Wizlish Wizlish It's a bit like emergency brakes that make trains 'stop short' at crossings. There is technology that could do this, and we've discussed it in a couple of fairly long threads. I don't recall anything like that ever being discussed here. How would this be accomplished? Under what terms would "stopping short" occur? Dave Klepper, I think, started a thread on electromagnetic track brakes, in which I think you participated. Interestingly enough, erikem did an engineer's analysis on the requirements to actually make such a thing work at full scale and concluded it could be made to function. The basic idea (correct me anyone if this is wrong) is to provide an electromagnet of appropriate field strength and characteristics that 'rides' along the railhead, with enough cross-sectional area to produce both 'clamping' friction and eddy current induction in the rail steel. When a high current is applied through this magnet it produces a strong retarding force, independent of any braking being applied to the wheels, and under some circumstances it can exert a restoring force to keep a carbody (or truck frame, perhaps) in line with the rails. There is a limit as to energy dissipation (in part set, I think, by the Curie point of the railhead) and there are some potential problems with rail lifting or activation on curves or crossovers, and of course with dramatic problems in train handling if there is any differential slack or disproportionate braking level in different parts of the train. A version of this is and was applied historically to streetcars, where it has I believe been demonstrated to work quite well, at the (unexpected to me) cost of increasing rail corrugation when used more than intermittently. Apparently modern Li-ion battery tech has gotten to the point that it can store enough energy to brake a given loaded car down to a reasonable speed in a respectably short time ... once. It may not matter if parts of the system, or indeed parts of the track, require even expensive repairs when the emergency track brake is 'fired'. The more important consideration -- which as I mentioned would be difficult if not impossible to substantiate -- is whether the system introduces more danger or risk than it relieves.
Wizlish It's a bit like emergency brakes that make trains 'stop short' at crossings. There is technology that could do this, and we've discussed it in a couple of fairly long threads. I don't recall anything like that ever being discussed here. How would this be accomplished? Under what terms would "stopping short" occur?
Wizlish It's a bit like emergency brakes that make trains 'stop short' at crossings. There is technology that could do this, and we've discussed it in a couple of fairly long threads.
I don't recall anything like that ever being discussed here. How would this be accomplished? Under what terms would "stopping short" occur?
Dave Klepper, I think, started a thread on electromagnetic track brakes, in which I think you participated. Interestingly enough, erikem did an engineer's analysis on the requirements to actually make such a thing work at full scale and concluded it could be made to function.
The basic idea (correct me anyone if this is wrong) is to provide an electromagnet of appropriate field strength and characteristics that 'rides' along the railhead, with enough cross-sectional area to produce both 'clamping' friction and eddy current induction in the rail steel. When a high current is applied through this magnet it produces a strong retarding force, independent of any braking being applied to the wheels, and under some circumstances it can exert a restoring force to keep a carbody (or truck frame, perhaps) in line with the rails. There is a limit as to energy dissipation (in part set, I think, by the Curie point of the railhead) and there are some potential problems with rail lifting or activation on curves or crossovers, and of course with dramatic problems in train handling if there is any differential slack or disproportionate braking level in different parts of the train.
A version of this is and was applied historically to streetcars, where it has I believe been demonstrated to work quite well, at the (unexpected to me) cost of increasing rail corrugation when used more than intermittently.
Apparently modern Li-ion battery tech has gotten to the point that it can store enough energy to brake a given loaded car down to a reasonable speed in a respectably short time ... once. It may not matter if parts of the system, or indeed parts of the track, require even expensive repairs when the emergency track brake is 'fired'. The more important consideration -- which as I mentioned would be difficult if not impossible to substantiate -- is whether the system introduces more danger or risk than it relieves.
I've just read the whole thread, and I just can't keep from commenting on Wizlish's post, where he has described the system quite well. Probably this has already been said somewhere else, but here is the state of things in Germany.
Trams have been legally required since 1950 to use track brakes to achieve the mandated deceleration (about 9 feet per second squared). On heavy rail, every vehicle except locomotives which is certificated for speeds above 140 km/h (87 mph) has to have such a brake; brakes on individual vehicles are powered by batteries. This applies, of course, mainly to passenger equipment, as very few freight vehicles operate at speeds above 120 km/h or 75 mph.
(I hope the picture landed in the post, have no experience doing this).
This kind of brake is, AFAIK, used in this fashion everywhere in the UIC region (Europe without former USSR). DB ICE3 EMUs have a contactless brake system using only eddy currents for braking.
Electrically-signalled pneumatic brakes are in widespread use on passenger equipment; they can be controlled either by electric signals transmitted by cable or in the traditional way by pressure in the air line. In the UIC version of this system, the locomotive provides electrical power to the train via the UIC-standard cable link and is therefore suitable for freight equipment. the cable link is constantly monitored for malfunctions or train separation. The UIC region once planned to convert all rail vehicles to automatic couplers (the AK69E); tests had shown that conventional pneumatic brakes in conjunction with these couplers led to inacceptable slack action up to derailments, so it was planned to use EP brakes throughout. The conversion to automatic couplers was planned to happen on an Easter weekend (I think in 1972) all over Europe at the same time, but a number of countries balked - understandably. As the new coupler was completely incompatible with the traditional screw-link coupling every vehicle would have to have been converted at the same time, and the total number of vehicles concerned was certainly at least comparable to the number of rolling stock in the US, with a considerably larger number of locomotives. I seriously doubt the changeover would have been possible in much less than a month, with a resulting disruption in rail traffic. So the automatic coupler didn't happen, except in special applications (mainly ore unit trains above 4000 tonnes) and EP brakes in freight applications died with it.
EuclidOne factor in particular to consider for reason to modify the brake force on the leading cars is the train speed. If the train deceleration happens to be exceeding the initial expectation of the system, then the application on the leading cars might be increased; and vice versa.
I would be a little more careful about 'which cars' get the increased braking effort, both in terms of reduction of train speed in different ranges (especially high speed, where you will want to modulate 'loads' differently from empties even if you haven't telemetered the load sensing, e.g. with higher effort on the loads initially shifting to a 'bias' toward the rear end of the consist to keep the slack from run-in or oscillation as the speed comes down.
The situation with unexpected deceleration also has a couple of aspects. This does not only involve graduated reduction of the front end to 'keep out of the way', it might also call for quick reduction if derailed cars start acting as an anchor and you want to preserve your drawbar integrity through the 'forward section' to avoid more potential run-ins.
I have already addressed some of the 'reasoning' for why you would want brake modulation relative to deceleration on the rear part of the consist, and at least one way that could be achieved without compromising more than about 3% of the ultimate "ideal-case" stop time or distance.
Euclid However, if the electric train line parts after the two applications are triggered, the application behind the derailment cannot be further modified. Off hand, I cannot see any reason why the trailing application should ever be less than maximum force; or would need any further modification during the stopping process.
You're with buslist and some other very smart people in thinking this. I still think that it would be better to modulate the application, even if the 'ultimate result' is to have the brakes applied with full 'emergency' force and perhaps sliding, and I also think that this needs to be done with some sort of wireless enablement.
My current thought is to put simple metadata into the DPU protocol that could be 'read' by the (still connected) processors in a following segment. Presumably it is not a 'difficult' exercise to provide tail-end 220V emergency power from a device on the 'disconnected' part of a consist, whether or not actual DPU power is there, so the issue becomes how the rear brakes are modulated IN SPECIFIC EMERGENCY CONDITIONS -- this would probably involve a special reception mode and perhaps an 'emergency button' (tied in with PTC somewhere) that would actuate the DPU overlay recognition or priority if automatic systems did not do the job right.
This might only be good for 'a couple' of stops to rest, but it would assure as well as much more complicated systems that the rear end of the train progresses to maximum achievable safe deceleration and then a stop in minimum-time or minimum-distance in the 'best' way that jurisprudence, insurers, lawyers and newsworkers would recognize...
This would also coordinate overall as well as incremental braking rate relative to some 'conditions' ... most notably, if the rear of the train turns out to be stopping "less slowly" than expected, or if it were to start to move again after stopping for some reason, there could be an indication (both automatically and to the crew) that they should do a graduated release even before (as would apply in any other circumstance I could conceive) the front-end crew has not yet been able either to walk the consist or assess the 'state' of the cars at the point of separation.
I'm taking up the point in your last paragraph in a following post.
My short answer is that it makes no difference, as long as both reservoirs are fully charged. Both types of application can happen instantaneously. Both will produce the same braking force. Both can be initiated by the intact wire when the affected derailment detector sends the signal to the ECP controller.
(They appear to be tinkering with the automatic 'quote' function, so I used BBcode directly)
As a 'nit', the ECP "emergency" application with an intact trainline will set up on the part of the train away from any command control valves (engineer's brake valve, properly-equipped FRED or DPU, an 'air repeater' midtrain car or Dave Klepper's midtrain valves, etc.) a little quicker, and that is the source of most of the observed "3%" difference. The 'rest' would come in if the ECP system is 'reliable' enough that a faster application rate in emergency without overshooting pressure on cars that can't handle 'excess' developed pressure can be achieved.
I'm uncertain whether the latter effect is safe to assume if the control cable (or other modality) parts or is impaired; I also presume that an ECP system that sets a 'penalty brake' emergency when it detects something like LOS or a wrong checksum or whatever would default to some 'maximum safe rate' and not to what the system could achieve under full sensing and feedback authority.
If the trainline parts, and the ECP is a 'hybrid' system, I expect the actual setup may be slower for ECP than one-pipe, because the trainline pressure will have to fall further before the physical triple valve ports over and the brake cylinders start to fill. If, as I think highly likely, the control cable 'parts' at roughly the same time the air line does, any 'default' modulation of the individual valves would depend on the way the particular ECP system handles catastrophic power-off (as distinct from catastrophic LOS of analog, digital, or mixed control signals). But this is an extreme case!
In my opinion, 'differential braking' in this situation would control the ECP on the 'front' end SPECIFICALLY to a service brake and not emergency application, not least because that preserves some chance of controlling the front end to keep out of the way of the back end piling up and sliding. if you have a dedicated ECP system (no use of a drop in the 'constant' supply pressure to actuate physical triple valves as a backup) then you have the authority to command graduated release on part or all of the 'intact' forward part of the consist, and even apply power to accelerate the train 'away' at graduated release ... but I'd think this will be much easier with a 'service' application than the digital equivalent of big-holing.
Most (perhaps all) the part of any system that controls braking with a 'derailment detector' is going to be concerned with trains that have not broken in two; in fact one of the important parts of having derailment detection at all is to prevent actual breaks from happening. So I think it makes some sense to have the 'right kind' of derailment detector ... one that can discriminate between different things happening with the truck position or perhaps geometry, and that understands when flange force or wheel accelerations are wrong, and that can communicate with the brake controller to 'figure things out' as much as possible, and as long a lead time as possible, before catastrophe actually starts. And then develop some protocol, probably involving metadata codes, that signal 'events' of progressive importance (as if they were guide numbers) to the ECP controller, so that the choice of differential vs. progressively 'fuller' service braking vs. emergency can be made with better assurance.
As I understand it, both 'commercialized' and marketed systems of American ECP (WABTEC and NYAB) use a procedure to establish physical location of the different cars at 'startup' (and, presumably, in realtime any time a car is switched out of a consist, etc.) It is comparatively simple to establish a metadata 'protocol' that tags a derailment-detector code with data associating it with a specific car, and in turn with that car's position, derived weight or maintenance state, etc. So nothing radical should be required either in the equipment or the programming to add the important part of this functionality ... in my current opinion.
(I should insert here that I still think redundant wireless backup, and some self-power and modulation for the ECP valves, is a valuable thing in one of these systems, but of course YMMV.)
I confess I have never quite understood what the controversy with 'crosstalk' between trains using twisted-pair conductors was supposed to be. My father thought that some 'early' applications might not have twisted the conductors "enough" (perhaps not at all?), thinking they were power and not mixed power/data. Certainly none of the Australian adopters would be that dumb.
Wizlish If you remember M636C's comments about how crosstalk was supposed to be causing 'emergency' events -- if the computer 'thinks' it has lost control authority ... or if it in fact does lose authority and cannot modulate valves ... the default is essentially indistinguishable from a 'big-hole' emergency application.
If you remember M636C's comments about how crosstalk was supposed to be causing 'emergency' events -- if the computer 'thinks' it has lost control authority ... or if it in fact does lose authority and cannot modulate valves ... the default is essentially indistinguishable from a 'big-hole' emergency application.
My comments about crosstalk were that I didn't believe it has ever occurred in real life. I have seen hundreds of ECP trains pass eachother at full track speed or dead slow and there was never any interaction of any kind.
I have seen an empty 80 car coal train stop following an emergency application and it did so so quickly and quietly that I didn't believe what I was seeing. There was no run in of slack. The train which was doing around 40 mph just stopped in less than its own length. The faulty signal was visible from a bridge I was standing on but not visible to the crew before they cleared the bridge. The train stopped with about 1/3 of the cars still in rear of the bridge.
I don't know the specific arrangement of air reservoirs on those cars, but it worked well.
M636C
Wizlish Remember that 'Emergency' is just as pneumatic on a current ECP setup as it is for Westinghouse for any of the events BaltACD mentions. If there are reserve power means on the cars, the emergency will follow quick opening of the valves; if there is no reserve, the valves are arranged to 'fail safe' by going to maximum (design) application as they de-energize. If you remember M636C's comments about how crosstalk was supposed to be causing 'emergency' events -- if the computer 'thinks' it has lost control authority ... or if it in fact does lose authority and cannot modulate valves ... the default is essentially indistinguishable from a 'big-hole' emergency application. I do not think there is a way to design a truly 'failsafe' EC system (without distributed power and intelligence!)) that would not have such a behavior to some very significant extent. Part of this of course is that, as I understand it, the current kind of ECP dual-mode conversion leaves the triple valve intact and provides a secondary modulating valve that 'bolts to it'. So any catastrophic drop of the (kept continuous) line pressure in an ECP one-pipe system will, when the pressure drops to where the 'existing' triple actuates, cause things to behave 'just' like a regular Westinghouse-equipped car going to emergency after a trainline break (well, maybe a little slower if the line starts at a higher actual gauge pressure...). In that case there 'might' be a little more or less actuating pressure depending on where the ECP valve was when it 'de-energized' and it was 'designed' just to stop in whatever position it was when the lights went out, or automatically closed to avoid 'false actuation conditions'.
Remember that 'Emergency' is just as pneumatic on a current ECP setup as it is for Westinghouse for any of the events BaltACD mentions. If there are reserve power means on the cars, the emergency will follow quick opening of the valves; if there is no reserve, the valves are arranged to 'fail safe' by going to maximum (design) application as they de-energize.
If you remember M636C's comments about how crosstalk was supposed to be causing 'emergency' events -- if the computer 'thinks' it has lost control authority ... or if it in fact does lose authority and cannot modulate valves ... the default is essentially indistinguishable from a 'big-hole' emergency application. I do not think there is a way to design a truly 'failsafe' EC system (without distributed power and intelligence!)) that would not have such a behavior to some very significant extent.
Part of this of course is that, as I understand it, the current kind of ECP dual-mode conversion leaves the triple valve intact and provides a secondary modulating valve that 'bolts to it'. So any catastrophic drop of the (kept continuous) line pressure in an ECP one-pipe system will, when the pressure drops to where the 'existing' triple actuates, cause things to behave 'just' like a regular Westinghouse-equipped car going to emergency after a trainline break (well, maybe a little slower if the line starts at a higher actual gauge pressure...). In that case there 'might' be a little more or less actuating pressure depending on where the ECP valve was when it 'de-energized' and it was 'designed' just to stop in whatever position it was when the lights went out, or automatically closed to avoid 'false actuation conditions'.
Unless there are batteries on the cars, charged from the train-line electric cable. Then, loiss of signal applies the brakes from the batteries. Probably just as expensive and requireing greater maintenance than the above, however.
Euclid tree68 Euclid Therefore, with ECP, a “Service” application can stop the train just as quickly as an “Emergency” application—if the “Service” reservoirs are fully charged. However that might not be the case, so the system includes “Emergency” reservoirs that always hold a full charge in reserve just for an “Emergency” application. Therefore, with ECP, the one main difference between a “Service” application and an “Emergency” application is which of the two reservoirs provides the air. Point of order: The current system also includes an emergency reservoir. It's not a feature unique to ECP. I said that the difference between ECP and conventional air brakes was how they operated their “Service” and “Emergency” applications. I elaborated that ECP includes a dedicated “Emergency” reservoir only to clarify that point about ECP. I did not say that the “Emergency” reservoir is not included with conventional air brakes. When I said (with ECP) the one main difference between a “Service” application and an “Emergency” application is which of the two reservoirs provides the air; I was referring at that stage to the difference between the two types of brake applications within ECP. The point of that difference between the two brake applications was opposed to the case with conventional air brakes in which there are other differences between the two types of brake applications besides which of the two reservoirs provide the air. The main difference I had in mind is that a “Service” application with conventional air brakes must have a limited rate of reduction in order to avoid triggering a change to an “Emergency” application. I assume that this is not the case with ECP because its electrical control dispenses with the need for the pneumatic control feature of conventional air brakes that produces an “Emergency” application on the basis of the rate of reduction triggering the “quick action” of the triple valves.
tree68 Euclid Therefore, with ECP, a “Service” application can stop the train just as quickly as an “Emergency” application—if the “Service” reservoirs are fully charged. However that might not be the case, so the system includes “Emergency” reservoirs that always hold a full charge in reserve just for an “Emergency” application. Therefore, with ECP, the one main difference between a “Service” application and an “Emergency” application is which of the two reservoirs provides the air. Point of order: The current system also includes an emergency reservoir. It's not a feature unique to ECP.
Euclid Therefore, with ECP, a “Service” application can stop the train just as quickly as an “Emergency” application—if the “Service” reservoirs are fully charged. However that might not be the case, so the system includes “Emergency” reservoirs that always hold a full charge in reserve just for an “Emergency” application. Therefore, with ECP, the one main difference between a “Service” application and an “Emergency” application is which of the two reservoirs provides the air.
Point of order: The current system also includes an emergency reservoir. It's not a feature unique to ECP.
There is still the need for pneumatic actuated emergcy applications - when the train becomes uncoupled for whatever the reasons (slip over, unlocked coupler operation, broken knuckle, drawhead failure etc. etc.) as the airline will be opened to atmosphere and the electric line uncoupled.
Never too old to have a happy childhood!
EuclidTherefore, with ECP, a “Service” application can stop the train just as quickly as an “Emergency” application—if the “Service” reservoirs are fully charged. However that might not be the case, so the system includes “Emergency” reservoirs that always hold a full charge in reserve just for an “Emergency” application. Therefore, with ECP, the one main difference between a “Service” application and an “Emergency” application is which of the two reservoirs provides the air.
Larry Resident Microferroequinologist (at least at my house) Everyone goes home; Safety begins with you My Opinion. Standard Disclaimers Apply. No Expiration Date Come ride the rails with me! There's one thing about humility - the moment you think you've got it, you've lost it...
EuclidWizlish, When you speak of modulating the brake application, what characteristics of that modulation are you considering? You cite the modulation in reference to your father’s system for Conrail for the brake response to PTC. I assume that the point of this modulation was to prevent the braking train from derailing. I further assume that this would be controlled by providing strain gages on draft gear to measure the buff and draft and then tell the ECP system what to do in order to balance the buff and draft forces draft in a train in order to prevent a derailment during a hard stop.
It was actually something quite different, and much more 'simple-minded' (as might be expected from '80s technology).
The 'problem' was that the effect of PTC, at that time, was similar to ATC in triggering a full 'penalty brake' application (if not in fact an emergency application) when a signal was passed -- much the same, I think, as the operation of the PZB 'intermittent' train control would have done. There are movies of test trains, I think on New Jersey Transit, that show the result.
The problem is, or was, that the testing was done on passenger consists, whereas the problem, for Conrail after the Chase accident, was that the trains that were needing to be 'stopped quickly' were particularly long and heavy freight trains, of 'interchange' quality as far as maintenance or any sort of special equipment like ride-height braking load adjustment would be concerned, operating at high speed.
That meant that all sorts of 'cure-worse-than-the-disease' knuckle snatching, derailing, and perhaps even stringlining might be observed if the 'wrong' consist were to be automatically big-holed by a nominal "safety" system.
What my father proposed was a safety system that worked in 'parallel' with the service brake valve. but was implemented in a handheld system that would be assigned to a particular engineer (the special part of the system on the locomotive being a proportionally-controlled 'rotair' valve, not a modification to the regular brakes). The idea was that a given consist's information (weight, load, car type, etc.) would be translated from the manifest into a 'computer model' of how the train would be handled in fast braking by a good engineer, and a 'penalty' application from the ATC/PTC system, whether from overspeed or signal violation, would result in the rotair valve 'simulating' how that engineer would work the regular brake handle over time to make a best-distance stop for that particular train.
Of course a much more 'sophisticated' version of this would have additional sensors in the equipment, or modules put in the train every so few cars that could read brakeline pressure, load between the knuckles and coupler faces, etc. But that was not the 'point' of the actual system, which was simply to make the automatic system imitate a human engineer stopping an ordinary train under then-ordinary conditions, instead of just pretending that dumping the air was the answer.
BaltACD One thing to remember about US unit coal trains - in many cases the cars are owned by private interests - not the trailroads - they are also maintained by private interests - not the railroads. If the private owners thought there was a ROI sufficient to cover the costs and add to the owners bottom line - they would be shouting and demanding ECP for the operation of their equipment. You don't own and maintain several hundred upto several thousand railcars and not think about the bottom line of your ownership.
One thing to remember about US unit coal trains - in many cases the cars are owned by private interests - not the trailroads - they are also maintained by private interests - not the railroads. If the private owners thought there was a ROI sufficient to cover the costs and add to the owners bottom line - they would be shouting and demanding ECP for the operation of their equipment. You don't own and maintain several hundred upto several thousand railcars and not think about the bottom line of your ownership.
I perhaps should explain some aspects of Australian Railways.
The tracks are owned by the state governments. Outside Queensland and Western Australia the main lines are leased to the Federal Government and operated and maintained by an operating authority the Australian Rail Track Corporation which acts as a commercial entity.
As a result, an operator can lose the traffic from a mine on a straight commercial basis, since the tracks accessing any given mine are open to all operators.
One mining company, Whitehaven, purchased their own train (3 x EMD 4300 HP locomotives and 80 coal hoppers) which is operated for them by Pacific National, presumably at a lower rate than PN charge for supplying a train.
Another company, Glencore, purchased nine trains and thirty 4350 HP GE locomotives which carries all their coal and that of some others, and is operated by Freightliner, a subsidiary of Genessee and Wyoming.
Both Glencore and Whiehaven purchased only ECP equipped stock, and Glecore bought theirs in three batches over years, so they must have thought the additional cost was worth the effort.
In Central Queensland, the BHP Billiton Mitsubishi Alliance (BMA) a coal miner, purchased thirteen electric locomotives and nine trains all fitted with ECP braking to be operated by Pacific National.
The operator who had all conventional fleets are converting to ECP (not just testing one train) so they must be convinced of the return on investment too.
While conditions are different in Australia, return is important and cometition is serious. Pacific National lost a mine contract to Glencore/Freightliner (who run only ECP trains) and as a result PN have started to put non ECP cars and locomotives in storage.
The return on investment is there, both to operators and private owners of locomotives and cars or they wouldn't keep buying ECP equipment.
I can't believe that conditions are so different in the USA that the same ECP gear on basically similar rolling stock won't give the same favourable return on investment.
tree68 M636C I think that if USA railroad managers had closely inspected the operating ECP trains in Australia last June they would have had to reconsider their position. I would opine that there are only two things that will cause US railroads to adopt ECP: 1. Government Regulation 2. Proving that ECP has an ROI sufficient to make it worth their while (ie, increasing the bottom line). At this point, #1 hasn't come into play for general railroading, and #2 hasn't shown it's face or ECP would be delayed only for lack of parts.
M636C I think that if USA railroad managers had closely inspected the operating ECP trains in Australia last June they would have had to reconsider their position.
I would opine that there are only two things that will cause US railroads to adopt ECP:
1. Government Regulation
2. Proving that ECP has an ROI sufficient to make it worth their while (ie, increasing the bottom line).
At this point, #1 hasn't come into play for general railroading, and #2 hasn't shown it's face or ECP would be delayed only for lack of parts.
Despite the recent statements by our political leaders, Australia is not noted for innovation.
However, in 2005, what is now Aurizon introduced their first unit coal trains in the Hunter Valley and used the first AC traction locomotives and the first trains with ECP brakes.
Clearly, with no significant numbers of locomotives and cars in that area (and in their case, on that track gauge) they were free to start as they wished and they had the first regular ECP trains and the first AC traction locomotives.
In the eleven years since, every operator in that area has wholly or partially converted to ECP, even the operators who had substantial quantities of conventional locomotives and cars.
Also in 2005, Pacific National purchased four GT46CWM locomotives. These were effectively low clearance SD60s and EMD asked a couple of times "are you sure that's what you want?" Since then PN have purchased only AC traction locomotives and all those are fitted for ECP braking.
These are all private companies worried about the bottom line. They don't see their competitors internal costs but they see the published profits. They do see the ECP trains running with fewer brake and wheel defects, able to run faster due to reduced stopping distances for service applications.
When PN decided to enter the Central Queensland traffic to compete with Aurizon there, they purchased only ECP braked equipment and only AC traction locomotives (including AC traction electric locomotives).
Back in the Hunter Valley, there has been a downturn in coal and the ECP trains are running and conventional trains being stored.
A whole order of coal hopper cars from the 1990s suffered from cracking due to the stainless steel being unexpectedly harder than specification. As these are replaced with new car bodies using the old trucks and couplers and brake gear, the new cars get ECP instead of triple valves. It costs a little more but the operators see a tangible return on the investment.
Even visting USA executives should detect the trend.
EuclidBut aside from that ability, it won't actually be needed because there will always be some interval between the sensing of a derailment and the breaking of the wire in which this system can tell the controller where to differentiate the two sections of cars, thus telling the trailing section to go to full default “Emergency” response; and tell the leading cars to go to the specialized reduced force response.
This is an interesting control modality. I am tempted to suggest that you modify it slightly by commanding a full service brake on the rear, which then transitions to 'emergency' as soon as the brake gear has set up completely under control (or automatically, as currently provided, if the trainline continuity becomes lost. That will give you the full time benefit of ECP for commanded reduction without losing at least the ability to modulate both halves of the train under control until everything is predictably set 'equally' for max braking.
At some point there may be a period of time where the front end needs to 'stop' more positively than the max service set, but less heavily than full mechanical emergency. I see a potential problem with excessive draft-gear or coupler force in that instance if you are controlling differential only with the front railed 'end' of the train. (This may not be significant in a given accident's consist, but I worry that in a very wide range of contexts it might be).
EuclidI have not correctly stated my thoughts about modulation above. When I ruled out the wireless control for modulation, as I mentioned above, I ruled out modulation as well. I should not have used the term “modulated response” in my second paragraph. Instead, it would be a predetermined response of lower braking force on the leading cars compared to the maximum force response on the trailing cars. But it would be the same force throughout the stopping time. This predetermined response would be determined by the variety of train characteristics that I mentioned above.
This is a place where you and I wildly disagree.
Even my father's old system for Conrail circa 1988 'modulated' the train-brake application continuously over time, and he thought necessarily so, so as to avoid the then-significant problems that occurred with 'positive train control' brake application of the sort suitable for passenger or commuter consists when applied indiscriminately to a much longer freight with unpredictable makeup or braking characteristics.
In my opinion it would be completely pointless to even attempt doing stretch braking with two 'average' brake applications on two halves of a foundation-braked train, even if the physical brake action were continuously proportional and progressive (which it is certainly not!) You will very quickly have a systemic and increasing load (and remember it HAS to stay in tension; even a moment's compression or run-in means almost assured disaster) that will involve very substantial inertial mass backed up by a large amount of aggregate friction and contact area. So when the draft-gear travel goes out you will essentially have the whole force yanking on one of a very few couplers in the derailed portion of the train. And if the derailed portion decides to dig in, it will bring its considerable deceleration to bear on the leading coupler connection, which will snap something and result very quickly in 'divergence' of the now-unguided first car and probable establishment of accordioning.
In my opinion, we need a better (and more detailed) description of exactly what the 'derailment detectors' consist of, how they operate, and how they are arranged to gracefully degrade where possible. There are a number of different modes of derailment, each of which may 'deserve' or demand its own specialized response that needs to be implemented no less quickly than in seconds, with little control latency given the comparatively long response of the actual power-braking foundation on current freight cars. But modulation fore and aft of a 'differentially-braked' supposed derailed car, whether or not trainline continuity is lost, is desirable to preserve. (Lest the cure become worse than the disease...)
EuclidRegarding the point of whether this is an “Emergency” application or a “Service” application, I do not entirely understand how these are distinguished in an ECP system.
If you read the technical documentation for either the NYAB or WABTEC system, you will get this. In a typical system the ability to 'big-hole' the pneumatic part of the system is retained, essentially 'downstream' of the modulating electric valve. Emergency then consists of full line pressure to the cylinders, the 'difference' in setup speed being only that the electric "bypass" valves all operate immediately, and a bit more positively via their actuators, than a sonically-activated triple will. In addition, the available air pressure in the trainline is 'full' at all times, so there may be some makeup to the main reservoirs as they supply the cylinders and thereby a higher cylinder pressure can be established earlier or reached 'at the limit'.
Modulated 'service' application, on the other hand, can use all the available ECP advantages, including different braking limit rates detected for car types or loadings. It also includes the inherent ability to use very heavy or quick application of the brakes without having to 'worry' that you'll have to 'live with the consequences of your actions' afterward until you can recharge the system enough to release the excessive set.
In any case, in my concept, the application is not triggered by the parting of the wire. It is triggered by a signal sent directly to the master controller by the affected derailment detector.
I am not saying otherwise in any respect. The 'parting of the wire' is the point at which either modulation and control power is lost to the trailing consist, or some or all the sensor input from that part of the consist becomes 'indeterminate', and while not in the disturbing sense of the Browns Ferry fire, this may be problematic if a short between the conductors compromises either the 'last known good' or 'emergency default' positions of the ECP brake valves, or compromises how those valves subsequently act under autonomous control or self-power.
The 'emergency' here would actuate not at the time the first detector activates, but at the time a loss-of-control event occurs (I think a direct parallel to the kind of issue buslist was describing, where any anomalous event in the wireline signal integrity causes an immediate 'penalty brake' to full halt for reset.) The issue is then whether subsequent reaction of the trailing brakes in your system goes directly to 'mechanical' full emergency release, or to a graceful-degrade failover of some sort in which some of the potential advantages of ECP control can be retained.
This application would be as quick as possible with ECP; on both sides of the derailment; but will follow the modulation program with some reduction of force in the cars ahead of the derailment.
The point I was making here is that it will not be 'as quick as possible with ECP' on the trailing part of a wireline-controlled system (you say you no longer have either RF or inductive control modality) - as soon as the signal integrity is lost to the trailing consist, you will at best be using some sort of AI or expert-system emergency self-modulation of that part of the brake system. Much more likely, you will be in full emergency at that point in an all-out effort to keep even one car from accordioning into what is soon to be a very substantial and stopped obstacle to the trailing cars.
Now, what happens to the brake application of the consist ahead of the derailed cars, especially as those cars may be dragged by their couplers and in fact may be starting themselves to overturn, is a separate issue. Here buslist et al. also think full emergency (to get the momentum of 'the inevitable' as low as possible before the rolling starts) is the most appropriate response to go to. The 'catch' here is that the derailment sensors on any of the cars in this section aren't providing you with 'additional information' you can use for the modulation decision, other than to give a heads-up and a record of how the actual derailing progressed over time. What you need are the differential draft-gear tension and extension sensors, and I have been tempted (although not very far, up to now!) to modulate the 'differential braking' so as to actually try to get a knuckle as close as possible to where the derailed cars are connected to the as-yet-underailed part of the train... then actually retard the braking rate to ensure that an underailed leading portion won't be run into, and perhaps further damaged or derailed, by whatever is 'known off the track' to enough extent to part the line. (Railroaders will be laughing hysterically at this point, but it does represent one of the branches that control theory reaches when this extent of brake modulation can be provided.)
EuclidMy concept of differential braking retains the faster application advantage of ECP on the entire train, but withholds some of the braking force on the cars ahead of the derailment.
So does mine. I'm just pointing out where the differential braking attempt needs to be discarded and the throw-out-the-anchor type of emergency braking would start to take its place. In part this is because any developing derailment is going to part that communication/power line, and you MUST have the brakes in failsafe before that happens. Might as well be in mechanical emergency as be trying to modulate brakes that no longer 'hear' you or can respond electrically...
As I think both I and my father have argued previously, we think it is a 'better' idea to keep an ECP train in heavy service than in 'emergency', primarily to avoid the usual emergency-related further damage and derailment propensity. buslist and a couple of other actual industry participants disagree with this, and I for one do not even propose to suggest that this 'difference of opinion' means that they are 'wrong'. Note particularly here that my method of ECP control calls for some distributed 'backup' modulation and actuation method other than that from a patent 220V 'cable', so maintenance of some, even all necessary modulation for differential segment braking and wheelslide reduction is preserved even if cars twist enough to compromise the line or the train parts. Be sure that when yours does, too, that the previously-expressed concerns about crosstalk, false 'safety' or penalty applications, etc. are addressed...
EuclidWhat do you have in mind with regard to your above comment?
What I understood him to mean was that faster setup and action helps keep as much as possible of the trailing part of the train from 'running into' already-derailed cars that have come to a stop with the usual digging-into-the-ballast kind of alacrity, at most any rate in excess of what cars can achieve with braked steel wheels on steel rails (or even, as seen in the notorious 'tornado video' steel structure sliding on steel rails or on relatively smooth ties and ballast adjacent to rails...)
The differential braking concepts still, I think, apply, but this is a very different thing from either what M636C means or what buslist has said in previous posts on this general idea.
Some of this does hinge on the idea that as soon as derailed equipment starts to dig in it is desirable or even 'politically' (or legal-liability-avoiding) necessary to put the rear part of the train in emergency ASAP to lessen the magnitude of the pileup. This whether or not up to that point differential braking were being used to keep the derailed equipment from digging in. I think it can be easily established that current systems of ECP do NOT have the capability to modulate braking in the same way they do for service -- that is part of the reason I consider the actual emergency-brake distance and time difference between regular and ECP brakes to be so tiny -- and therefore when any part of the trailing consist has to go into 'maximum effort' braking to avoid further pileup, any attempt to control what the derailed equipment is doing via differential modulation will effectively end.
Much of the effect of very quick and positive braking of a consist following an embedding derailment can be achieved with multiple inline venting of a conventional brake pipe, at vastly lower cost. In my (not very humble in this context) opinion, the complaints that have been made about problems with practical implementation of such valves (including freezing under Canadian conditions of weather, design, and maintenance) can be overcome, or systems that directly address the problem can be designed that will work in normal railroaders' practice. Perhaps that ought to be at least an intermediate priority ... but of course it would probably become an excellent case study for 'good driving out better' in the evolution of train brake systems.
M636C In earlier posts I've suggested that the faster application of brakes on the trailing section of the train AFTER a derailment would reduce the damage, and this feature of ECP braking might not be reflected in a theoretical calculation of braking distance compared to a Westinghouse system. M636C
I am so damn tired of this stupid posting interface kicking you out for hitting the wrong arrow key by mistake, after it throws you somewhere else in the post at random when you try to use italics or boldface. It doesn't appear to be machine- or even platform-specific, so I conclude it is either a bug or misconfiguration in the Kalmbach site code. Why can programmers not figure out how to program for actual people?
I suspect much of the objection by American railroads to ECP involves the great investment in 'stranded cost' proprietary equipment, systems, and associated maintenance and training to make even a limited change within current operating models. In my opinion, the net effect on safety is reduced, rather than improved, in a scenario where train-handling comes to depend on improved ECP performance that 'may or not' be present in an actual train; I believe this mirrors the same phenomenon during the long introduction of air brakes 'as mandated' in the early days of Government 'safety regulation' involving Lorenzo Coffin et al.
I also suspect there is a certain industry reluctance to be the 'early adopter' that pays all the costs for the learning and acceptance curves, and the initial prices before production and aftermarket bring the unit costs down and 'commoditize' many of the parts and standards. With reference to 'reading between the lines' in the posts regarding the Heavy Haul conference -- where were the Australian papers and commentary/questions on the Chinese presentation? Where are the non-manufacturer papers and discussions, even now, that make the right 'business cases' for ECP adoption (even in the absence of likely support or even toleration from the AAR or cognate organizations)?
I personally find it outright astounding that either WABTEC or NYAB hasn't issued a pointed position statement or white paper that clearly defines how their system does both service and emergency braking, and equally clearly deconvolves the confusion about 'emergency' and 'service' stopping time and distance in clear, unequivocal English sentences. That one document, coming from a verifiable and reasonably expert source, would close down most of the waste of time, money, and talent in the current 'mandate' furball/cluster.
imho.
M636CI think that if USA railroad managers had closely inspected the operating ECP trains in Australia last June they would have had to reconsider their position.
Our community is FREE to join. To participate you must either login or register for an account.