General Discussion

Under the new PTC law, do all lines have to have PTC installed on them, or just those with passenger trains? I should know this, but I have missed it.

George

Just to clarify, TIH=Toxic Inhalation Hazard, yes?

 The law required:

1. All main tracks carrying intercity or commuter passenger, or toxic inhalation gases (TIH), excepting:
   
2. Lines with fewer than 5 million gross tons (MGT) per year;
3. Lines operated by Class II or Class III railways;
   
4. Trains operated by Class II or Class III railways on Class I railways if they are operating less than 20 miles on the Class I, but,
   
5. Other lines as designated by the Secretary of Transportation shall be equipped.

In practical terms, that is something like 80 percent of the Class 1 route miles in the U.S.

There is a considerable discussion ongoing as to the exclusions and inclusions, most of which has nothing to do with safety and everything to do with who is paying and how much it costs.

RWM

Thanks to all for the replies.

George

Railway Man:

  [snip; emphasis added - PDN]  There is a considerable discussion ongoing as to the exclusions and inclusions, most of which has nothing to do with safety and everything to do with who is paying and how much it costs.

RWM

As I used to advise my municipal clients to say to representatives from the state who came calling with news of various new programs, requirements/ mandates, and funding: 

''We'll take what you say seriously, in direct proportion to how much money you bring along to help us implement it.''

- Paul North.

Suppose I am a Class I road and I decide it is feasible for me simply to reroute any TIH cargo away from a given line.  However, Amtrak is a user of that line.  Can I tell Amtrak to pay for the PTC or forget using my line?  As RWM said, I'm sure there are some interesting discussions going on. 

Dakguy201:

Suppose I am a Class I road and I decide it is feasible for me simply to reroute any TIH cargo away from a given line.  However, Amtrak is a user of that line.  Can I tell Amtrak to pay for the PTC or forget using my line?  As RWM said, I'm sure there are some interesting discussions going on. 

You could.  But will the FRA agree to both halves of the decision?  That is the question.

RWM

Such a reroute and resulting choices may not be merely hypothetical, if you think about it a little further.

If we accept that Amtrak should or does want to go where the population centers are - in general those are the exact same places that we'd want to keep TIH away from as well, for the same reason - in a  rational world, that is.    So those reroutes should be encouraged, not thwarted, to minimize the probable effects from any TIH release that might occur, from any cause.

In my view, if only one could be paid for, a sensible decision would be to put the PTC on the passenger line, not on the TIH line.  But then the law would lead to / mandate that the TIH stay on the now-PTC protected line.  So what we have here is a pair of undesirable situations - either PTC and both Amtrak and TIH on the same line, with the railroad paying most or all of that line's PTC cost; or on separate lines, with the railroad certainly paying the cost for one PTC installation for the TIH line, and possibly some or all of the cost of the PTC on the other Amtrak line.  Although a sensible compromise solution would be to allow the railroad to separate the TIH at the price  of / only if it pays for the cost of putting PTC on the Amtrak line, then the railroad doesn't seem to get much benefit from that - it has to pay for at least part of a PTC installation that it can't use.  So this seems to be something of a 'Hobson's choice' for the railroad.

- Paul North.

Perh

Paul:  Embedded in your first sentence is a premise, that cost-benefit comparison between passenger train accident prevention and TIH accident prevention favors passenger.  For that comparison I have seen no evidence.

The law does not proscribe an allocation formula between TIH and passenger for the installation of PTC.  Some railroads have proposed an allocation formula on a per-train basis, e.g., if a line has 1 train per day carrying TIH and 3 passenger trains, the allocation formula is 1 part to the host railroad, 3 parts to the passenger carrier. The law doesn't enable the host railroad to charge back to Amtrak 1971 agreement trains their share of the PTC, but some post-1971 Amtrak agreements, as well as most commuter line agreements, do have language that enables the host railroad to charge back an allocated PTC cost or even the full cost of the PTC.

But in reality most of the intercity passenger lines have TIH on them that isn't going to go away quickly, so the choice doesn't occur.

RWM

Well, since the thread that had the debate going about WW I rail ops vs. same in WW II is now locked, I'll have to look here instead for my intellectual exercise for the day:

Railway Man:

  Paul:  Embedded in your first sentence is a premise, that cost-benefit comparison between passenger train accident prevention and TIH accident prevention favors passenger.  For that comparison I have seen no evidence. 

 True enough.  That sentence is: ''In my view, if only one could be paid for, a sensible decision would be to put the PTC on the passenger line, not on the TIH line. [emphasis added - PDN]  I claim or rely on no formal studies or other 'evidence', other than my own personal seat-of-the-pants 'value judgment'.  That in turn is based on my recollection that over the past 40 years or so, it seems that far more lives have been lost in collisions involving passenger trains and other trains, than in TIH incidents that involve trains - although to be fair, I think grade crossing collisions have claimed more lives than either of those.  But other than that historical experience - nahh, I got nothing.  Now, I wonder what the cumulation of FRA / NTSB statistics would show on this ?

Railway Man:

  The [PTC] law does not proscribe [prohibit] an allocation formula between TIH and passenger for the installation of PTC.  Some railroads have proposed an allocation formula on a per-train basis, e.g., if a line has 1 train per day carrying TIH and 3 passenger trains, the allocation formula is 1 part to the host railroad, 3 parts to the passenger carrier. The law doesn't enable the host railroad to charge back to Amtrak 1971 agreement trains their share of the PTC, but some post-1971 Amtrak agreements, as well as most commuter line agreements, do have language that enables the host railroad to charge back an allocated PTC cost or even the full cost of the PTC. [edits and emphasis added - PDN] 

Interesting formula - makes sense, since 'train count' or frequency is probabaly the most important single driver or parameter that actually causes the need for PTC for collision avoidance - less trains = less potential conflicts.  And while speed control, signal compliance, and 'track condition' functions are largely on the basis of individual trains - there too, the operator who runs the most trains gets the most benefit. 

Yes - merely for simplicity, I was addressing/ assuming that the 1971 Amtrak law's prohibition on additional charges was generally applicable - though it is not, as you have pointed out. 

Railway Man:

  But in reality most of the intercity passenger lines have TIH on them that isn't going to go away quickly, so the choice doesn't occur.

RWM

I'll defer to your observation and experience on this point.  The need or motivation to reroute the TIH may vary with the likely magnitude of the PTC expenditure needed to keep it where it is . . .

- Paul North.

Paul:  The problem with cost-benefit analysis for PTC is that standard statistical analysis tools do not provide sufficient precision.  The rate per train-mile of human casualties and property damage caused by authority excursions, overspeeds, or open switches is too small to find a causal relationship that can statistically be distinguished from random chance.  Thus, which types of loss-preventions one would want to preferentially target the PTC system toward, that cannot be predicted with any accepted degree of certainty.

RWM

RWM -

OK, thanks for that thoughtful and in-depth - very deep ! - explanation. 

At the risk of possibly over-simplifying, it seems ironic to me that the rail system is so safe - i.e., ''the rate per train mile . . . is too small to find a causal relationship'' - that we can't now use the casualty/ damage data to identify what's broken or causing the losses - as opposed to those caused by random chance -  and hence we can't figure out what 'fix' might work on those problems.  And, that the problems are so 'diluted' that a specific 'fix' seemingly likely won't work even on an identifiable problem with an acceptable degree of certainty or reliability.

And just to clarify (not argue with) your point:  It seems you're saying that - for instance, 'authority excursions' [= running past Track Warrant limits or an absolute red = Stop signal] - are so rare, that the statistical science can't predict what percentage of those incidents would be prevented by implementation of PTC ?  Please forgive my naivete here, but isn't PTC designed to prevent all of those incidents ?  Is there a credible scenario where that could occur which PTC - operating as intended - wouldn't be effective to stop the train before that occurs ? 

Or, is this a database collection and characterization problem - one where the collected historical data on the causes of past collisions hasn't been 'sliced and diced' in quite the way that is now needed to identify the 'root cause' kind of thing, so it can't be said how well PTC will work on an undifferentiated broad spectrum of loss causes ?

Finally, from previous threads on PTC I understand that you abhor the government = FRA's insistence on near-perfection of that or any other new/ replacement/ supplemental safety system - truly, it seems that ''The perfect is the enemy of the good'' here.  But I'm not picking up that's the difficulty with this particular analytical problem - correct ?

Interesting dilemma, anyway.  Thanks for sharing you insights on it.

- Paul North.

Paul_D_North_Jr:

And just to clarify (not argue with) your point:  It seems you're saying that - for instance, 'authority excursions' [= running past Track Warrant limits or an absolute red = Stop signal] - are so rare, that the statistical science can't predict what percentage of those incidents would be prevented by implementation of PTC ?  Please forgive my naivete here, but isn't PTC designed to prevent all of those incidents ?  Is there a credible scenario where that could occur which PTC - operating as intended - wouldn't be effective to stop the train before that occurs ? 

"Designed To" and "Will" are two different things.

As to a "Credible Scenario", if it can happen it will happen.  There's no way to design a perfecxt system.   

Paul - As I read RWM's post I kind of draw the conclusion that it's difficult to come up with a cost/benefit analysis. 

IMHO, RWM posits that we don't even know how many incidents PTC will prevent.  The big ones, no problem.  It's the little stuff that didn't result in any loss (ie, running past authority limits, speeding) that is hard to quantify since it probably rarely gets reported to anybody who will make a big deal of it.  Granted, somebody may get a brief vacation, but the Daily Blowhard won't be covering it.

This brings up the very real possibility that PTC will result in a flood, nay - a deluge, of "incidents" as we become aware of those otherwise insignificant events.