Regarding the claims made by NYAB in ad

Trains Magazine General Discussion
1

Preface: I am in no way trying to bash one of Trains advertisers; I am mostly seeking information and venting some irritation.

Since I’ve been retired for a while, I’m not too hip on the latest railraod tech. Which is why I ask my fellow railroaders for their input regarding the claims made by New York Air Brake in the ad on the back cover of the July 2017 Trains magazine.

Specifically, the claim that it “has never ‘broken a train’ over 1,200,000 trips and 190,000,000 miles”. Are these real, on the road miles or are they ‘simulated’ software miles. Also, if they are indeed real miles, what types of trains are they quoting? Mixed manifest mega-tonnage or 100-car trains of empties? On flat terrain or in ‘hogback’ territory?

While I’m certainly no Luddite, I find it hard to believe any software could be written that could handle the many variations of tonnage, load/empty distribution, track profile (grade ups and downs, curves), wet rail, snowy rail, leaves on rail, cold temperatures, hot temperatures, malfunctioning locomotives (such as no sand, wheel-slip issues, load dropouts, etc), trainline ‘kickers’, air hose parting on crossings or from debris placed between the rails, train striking a vehicle or person, signal malfunctions (such as dropping from clear to red just as the train approaches); just to name a few.

Since each of the items mentioned above has various effects on a train, the combination of possible multiple events would seem to lead to such a large number of permutations of possible hazards that I cannot imagine any software being able to handle that level of computation in the time frame necessary to act.

I understand that to what I am referring is an advertisement (with all the veracity that goes with it), and maybe the NYAB really does have such a superb product.

But as a former engineer, and in defense of my fellow rails, it rather irritates me that the language in the a

2

We already, to some extent, have trains operating in the US that operate themselves. I know UP, NS and CP are using them. They are used for “fuel management”, but I don’t think they’ve been saving as much fuel as the salesman said it would. They currently are a clear block systems. That is, they don’t see restrictive signals, they assume you are the only train out there. They do recognize permanent and temporary speed restrictions and work zones, provided they are in the system when intially logging in. Unforseen restrictions, radioed to the train, have to be acted upon manually.

UP has both NYAB’s LEADER and GE’s Trip Optimizer. I believe NS has LEADER and CP Trip Optimizer. (Not all routes of the railroads listed may be equipped.) IMO, both are scary, for different reasons.

LEADER has two versions. One only prompts the engineer to make throttle/dynamic settings. (Both systems prompt the engineer if using air brakes is needed.) If you are DP, it assumes the DP is in sync mode. The newer version, and slightly better, is an auto-throttle/dynamic operation. It actually runs the train. In auto throttle mode, it operates any DP separately from the lead. I feel it’s scary because it’s train handling isn’t good. Especially the first version. It’s caused people, myself included, to tear trains apart when following the prompts. About the only good thing about LEADER I can think of is the display. It’s better than the Trip Optimizer display and will be close to what the PTC displays will be.

Trip Optimizer has been an auto-throttle system from the beginning. We just got it on our part of the system a few months ago. It’s scary because it’s pretty good when it’s working. My MOP said the algorithims for our PTC will be the same used by the Trip O

3

To what Jeff, said; SCARY, might be an underatement !

Just an observation: Train Handling is an ART… Irequires a level of skill, and a sense of understanding what is going on in each train. Things that cannot be learned in a classroom, they come from experience. IMHO.

As nothing more than an interested observer, who lives beside a rail line, that line has a slight up grade, we have a number of trains,in eiither direction, passing each day. In a short time, it becomes loudly, obvious, when there is a rookie in the cab, as well as the experienced one who can handle a train smoothly…The differences between "Spike Jones and his Band’ and a Lawrence Welk Band, playing similar pieces.

4

I suspect much the same was said about autopilots on aircraft.

5

It’s still a much - talked about topic. When autopilot systems fail- it goes to manual mode. And when the person that is supposed to be the safety manual backup has lousy skills because they barely ever operate, well…

6

True. But in some crashes, it would have been better to let the autopilot continue even though it seemed counter-intuitive. Human interference made a bad condition worse.

7

Developing an autopilot is a much simpler problem than developing control software for a long freight train. An airplane can be treated (mostly) as a rigid body with relatively simple flight dynamiics. As long as the plane is under maneuvering speed, flight control inputs are unlikely to break the airframe. The impetus for autopilots came from reducing the workload on the pilots, or in the case of high performance aircraft were stability augmentation systems that responded much faster than any human could.

A long freight train is a whole different beast as it isn’t a rigid body with respect to curves (horizontal AND vertical) and slack action. With regards to the latter, it is very easy to break the train with improper control inputs, either through slack action or stringlining (tank car of metam sodium?). In addition, the varations in train makeup will make for very significant changes in train dynamics.

That said, even very recent autopilots have had problems dealing with unexpected events such as icing of the pitot tubes. (omitting long rant on control yokes versus side sticks on transport aircraft)

8

You are the engineering guy, but surely weather, velocity, acceleration, and horizontal and vertical changes are complicating factors not seen on trains. In the history of science and technology, generally certain developments happened earlier because they are simpler/easier.

9

I believe that if Spike Jones were to play the same piece that Lawrence Welk played, he would apply his talent to the performance. He was unique; who else could describe a horse race as he did–“Cabbage is ahead, Girdle is in the stretch, Tomato is trying to catchup–and it’s Hankerchief by a nose!” He was not inept.

A better comparison would be that of Florence Foster Jenkins and Beverly Sills singing opera. Mrs. Jenkins could not sing, and she had to hire an opera house for her performances.

Edited to correct spelling mistakes.

10

You forgot Mother-in-law nagging at the rail.

11

The main issues with weather and autopilots are extreme turbulence (e.g. thunderstorms) and icing. Both are usually dealt with by avoiding them, with the autopilot allowing the pillot(s) to concentrate on naviagting around those conditions. Having said that, icing is the more serious of the weather related issues for autopilots as many aircraft are certified to fly into known icing conditions and that icing WILL change flight dynamics in various unpleasant ways.

Freight trains are also affected by weather, with rain, snow and ice affecting adhesion along with wind affecting train resistance in much more complicated ways than winds aloft affect flying.

Keep in mind that an autopilot is pretty simple device, a couple of gyros coupled to the flight control surfaces to keep the wings and nose level. There’s a bit of control theory behind this, with the impetus coming from problems of the interation between the constant speed props and turbochargers on the B-17. A flight management system is a bit more complicated, controlling the throttles, iniating turns, climbs and descents.

To do the train right, every car in the train needs to be modeled.

12

WARNING: OFF TOPIC RESPONSE!!!

Actually the song you are referring to; Spike Jones manic take on the classic “William Tell Overture”, featured narration and lyrics written by City Slickers associate “Professor” Doodles Weaver

https://en.wikipedia.org/wiki/Doodles_Weaver

He was a regular contributor to Jone’s radio show and interestingly, the uncle of well known actress Sigourney Weaver…

https://youtu.be/aupQysNpMzk

13

Thanks,

Thanks. I completely forgot about her. She did not even show.

14

At least it was a photo finish… or an oil painting.

15

[quote user=“schlimm”]

erikem

schlimm

I suspect much the same was said about autopilots on aircraft.

Developing an autopilot is a much simpler problem than developing control software for a long freight train. An airplane can be treated (mostly) as a rigid body with relatively simple flight dynamiics. As long as the plane is under maneuvering speed, flight control inputs are unlikely to break the airframe. The impetus for autopilots came from reducing the workload on the pilots, or in the case of high performance aircraft were stability augmentation systems that responded much faster than any human could.

A long freight train is a whole different beast as it isn’t a rigid body with respect to curves (horizontal AND vertical) and slack action. With regards to the latter, it is very easy to break the train with improper control inputs, either through slack action or stringlining (tank car of metam sodium?). In addition, the varations in train makeup will make for very significant changes in train dynamics.

That said, even very recent autopilots have had problems dealing with unexpected events such as icing of the pitot tubes. (omitting long rant on control yokes versus side sticks on transport aircraft)

You are the engineering guy, but surely weather, velocity, acceleration, and horizontal and vertical changes are complicating factors not seen on trains. In the history of science and technology, generally certain

16

Murhpeys Law- There is no such thing as fail safe

17

In addition to uniform brake valves and draft gears, I would add uniform car lenght (yes, I know we have that in unit trains–I’m referring to mixed manifests), uniform train blocking, uniform power, uniform brake shoe height, …

To Schlimm:
You said that “surely weather, velocity, acceleration, and horizontal and vertical changes are complicating factors not seen on trains”. While I agree, in principal, I submit that an airplane (pilot) does not have the myriad things on the ground (as mentioned in my OP) that can change so rapidly.

I’m not trying to be sarcastic, but I am being rhetorical; have you ever actually been on a main-line freight train? Have you ever felt the incredibly immense instant acceleration that results from the run-in of slack on a very heavy train? It is a feeling like no other I’ve ever experienced. Using an aviation analogy, I would imagine it is somewhat like launching off of an aircraft carrier (except the speeds are a bit lower). One minute you’re drifting along in the 4th notch, letting the friction of rolling resistance to slow your train for the planned upcoming slow order, when all of the sudden, WHAAM!, the rear 40 loads of lumber have suddenly tried to reach their destination ahead of the locomotives, and your speed has jumped 20mph. Now you are going too

18

The effect of slack is that unless the train is either stretched or compressed - and stays that way for the relevant time frame - the train braking performance isn’t ‘linear’. That means it can be ‘discontinuous’, which means not always and wholly predictable.

Even then, there’s an element of ‘dashpot’ in the cars because the draft gear allows some motion with varying degrees of resistance. Compare to the slack between the coupler knuckles, which is pretty ‘loose’ as that space opens and closes. But it’s better now that the Hydra-Cushion cars aren’t so prevalent . . . [:-^]

As to engineer proficiency: One of Malcolm Gladwell’s books had a chapter on how the best correlation with aircraft crashes was the culture of the crew, not weather or anything like that. As I recall, the US pilots are the safest because they’re not afraid to challenge something that seems not right to them, whether it’s a wrong decision by the command pilot or a bad situation. Jeff’s recounting of how he observes and overrides the system seems to be of like kind.

  • PDN.
19

Electronic braking that has been discussed on here before is supposed to reduce slack problems considerably. I think an engineer’s job is hard but I also think much of it can and will be automated within the next 10-15 years, at least on mainlines.

20

Yup.

And when we see a train, will we be waving at C3PO at the controls?

And will he wave back?