Transcon: Delayed Trains

[Overmod]

At least some of the push for ECP in the early days of Feinberg's tenure at FRA used the same secret methodology as the use of artificial NOx standards in Tier 4 final: to "induce" the railroads to start making changes undesired by them without requiring a funded mandate.  (In the pollution case, policy at EPA was toward requiring SCR/DEF on all locomotives, which of course railroads have been resisting for all the applicable reasons, and in my opinion this was central in their denying EMD's waiver application for the 2.5% exceeding of NOx emissions on a small outlying part of the test cycle that their EGR-based solution could provide.)

As I recall, this was back in the Blast Zone days before NIHSA essentially solved the exploding oil-train issue correctly, by degassing the crude before it went in takn cars in the first place.  The discussions at FRA did clearly understand that the advantages of ECP were primarily in service braking, and they did understand the value of simultaneous application of service brakes as well as proportional release.  Perhaps "unfortunately" the effective solution of the explosions occurred before ECP could be encouraged or mandated on revised stock, or on new or rebuilt cars during their (re)construction.

I continue to think that ECP conversion via the manufacturer's option should take place funded by dollar-for-dollar setaside of income or other tax paid by a railroad or private shipper.  This expense comes to only a couple of thousands of dollars per car, and the car then operates normally as a one-pipe Westinghouse car until converted, which I believe is at most a 20-minute process including testing.  This work could easily be done at the same time a 'power parking brake' system is installed on a given car, which I understand is now mandated in Canada.

 

[Euclid]

Apparently due to the need to convert the entire North American railcar fleet and locomotive roster to ECP in order for universal interchangeable equipment operation in trains, the railroad industry sees an enormous cost with too little benefit.  In my opinion, the industry’s total disdain for ECP brakes is the perfect opposite for their love of Autonomous freight trains, and their already adopted practice of monster trains made possible by DPU. 

 

Around 2015, we almost had a federal mandate requiring ECP brakes on crude oil trains.  At that time, the industry’s position on ECP became clear.  Unfortunately, the case for ECP had not been clearly made because it focused, as it usually does, only on stopping trains quicker than conventional PCP braking.  And that ECP advantage has been hopelessly discredited due the long running omission of the fact that the advantage of quicker stopping only applies to trains making a “Service” application and not to those making an “Emergency” application.  Yet, it is the Emergency application that needs the quickest possible stopping, and not the Service application.   

 

But despite that omission, perhaps the greatest fundamental benefit of ECP braking is its simultaneous brake application.  This better controls and mitigates “in-train” forces that can result from the time for brake application propagation associated with conventional PCP brakes.  Brakes setting up sequentially over a period of time can cause hard slack action, which can derail trains.  This is perhaps the greatest threat to oil trains, and ECP basically eliminates that threat. 

 

But, instead, the railroad industry confines their argument only on the point that ECP has no advantage over PCP in stopping quickly to prevent collisions.  Yet that is completely beside the point of requiring ECP brakes on oil trains. 

[BaltACD]

blue streak 1

So   what happens to cars in Mexico especially in south Mexico and even a very few even further south ?  There are many more cars going there now with the new  car ferrys with much higher capacity,

Cars that go to Mexico eventually return to the USA.  Maybe they are 'picked clean', maybe they aren't, same as today.

[blue streak 1]

So   what happens to cars in Mexico especially in south Mexico and even a very few even further south ?  There are many more cars going there now with the new  car ferrys with much higher capacity,

[Overmod]

The idea of contactless ECP actuation has taken two relatively similar forms: the use of distributed radio, and the use of hardwired harnesses with redundant short-range wireless connection between cars.  Both of them have obvious potential problems 'in the real world' and I suspect the actual system used would involve both (with the wireless commands riding on the DPU control signal)

Some of the SDR technology used for PTC might be applicable to in-train communication and solve difficulties of 'crosstalk' error between trains.

[Euclid]

Overmod

The issue of the 220V trainline with powerline data modulation vs. distributed battery power and wireless links via short or long-range radio is interesting technically,...

My understanding that is the first approach you mention is what has been put into use with users such as Rio Tinto, and others, if any.  So that was the original concept with ECP development. 

 

But then I have also heard generally about the second approach you mention, which may equip each railcar with storage batteries, solar panels, and generators driven by railcar wheels. 

 

So each railcar has its own onboard source of electric power and an electrically powered control that actuates the pneumatic control valve. 

 

Then as that valve is actuated, it admits or vents air pressure to apply or release the airbrake pneumatic cylinder on each railcar, which applies or releases the car brakes. 

 

With this approach, there must also be a means for the engineer in the locomotive cab to actuate the electrically powered control valve on each railcar.  What is that means used by the engineer to collectively actuate the electrically powered brake control valve on each railcar?

 

Has this second concept ever been used in a working ECP system?  What advantage does it have over the first of the two concepts you mentioned above?

 

[Overmod]

The issue of convertibility has been discussed extensively by the manufacturers.  The problem is not with the ECP-equipped devices, but with the evolution of one-pipe Westinghouse, most particularly the introduction of pressure-maintaining (which eliminated any chance for proportional release).

Current 'state of the art' is to provide equipment that can easily be changed in the field between ECP actuation and traditional Westinghouse.  The idea both manufacturers had was that equipment would be gradually converted and then, as complete sets of dedicated cars and then lanes of interchange developed, on a particular 'flag day' the equipment would be cut over to use the ECP.  For legal reasons no one should have trouble guessing, I doubt we will see a return of the old 1908-era practice of equipping unbraked cars with 'through' brakelines and putting unbraked cars on the rear of the brakepipe-equipped stock...

The issue of the 220V trainline with powerline data modulation vs. distributed battery power and wireless links via short or long-range radio is interesting technically, but I can't imagine depending on individual car operation to implement braking safely.  A more likely approach is to use distributed battery power for the power 'parking brake' but you would NOT want this automagically linked to brake applications!  (In fact you'd have to take pains to ensure that rocks, teenagers, etc. did not have free access to the power-brake application switches on individual cars, while still ensuring full crew access without cards or keys -- not a trivial problem!) 

[Euclid]

challenger3980

Would it work to use a midtrain DPU, then have all the ECP cars trail the lead power, and have all the PCP cars trail the DPU?

It sounds like a PITA, especially for Local service, but as a transitional measure, would that work?

Doug 

Well, as you know, the overall objective would be to convert all equipment to ECP, but be able to do that over a reasonably ample period of time, so that no single car has to be out of service for any more time than it takes to convert it.  One way to do that would be to convert every car to dual purpose braking with a combination of both ECP and traditional PCP brakes. 

 

Then as these dual purpose cars show up in trains, the majority of cars will still be straight PCP.  So you set the minority of cars being dual purpose to the PCP purpose.  So, then when the train runs, it will run with 100% PCP.  Some cars will only have PCP and will use that.  New dual purpose cars will be set to run with PCP.

 

Eventually, still before all cars are converted, some trains will, by chance, have all cars having the dual purpose brake system.  In that case, they will all be switched to their ECP settings, and the whole train will run with ECP.

 

Eventually all cars will have been converted to the dual purpose brakes.  Then as those brakes need maintenance or replacement, they will be replaced with dedicated ECP brakes without the dual purpose feature.  So for some long time of that conversion, trains will be running with a combination of pure ECP systems and the older dual purpose systems set to ECP operation.  Eventually, over a long time, as the dual purpose systems are replaced during routine maintenance and repair work, the entire North American fleet will be pure ECP.

 

However, what you are suggesting is that, as cars are converted to ECP, they are immediacy able to run with active ECP as their braking system while in a train largely consisting of traditional PCP. 

 

One benefit if this would be to be able to get the improved braking work out of the ECP cars without waiting for the day when trains are mostly all ECP cars.  So with this, a train would be running with some cars using ECP and some using PCP.  Blending these two systems in one train would be problematic because the performance speed difference that Balt mentioned.   

 

But even though ECP can set faster than PCP, I think it could be programed to set slower in order to synchronize with the setting of PCP.  A train operating this way would have both brake systems, and to the extent that it had ECP cars, it would have an overall better brake system operating than if had only PCP brakes.  Thuse the added benefit of ECP brakes could begin to be reaped earlier in the conversion process.  Although I suspect this combination of ECP/PCP might have complications even it the ECP applciation were slowed down as a fixed condtion.  It may have to be programed for each specific train according to the exact number of PCP and ECP cars.  And even the distrubuition of these cars might have to be controlled for overall performance.       

 

[BaltACD]

Pneumatic braking operates at the speed of sound.  Electronic Controlled braking operates at the speed of light.

To my mind, it should be POSSIBLE for someone to design an car air brake valve that can operate with both operating conditions.  The electronics could be powered by some form of solar panel with a small rechargable battery being a part of the air brake valve itself.  Charging could also be done with some form of car mounted generator/alternator, powered by the air flow necessary in charging up the overall brake system on the train (many of the 2-way EOT's being used are already air powered to charge their own battery).  Each air brake valve could also contain a small radio transmitter, similar in concept to Bluetooth to act as a repeater to propagate brake valve operation as commanded by the engineer.

To my knowledge no one is working on such a brake valve.  Change of this character has to be both forward and backward compatible for the industry to accept the investment necessary to implement it.

[challenger3980]

Would it work to use a midtrain DPU, then have all the ECP cars trail the lead power, and have all the PCP cars trail the DPU?

It sounds like a PITA, especially for Local service, but as a transitional measure, would that work?

 

 Doug 

[Euclid]

blue streak 1

not necessary for an all or nothing.  Maybe just the ECP equipped cars run next to front locos equipped and DPU consists as well.

Well, the ECP and PCP systems cannot share the same pneumatic trainline because the PCP line supplies air pressure to charge the car reservoirs and to act as a control line for operating the brake valve on each car by reducing and recharging the line.  Whereas the ECP pneumatic line only supplies air pressure to charge the reservoirs.

 

So there would need to be a PCP brakepie trainline extending throughout the train to serve the blocks of cars equipped with PCP brakes, and then each block of ECP cars would extend from the lead locomotive and also from each DPU, to the end of the ECP cars connected to those individual sets of locomotive units.

 

I guess you would just equip the ECP cars with an ECP pneumatic trainline, and equip the PCP cars with a conventional PCP trainline/brakepipe.   

 

But aside from all that, I am not sure how those two different braking systems could be coodinated in one train.  The concept would be similar to how they blocked airbrake equipped cars within a train otherwise having only hand brakes back when airbrakes were being gradually added to all rolling stock.

[blue streak 1]

not necessary for an all or nothing.  Maybe just the ECP equipped cars run next to front locos equipped and DPU consists as well.

[Euclid]

charlie hebdo

[from Global Railway Review]

Hitachi Rail is a world leader in autonomous and signalling technology and recently, also in Australia, began a major contract to install innovative technology to automate elements of Queensland’s New Generation Rail (NGR) fleet. The $107 million contract will see Hitachi Rail install Automatic Train Operation over European Train Control System Level 2 technology on all NGR trains.

 

I don't know if that trade mag is a reliable source. However, if so, one can infer that the existing Autohaul operation on Rio Tinto has not been plagued with problems.

 

The investment in an ECP brake system like that must have cost billions of dollars.  So I think they would have found out ahead of time whether it would be plagued with operating problems.  If they found it would be, they would not have bought it.  It is not like buying a used car.  I am sure the manufacturers have a stake in the reliable performance of the ECP equipment. 

 

The primary problem with converting to ECP is that it is an all or nothing proposition; meaning you have to convert all of the rolling stock and locomotives before it can be put into use within the system that is entirely standardized with interchangeable rolling stock equipped entirely with PCP braking.  This requires a near instantaneous conversion of the entire North American fleet of rolling stock plus many locomotives as well.  The only solution to that problem is make the change by installing dual purpose braking, so that trains can operate with each of their consists having a mix of ECP and PCP braking.   

 

That would perfectly solve the practical problem by allowing the conversion of equipment take place over time at a reasonable pace rather than converting all at once. 

 

But the problem with the dual purpose brake equipment is that it will greatly increase the cost over single purpose ECP braking.  And then once the dual purpose system is fully installed on all equipment, it will be unnecessary.  Also, as this first conversion to dual purpose ECP/PCP progresses, there may be potential improvements found desirable for the ECP side, but then held back because the PCP side interferes with the improvement; and that side is only temporary.  So it may be better to delay the improvement to the ECP side until the temporary PCP side is phased out. 

 

There may also be problems involved with mixing the two types of braking within one train consist.  Would train makeup have to be adjusted for this mix of braking types?

[charlie hebdo]

[from Global Railway Review]

Hitachi Rail is a world leader in autonomous and signalling technology and recently, also in Australia, began a major contract to install innovative technology to automate elements of Queensland’s New Generation Rail (NGR) fleet. The $107 million contract will see Hitachi Rail install Automatic Train Operation over European Train Control System Level 2 technology on all NGR trains.

 

I don't know if that trade mag is a reliable source. However, if so, one can infer that the existing Autohaul operation on Rio Tinto has not been plagued with problems.

[Euclid]

tree68

Even with fixed consists - you've got to deal with potential detours and the like.  And the possibilities with a train of PRB coal headed to South Carolina are endless...

 

Detours would be a major contingency, so I would expect detouring trains would be manually operated.  In any case, I am sure railroad companies realize that a change this complicated will not come without some teething problems.  I would expect them to begin with very limited operations of ideal trains on the most ideal runs to see how it works. 

[jeffhergert]

I've discussed with co-workers a couple of things the railroads might try.

One.  Get the Government to temporarily suspend the RSIA that caps monthly working hours, requires undesturbed rest and 48 or 72 hours off after working 6 or 7 consecutive days.

One A.  Temporarily put off rest day requirements in the new labor agreements.

Two.  Temporarily increase the Hours of Service back to 16 hours a day.

I'm sure their argument would be the service crisis requires more work out of the employees they have until they can hire enough.  Of course the crisis, of their own making, will never be resolved.

Jeff 

[jeffhergert]

Euclid

Broadway Lion said:

 

4. Move to automatic operation with no crews on certain trains. My model railroad (a subway layout) operates across twelve scale miles of track and I keep ten trains running at once.

 

Does anybody believe this is impossible?  Certainly the technology for autonomous trains is proven to work.  For trains, it is fundamentally easier than for road vehicles.   The railroad application is half way there with its mechanically self-guiding track.  Whereas cars and trucks need to assimilate a wide field of ever-changing features, and make constant decisions about whether it is safe to proceed. 

 

This challenge is non-existent with trains.  In my opinion, full autonomous driving of cars and trucks is overpromised for marketing purposes. Ultimately I expect it will begin with only certain roadways which are sufficiently upgraded to play a part in the autonomous concept.

 

For railroads, the main obstacle is to fit all of the complex train handling moves into the autonomous program.  I think that is why Lion confines his prediction to applying automatic running to “certain trains.”  In other words you start with certain trains, and move forward with further development. 

 

People always shoot down the viable comparison of U.S. railroading to Rio Tinto by saying the two are not comparable.  Rio Tinto proves that the lack of comparability is not train size.  Instead the lack of comparability is move complexity, which is relatively simple with Rio Tinto.  With autonomous running of U.S. railroads, the changes in train operation will be larger than the change to running trains autonomously. 

 

Has anyone ever seen anything about Rio Tinto's automated train failure rate?  They, nor their suppliers are going to want that public so they're not going to put out the numbers.  But someone from that area has to have an idea of what it is. 

And there has to be one, and it may be within their comfort zone.  Not necessarily catastrophic failures, but glitches that require a human to reset.  We use some of the same equipment for our auto throttle energy management systems.  And it often has glitches, usually just disengaging itself for no reason and then after a half mile (or more) wants control back.  Rarely, but it has happened, it will disengage without giving control back to the engineer. 

The easiest trains for automation to run are trains that are almost the same all the time.  Unit trains that have the same type of equipment, empties being handled better than loads.  Like with a human (actually most humans are better) the shorter the train the better auto throttle handles a train.

Jeff

[tree68]

Even with fixed consists - you've got to deal with potential detours and the like.  And the possibilities with a train of PRB coal headed to South Carolina are endless...

[Euclid]

Sure, there will be plenty of labor needed to move trains even after converting to autonomous running, which only eliminates train crews over the road.  And even that is not going to happen all at once like dieselization.  All contingencies such as broken knuckles and other technical failures will have to be dealt with by ground-based personnel. 

 

I believe that the management side of railroading wants autonomous trains for the same reasons they have always wanted reductions in crew size down to one person.  Now they see the temptation and promise of going to zero persons at the cab controls, and operating by a program.   

 

Fundamentally, autonomous operation is easier to do with trains than with road vehicles, because road vehicles require constant steering adjustment in order to follow a path.  On the contrary, trains are automatically steered by the path. 

 

For this reason, trains facing a myriad of variable operational conditions and choices will be as difficult as, or more difficult to automate than road vehicles; so trains will be selected for automation on the basis of their individual routing and work simplicity of operation.  I assume that this is why Broadway Lion said: “Move to automatic operation with no crews on certain trains.”  I think these must be trains that are as point-to-point as possible, and having the fewest contingencies. 

 

For automation to continue into the realm of the trains facing more complex travel patterns, there will have to be a reduction of those complications.  I believe that resolving this challenge will be the biggest obstacle facing autonomous running.  It will require eliminating the routing complications as much as possible.  I think this will be aided by trains becoming shorter and lighter, as permitted by autonomous operation.  So the success of full autonomous operation will force massive operational changes in yard and terminal handling, and in ground based emergency support.    

 

These operation changes are likely to be a larger project than the actual automation of train movements. 

[Backshop]

There are still crew costs to autonomous trains.  Who assembles them and yards them?  Who does repairs on the road? Remember, the Australian ore trains are captive operations with set consists, these aren't.

[Euclid]

Psychot

 

 

Euclid

Broadway Lion said:

 

4. Move to automatic operation with no crews on certain trains. My model railroad (a subway layout) operates across twelve scale miles of track and I keep ten trains running at once.

 

Does anybody believe this is impossible?  Certainly the technology for autonomous trains is proven to work.  For trains, it is fundamentally easier than for road vehicles.   The railroad application is half way there with its mechanically self-guiding track.  Whereas cars and trucks need to assimilate a wide field of ever-changing features, and make constant decisions about whether it is safe to proceed. 

 

This challenge is non-existent with trains.  In my opinion, full autonomous driving of cars and trucks is overpromised for marketing purposes. Ultimately I expect it will begin with only certain roadways which are sufficiently upgraded to play a part in the autonomous concept.

 

For railroads, the main obstacle is to fit all of the complex train handling moves into the autonomous program.  I think that is why Lion confines his prediction to applying automatic running to “certain trains.”  In other words you start with certain trains, and move forward with further development. 

 

People always shoot down the viable comparison of U.S. railroading to Rio Tinto by saying the two are not comparable.  Rio Tinto proves that the lack of comparability is not train size.  Instead the lack of comparability is move complexity, which is relatively simple with Rio Tinto.  With autonomous running of U.S. railroads, the changes in train operation will be larger than the change to running trains autonomously. 

 

 

 

 

As the railroaders on this forum can attest, handling a long, heavy train is a tremendous challenge for both humans and computers. There's a reason why aircraft have been automated for decades while trains are still a work in progress. Gathering the enormous amount of data necessary for properly handling a long train and feeding it into the computer is a much bigger challenge than the limited number of parameters required to handle an aircraft.

I'm guessing that if railroads go to autonomous operation at some point, the trains are going to get much shorter so it's easier for computers to operate them -- especially since automation would remove one of the main imperative for these grotesquely long trains: labor costs.

 

I totally agree that autonomous operations will end the monster trains.  The point of monster trains is to reduce labor cost by moving more cars per crew.  But that point goes away with autonomous operation because there is no crew cost with crewless trains.  So there is no point in coupling shorter trains together just to handle more tonnage with one crew.  It may prove that the shorter, lighter trains cost less per ton to operate than the monsters, once you take crew costs out of the equation.  I think it may usher in a new era of lighter, faster, more nimble, and more frequent trains.  Ironically, this will redefine PSR to mean Precision, Scheduled, Railroading.   

[rrnut282]

Automated trains will not be able to deal with a broken knuckle or emergency applications of the brakes (kicker).  There will still be a need for someone at the very least, very near every train.  Automation will not fix terminal issues of putting trains together and getting cars on the right trains.  And we all know railroads will cut this area, too, because savings are savings, customers be damned.  

[charlie hebdo]

Psychot

I'm guessing that if railroads go to autonomous operation at some point, the trains are going to get much shorter so it's easier for computers to operate them -- especially since automation would remove one of the main imperative for these grotesquely long trains: labor costs.

At least in major meteo areas, it would make rails better citizens 

[Euclid]

zugmann

 

 

Euclid

I conclude that the railroad management is exceptionally motivated to adopt autonomous train running just for the labor it would eliminate. 

 

 

We're not going to need hundreds of managers to oversee computers running trains.  So I don't think they should be that motivated. 

Wasn't that a major issue (for some) with permanent work-from-home?  Countless middle managers were pretty much as useful as certain pieces of anatomy on a bull?

 

I only use the term management or managers to distinguih management from labor. In this case, I mean the management that would advocate Autonomous running and have a say in the decision to implement it.  It would be the same people who are in favor of smaller crews.     

[zugmann]

Euclid

I conclude that the railroad management is exceptionally motivated to adopt autonomous train running just for the labor it would eliminate. 

We're not going to need hundreds of managers to oversee computers running trains.  So I don't think they should be that motivated. 

Wasn't that a major issue (for some) with permanent work-from-home?  Countless middle managers were pretty much as useful as certain pieces of anatomy on a bull?

[Psychot]

Euclid

Broadway Lion said:

 

4. Move to automatic operation with no crews on certain trains. My model railroad (a subway layout) operates across twelve scale miles of track and I keep ten trains running at once.

 

Does anybody believe this is impossible?  Certainly the technology for autonomous trains is proven to work.  For trains, it is fundamentally easier than for road vehicles.   The railroad application is half way there with its mechanically self-guiding track.  Whereas cars and trucks need to assimilate a wide field of ever-changing features, and make constant decisions about whether it is safe to proceed. 

 

This challenge is non-existent with trains.  In my opinion, full autonomous driving of cars and trucks is overpromised for marketing purposes. Ultimately I expect it will begin with only certain roadways which are sufficiently upgraded to play a part in the autonomous concept.

 

For railroads, the main obstacle is to fit all of the complex train handling moves into the autonomous program.  I think that is why Lion confines his prediction to applying automatic running to “certain trains.”  In other words you start with certain trains, and move forward with further development. 

 

People always shoot down the viable comparison of U.S. railroading to Rio Tinto by saying the two are not comparable.  Rio Tinto proves that the lack of comparability is not train size.  Instead the lack of comparability is move complexity, which is relatively simple with Rio Tinto.  With autonomous running of U.S. railroads, the changes in train operation will be larger than the change to running trains autonomously. 

 

As the railroaders on this forum can attest, handling a long, heavy train is a tremendous challenge for both humans and computers. There's a reason why aircraft have been automated for decades while trains are still a work in progress. Gathering the enormous amount of data necessary for properly handling a long train and feeding it into the computer is a much bigger challenge than the limited number of parameters required to handle an aircraft.

I'm guessing that if railroads go to autonomous operation at some point, the trains are going to get much shorter so it's easier for computers to operate them -- especially since automation would remove one of the main imperative for these grotesquely long trains: labor costs.

[Euclid]

I conclude that the railroad management is exceptionally motivated to adopt autonomous train running just for the labor it would eliminate.  To them, it would almost be like winning a war.  In addition, the manufactures of autonomous running equipment are exceptionally motivated by the lucrative new market that will suddenly open up with such a large, deep pockets, customer base as the railroad industry.  So the manufacturers will subsidize the implementation. These two factors combine to form a powerful current that will make a way for itself.

 

The biggest challenge is for the railroads to reorganize their operations for autonomous train running. 

[tree68]

zugmann

Sure if everything is near perfect.

And there lies the rub, as they say...

[zugmann]

tree68

Driverless trains are certainly within the realm of possibility technologically.  

Sure if everything is near perfect.

https://www.railwayage.com/safety/fra-issues-unintended-train-brake-release-safety-advisory/?fbclid=IwAR3_jn3ZTpC66Ds8rOO-CLz3fJRnTxfbrbQs_dhdz6a8Dj_jx06tzDWtCYg

 

From the above:

"FRA reported that its investigation of the rolling train’s event recorder, positive train control (PTC) system and engine data logs revealed “the PTC system had operated properly and would have initiated an emergency brake application upon reaching the signal; the Trip Optimizer was off; and the lead locomotive and consist did not cause the unintended brake release.”

 

If nobody was on board, we'd hope the emergency application would hold - a gamble since the brakes already aren't acting in a perfect way.  That's a bet I wouldn't lay money down on.  I've never seen a PTC computer hop off and tie on some brakes, but what do I know?  I don't have a train layout.