he is one who believes in an emergency application to get the maximum possible 'way' off the train before accident kinetics begin to cause damage.  That point of view may still apply even though the apparent differential in stopping distance is small.

[wanswheel]

BaltACD

The slide show doesn't explain the nomenclature for all the lines in the graphs displayed.

 

Alphabet soup about various valves

http://techinfo.wabtec.com/DataFiles/Leaflets/ABDX%20ID%20Chart%20-%20Feb%202011.pdf

 

[wanswheel]

Excerpt from New York Air Brake brochure

http://www.nyab.com/media/banner/documents/downloadsservices/products_1/ep60_1/ep60.pdf

 

EP-60 is an electronically controlled pneumatic (ECP) train brake that provides a dramatic improvement in train handling for long, heavy haul trains…

The EP-60 design provides a unique feature which achieves a uniform brake ratio at all car loadings. This assures that light cars do not get over-braked. Improved train handling results from a reduction of in-train forces generated while braking a mix of empty, partially loaded, and fully loaded cars. The use of analog solenoid valves provides proportional brake cylinder control, as well as graduated application and release.

[dehusman]

First a couple terms.  A UDE is an undesired emergency which when the train goes in emergency it was not engineer induced.  Engineer induced is when the train is placed in emergency by the engineer.  Train induced emergency is when something other than the engineer induces the emergency application.

Most the studies appear to be describing an engineer induced emergency.  They include the time it takes for the application signal to go the length of the train.  With a train induced emergency the time will be less because the emergency application propagates out in both directions from the location in the train that caused the emergency application.  The cause is exactly in the middle, the propogation time will be about half that of the propagation time of a conventional configured train with an engineer induced emergency.  With a DPU it gets more complicated.  The first emergency application that reaches a DPU will signal the other DPU's to go into emergency and the emergency application will propagate out from each engine.

If it is an engineer induced emergency, then there is a a reason, usually a severe threat to life or property, that the engineer is trying to avoid or mitigate.  If your kids were on a bridge 2500 ft from where the engineer saw them and emergency could get the the train stopped in 2400 ft and service could get the train stopped in 2600 ft, probably you, your kids, the train crew and the railroad would rather get the train stopped in 2400 ft.

If its a train induced emergency for a "real" emergency situation (such as a derailment) then that means that something is bad wrong already,  the derailment is already happening, cars are on the ground and they are far enough out of line tha the air hoses have parted.  Stuff be happening.  Bad stuff.  At this point worrying whether an emergency application will cause the train to derail is not the primary concern because the train is already derailed.  At this point getting stuff stopped is the primary concern.

The place where an emergency is bad is when there isn't a problem and the train goes UDE.  With modern train placement rules the chances of a train derailing when it goes in emergency is much less (though not impossible).  I would think that a figure of 50 to 75 trains go UDE across the US every day would not be unreasonable.    If there were more than 10 or 20 derailments yearly nationwide as a result, I would be surprised. Sometimes a train goes UDE and then immediately recovers its air. 

Railroads spend lots of money eliminating UDE's.  Some roads have programs that compare the consists of all trains that go UDE and look for common cars, a very small percentage of cars cause a high percentage of UDE's.  If one car has been in 8 trains that went UDE, maybe that car needs to be looked at.  They spend lots of time improving the train lines and reliability of the brake connections  and components.  One of the latest things being tried is mounting the air hose to the side of the coupler so it is in a more consistent relationship to the other car, which reduces UDE's.

Solving the root cause eliminates the failure.

ECP will not prevent oil train fires.  ECP will not eliminate derailments.  It *might* reduce a derailment by a car or two.  Would it really make a difference if there are 6 cars in a pile and on fire or 7 cars in a pile and on fire?  I'm thinking not that much.  Zero cars in a pile or on fire is the goal.

[Euclid]

I do not agree with the AAR conclusion that ECP brakes offer nothing to improve oil train safety and nothing to prevent derailments.  I think that is extreme rationalization for the purpose of posturing against the regulators who have the power to impose requirements on the railroads.  When you don’t want to pay for something, it is easy to come up with arguments about why that something is not necessary.

 

These are the advantages of ECP brakes that would contribute to oil train safety:

 

Quicker Stopping In Emergency Application:

 

The 7% shorter stopping distance for an emergency application is not much, but it is better than nothing.  It will reduce the number of cars entering a pileup, reduce tank car breaching and slow fire development.  It will also reduce the chance of collisions.

 

Fewer UDEs:

 

The AAR says that ECP does nothing to prevent derailments.  Yet derailments can be caused by a UDE, so fewer UDEs will result in fewer derailments.

 

Reduced Slack Action and Reduced Buff and Draft Force:

 

Slack run-in and run-out can cause derailments directly.  Slack run-out can pull a drawbar that falls on the track and derails a train.  Slack action can also perturb a derailed dragging car into a jackknifing and pileup.  Buff and draft forces can develop without actual slack run-in or run-out, and have the same derailment causing effects.

 

***************

 

One factor entering into the AAR opposition to ECP brakes on oil trains might be the prospect of having that mandate ultimately expanded to a full system conversion to ECP.  After all, if safety is good for oil trains, it is good for all trains.

 

I am not sure of the exact numbers, but say the PTC mandate will cost 15-billion.  I understand the ECP tank car mandate will cost 2-3-billion.  What would be the cost of converting all locomotives and rolling stock to ECP?  Would it be 50-billion or higher?   

[Euclid]

Wizlish

I am working up some notes and comment on the WABCO slide presentation from seven years ago, but wanted to note something else first.  There's a pretty good presentation on load sensing devices here:

http://slideplayer.com/slide/1592622/

I think this may contribute materially to the discussions we were having about these earlier in the thread, and in other threads that touched on empty/load proportioning of maximum brake effort.

 

 

 

That is an amazing source of technical detail about empty/loaded sensors.  I wonder how much shorter the stopping distance of an oil train would be in an “Emergency” brake application if the train were equipped with these empty/loaded sensors, compared to the same train without the sensors.  I have a hunch that it would make a big difference; like maybe 50% or more.  Does anybody have a source to answer that question?

 

I also suspect that it would be far less costly to produce the effect of the empty/loaded sensors on an ECP braked train compared to a Westinghouse braked train.  It seems to me that the load sensing component of the empty/loaded sensor would not be needed with an ECP braked train operating as a unit train, as oil trains are.  You only need the actual sensor devices on trains that have a mix of loads and empties. 

 

All you would need to produce the effect of the loaded/empty sensors on an ECP oil train is a switchable brake force selector inside of the brake valve on each car.  One master switch would control that brake force selection for the whole train, based on whether it is loaded or empty.  It would instantly toggle the brake valve on each car simultaneously from a low brake force for empty trains to a high brake force for loaded trains.  The car brake valves in the ECP systems are signaled by wire electronics, so it would be easy to communicate a signal simultaneously to all the car brake valves to switch their brake force according to the loaded or empty status of the train.

 

This master selection valve could be operated manually by the engineer, or it probably could just be set automatically according to the data that accompanies the train shipment.  That way nobody could make a mistake with the switch setting.

 

What would not be needed for this empty/loaded brake force changer on an oil train with ECP is the load sensing device on each car.  There is no need for a tank car to sense its empty/loaded status and communicate it individually to the brake force selection valve on that car.     

 

I expect that this type of empty/loaded brake force selection system already exists for ECP applications as an available product option. 

 

We know that ECP brakes reduce stopping distance by 7% compared to the most enhanced Westinghouse system.  How much would the stopping distance be shortened by this empty/loaded switch combined with ECP?

[oltmannd]

Euclid

One master switch would control that brake force selection for the whole train, based on whether it is loaded or empty.

And, the first time someone forgot to turn "load" back to "empty"?  

Think that through and tell me what you think.

[oltmannd]

Euclid

What would be the cost of converting all locomotives and rolling stock to ECP?  Would it be 50-billion or higher?   

There are 1.7M freight cars in UMLER.  At $10K a pop, that's $17B.  Then 40,000 locomotives - guess $100k per = $4B.  And 70K EOTs -guess $5k per = $350M.  Plus training a couple hundred thousand employees - figure 40 hours at $100/hour = $800M.

Then, for at least the five year transition, you'd have to add about a minute to each outbound car inspection per car to make sure each car is set for ECP or non-ECP operation.  Guess at 20 trips per car per year for 1M cars, 5 inspections per trip, $50/hr ~ $100M/yr.

...provided you get everything right on the first pass.

[Euclid]

oltmannd

 

Euclid

One master switch would control that brake force selection for the whole train, based on whether it is loaded or empty.

 

 

And, the first time someone forgot to turn "load" back to "empty"?  

Think that through and tell me what you think.

 

Don,

 

I have thought that through, and I was referring to that issue above when I said:  “This master selection valve could be operated manually by the engineer, or it probably could just be set automatically according to the data that accompanies the train shipment. That way nobody could make a mistake with the switch setting.”

 

I agree with your point that the setting of the selector switch must be infallible. Yet, considering the incredible complexity and responsibility of running a train, certainly the need for setting the load/empty switch cannot be a showstopper.  But I do tend to agree that it may be too much to ask of one person to get it right all the time.  One mistake could flatten 800 wheels. 

 

I imagine that there must be a ton of data transmission that accompanies the movement of oil trains.  Surely this data would indicate whether the train is loaded or empty.  So I believe that data could be used to directly set the load/empty switch in order to remove the chance of human error.    

[zugmann]

Load/empty data is sent by humans. So human error won't be eliminated, just falls to a different department.

[Euclid]

Zugmann,

 

I am thinking that the data that is entered along with any train shipment will contain lots of duplication and redundancy, and also be the result of several different people adding the data.  So it may confirm the load/empty status in 100 different ways by multiple inputs.  Out of that, it ought to be possible to get an accurate answer to the question of whether the train is loaded or empty. 

 

As I mentioned, I think there is a good possibility that this system already exists as a standard optional feature if ECP.  Maybe I will call Wabtec and ask them.  It would seem like an incredible waste to equip a 100-car unit train with 100 mechanical sensors just to tell what is already known.  And, as somebody previously mentioned here, those mechanical sensors are a maintenance issue in addition to their outright cost.    

[zugmann]

I can tell in about 3 seconds if a unit train is loaded or empty. Can also tell you from the head end if that rear buffer on an empty unit train is loaded or not. That takes a little more than 3 seconds though.

[Euclid]

I know what you mean.  The locomotive can tell whether a train is loaded or empty, so maybe it could be set up where the locomotive sets the load/empty switch for the train if that is the only reliable method possible. 

[zugmann]

Locomotive can't tell nor does it care. The operator can, though. 

[Euclid]

I think locomotives could be modified in a way that they could tell whether the train is loaded or empty, but they would have to be under way in order to decide.  But that too would probably cost a trillion dollars.  In any case, I don't see inability to properly set the switch as being a showstopper. 

[tree68]

Euclid

I think locomotives could be modified in a way that they could tell whether the train is loaded or empty,

While strain gauges could be used to determine drawbar pull, it takes a seat in the seat to reckon out all the variables.  F'rinstance, if a train is draped over the crest of a hill (or several, in hogback country), it's going to be completely different than a train that's on a downgrade, or an upgrade.  

I'm sure a computer could could figure all of that out, but it would need to know the gradient of every foot of track, the loaded and empty weights of each potential car, and a whole lot of computing horsepower to crunch those numbers.  

I will grant that it would be relatively easy on a unit train, but on a manifest freight, the possible variations are virtually infinite.

I just looked at that slide presentation on stopping distances.  The difference for a full service stop is indeed impressive - on the order of 50%.

However, on page 8 of the same presentation, the difference between a full service stop and an emergency application is just 12%.  If some obstacle is within that 400 feet, ECP just saved the day.  But for the rest of the 2700 feet, not so much.

If we're talking a derailment (and we have been) and stacking cars up, ECP will make little difference in the end result.

[Euclid]

Yes the improvement in ECP stopping distance for a service application is substantial, and must be what the often repeated claim of up to 70% shorter stopping distance for ECP is based on.  In one of my above posts, based on the Wabtec chart, I did calculate that the stopping distance for an emergency application is 14% shorter for an ECP train compared to the base line, non-ECP train; and 7% shorter for an ECP train compared to their most enhanced non-ECP train.  I assume that is the basic train that the AAR is referring to when they argue that ECP will not help safety. 

 

Therefore, when the USDOT says that ECP will increase oil train safety due to shorter stopping distance; and when the AAR says the shorter stopping distance is negligible and not worth it; they are both referring to a stopping distance with ECP that is 7% shorter than the stopping distance of the same train with conventional air brakes that the AAR prefers for oil hauling.  So USDOT must be justifying the ECP mandate based on that 7% stopping advantage of ECP. 

[Euclid]

Regarding the load/empty switch for an ECP train, I would not propose that this switch needs to be set by the locomotive in sensing its load in pulling.  There has got to be a simpler way.  In any case, the train switch for load/empty is only possible with a unit train with ECP brakes.  A manifest train would require a mechanical load sensor for each car, and an addition to the “triple valve” of each car, so that each car could adjust its brake force according to whether it is loaded or empty. 

 

Now that the ECP mandate his here, we have a 7% reduction in stopping distance.  The simplified load/empty control that I have suggested does require ECP as the basic platform, and it would further reduce the stopping distance.  So maybe there is more benefit in the fact that ECP allows this simpler load/empty brake force control than there is in the fact that ECP alone stops quicker.    

[Wizlish]

Euclid

I also suspect that it would be far less costly to produce the effect of the empty/loaded sensors on an ECP braked train compared to a Westinghouse braked train.

Maybe so, but the effect is a bit different.  The ultimate purpose of any 'load/empty' sensor, as we discussed (in two senses, which converge to the 'right' answer), is to maximize braking effort to the point just before wheelslide occurs, for any given car weight or loading characteristics.  There is more precision to how an ECP valve can control that braking, therefore more reason to use a modern kind of load-sensing device on an ECP-equipped car, and implement the brake control on a car-by-car basis depending more or less automatically on the weight of each car at the moment.  I think the hardware (and programming) needed to accomplish this is only a slight addition to the existing cost of ECP equipment (or conversion of an existing car) and the additional maintenance issues would be relatively slight.  The alternative (as with some types of load sensing) would be to 'gag' the load sensor so that the brakes only produce the 'minimum safe' braking that does not cause wheelslide when the car is unloaded.

It seems to me that the load sensing component of the empty/loaded sensor would not be needed with an ECP braked train operating as a unit train, as oil trains are. You only need the actual sensor devices on trains that have a mix of loads and empties. All you would need to produce the effect of the loaded/empty sensors on an ECP oil train is a switchable brake force selector inside of the brake valve on each car. One master switch would control that brake force selection for the whole train, based on whether it is loaded or empty. It would instantly toggle the brake valve on each car simultaneously from a low brake force for empty trains to a high brake force for loaded trains.

I have a very different problem with this than Don Oltmann did.  All it takes is one leaking valve or punctured car, or one vandal releasing the contents, for a potentially very dangerous situation to arise.  And your system would have only a ponderous, and non-automatic, ability to deal with such a situation.

I have to wonder whether this system is a bit like automatic retainers, translating the retention of pressure in the brake cylinders into the degree to which the brake cylinders will be allowed to pressurize.  Thing is that you still need the whole electric trainline, resistance to breakage or interruption, signal integrity against improper actuation, etc. that is the lion's share of an ECP application, and all you have is a defective implementation of what simple mechanical devices on regular Westinghouse-braked cars can do.

 

The car brake valves in the ECP systems are signaled by wire electronics, so it would be easy to communicate a signal simultaneously to all the car brake valves to switch their brake force according to the loaded or empty status of the train.

Wouldn't this require an extension to the existing standards and protocols for ECP?  Not saying it couldn't happen ... but do you have any idea how to get a standards revision set up and achieved?

[BaltACD]

Fertilize the money tree!

[CMStPnP]

I think this latest unfunded mandate from the Feds is rather sad.   Much cheaper to the railroad industry if they allowed the railroads the ability to refuse to handle hazardous cargo or allow the railroads to charge a risk premium on the increased risk or liability they are taking on by hauling said hazardous cargo.    That would have been my approach.     The shipper should assume more of the risk here.