Where are the battery-electric locomotives going to be charged?
Frankly, even if greyhounds' approach were to be implemented only on the scale of CP Expressway, it would be worth trying -- if he could secure the financing to build the specialized equipment and put it in place, and if he can secure enough ongoing, regular business willing to pay for its extra cost (or its cost over autonomous electric trucking, which is the 'default alternative'). Note that a suitable 'power car' could be made low-deck, to match rakes of flatcars on either side (like the arrangement in some of the British Autocoach trains) with hardpoints to mount an appropriate guidance sensor suite on the ends of the set. As he said, none of the technology that would have to be used, aside from the guidance, has to be reinvented under conditions of uncertainty.
In my opinion, what Parallel Systems proposed, and what they now seem to be doing, have more operational and logistic holes than moldy and rat-gnawed Swiss cheese. It doesn't take the scientific method to establish that the fundamental engineering can't make the fundamental economics work. Keep in mind this is with level 4 autonomy for road vehicles stipulated as practicable by the time any commercial adoption of a Parallel Systems-style operation is actually marketed.
Exactly how did the RoadRailer people propose their equipment would be used with other intermodal consists, other than appended onto the back end as with other conventional equipment? (Something I confess I was hoping to see with that Canadian module thing was the use of a 'hostling' module at the head of the RoadRailer segment, where the transition truck would be, that could be used to allow cutting-off and moving of the RoadRailers if the end of a train of intermodal flats needed to be exposed quickly to back it to a ramp...)
That's why I think that the role, if any, for autonomous railcars is in the terminal (think self assembling and self disassembling trains). Individual railcars traveling down a line break two advantages of a train: first is more efficient use of track space/time; second the frontal area of a single car is about the same as a train. The trains would most likely be locomotive hauled.
Side note about braking: From what I've read, the highest reliable factor of adhesion for braking is 10%. At 30mph, this translates into a 300' stopping distance from when the brakes take hold (I'm in no way saying that a 100 car freight train is going to stop in 300' from 30mph). At 15mph and 25% factor of adhesion, stopping distance would be 30' assuming the brakng effort was applied instantly. I would expect an autonomous rail car to use very quick acting regenerative braking, but that assumes the sensors are up to snuff (right now, they're not).
Railroading as mass transportation, is not a new concept, Martin Stevers 1933 book, Steel Trails, makes frequent mention that railroading works best with large scale transportation, i.e. wholesale, not retail. OTOH, the problem with not pursuing revenue enhancement is that it may enventually lead to revenue declines.
jeffhergertObviously the way to get freight lost to trucks back on the rails is to act like a truck. Only without the flexability, convenience, and maybe once all is said and done, the cost of a truck. One other thing not mentioned. In the last 30 years or so many facilities, either production or distribution, have been built away from a railroad. Some locations with rail service have developed industrial parks away from the railroad, but with great highway access. This is a system that will either completely change the way that railroads are structured and operated (unlikely) or be such a small part of the transport picture that it's almost irrelevant (more likely). Jeff
One other thing not mentioned. In the last 30 years or so many facilities, either production or distribution, have been built away from a railroad. Some locations with rail service have developed industrial parks away from the railroad, but with great highway access.
This is a system that will either completely change the way that railroads are structured and operated (unlikely) or be such a small part of the transport picture that it's almost irrelevant (more likely).
Jeff
21st Century railroads are not interested in car load customers. Train load customers are all they really are looking for. Car load customers create the requirement for too much terminal support and thus too much cost.
21st Century railroading is all about cost avoidance more that it is in revenue enhancement. They want a much higer percentage of all revenue generated brought straight to the bottom line without incurring costs along the way.
Never too old to have a happy childhood!
Obviously the way to get freight lost to trucks back on the rails is to act like a truck. Only without the flexability, convenience, and maybe once all is said and done, the cost of a truck.
I would rather some answers be thought about before we commit the billions of tax dollars (you know the gov't will be the ones footing the bill one way or another) in a failing enterprise.
A to B is the easy part. Yarding, switching, delivering, inspecting, etc is the part that will be a bit more difficult.
The opinions expressed here represent my own and not those of my employer, any other railroad, company, or person.
greyhoundsI think some people are missing some important points here. Let's start with the reasonable belief that the folks at Parallel Systems are quite knowledgeable with things such as AI, computers in general, engineering, and sensors. Let us also acknowledge that they probably aren’t too familiar with how to move freight. Everyone also needs to realize that rail movement, as it’s done today, is noncompetitive for most freight revenue in the U.S. That’s because most freight moves less than 500-600 miles. Except for unit trains, current rail systems incur terminal costs and/or drayage expenses that are largely nonexistent for truckload movements. Rail generally does have a per mile linehaul cost advantage, but it takes a long haul for this cost advantage to offset the extra terminal/dray costs necessary for rail movement as it’s done today. This misses most of the market because most of the market is shorter haul. How can this be fixed? Well, if the automation technology of Parallel Systems was to be placed on a small battery electric locomotive, operated with no crew, small unit trains of maybe 6 T’d up containers could be economically viable. This would reduce the rail’s terminal cost disadvantage. Loading/unloading of the railcars could be done with autonomous electric yard tractors. Think of the CP’s Exxpressway terminals writ small. This does have possibilities. Think of how to use it and make it work. Not of reasons it won’t work. Problems are something to be solved.
Responses on here are largely reflective of a siege mentality, i.e., repel any progress to preserve the status quo. To quote John Cleese, "Typical, just typical."
Most respondents overlook your major point, which was that railroads of today have lost most of the transportation market to trucking because it is shorter haul. You are simply proposing a possible method to recapture some of that market segment.
BTW: Nitpicking as seen here does not resemble the scientific method.
Refresher:
1. Ask a question.
2. Background relevant research.
3. Formulate a hypothesis.
4. Test hypothesis, typically with an experiment but possibly correlations.
5. Analyze data and draw conclusions.
6. Communicate so others can attempt to replicate research results.
Duplicate
We've had new technology and equipment proposed a few times over the years. Usually it's something that's incompatible with current equipment and practices. Some say that's great. Yes, maybe the ideas proposed were sound, IF you were starting from scratch.
This autonomous vehicle isn't compatible with current equipment or practices. I think to try to use it in the existing network will only be detrimental to both conventional and the proposed systems. You will have to delay one system to allow the other to move.
The ones proposing it seem to have no idea how the current railroads are set up. They are not a public highway system. Sure, some loads might be able to travel a single line. Most are probably going to be interchanged at some point. You either need buy in from all railroads or government ownership of the infrastructure with open access. Good luck for either.
Even if with government ownership, automation of the vehicle, I'm not sure it will be available at a price to compete with trucks. Highways are public funded from user fees, but this no longer covers the costs. Funds from other taxes are used for the short fall in road funds. Don't expect that with rail infrastructure. Expect very little funding outside of user fees from the rail vehicle users.
The automating part is the easiest part.
If you're going to propose a new technology, it is best to flesh it out with a lot of details, like you put some thought into it. When you put out a bare-minimum statement, with no regard for real-world operations, you should expect to be torn apart. It's like me saying "there should be a passenger train between Detroit Birmingham, Alabama" and not knowing anything about what routes are available, what the track speed is on different routes, how many trains are on each route, etc. Anyone can come up with pie-in-the-sky ideas but if you don't do the work to show how and why they would be practical, from both an operational and financial perspective, don't expect them to get a warm welcome.
Overmod zugmann Is that the other side of the coin that says "this is a new way - it HAS to be better!"? No, the other side says "this is why we've always done it our way; show us how your solution does it better without the pitfalls we know too well..."
zugmann Is that the other side of the coin that says "this is a new way - it HAS to be better!"?
No, the other side says "this is why we've always done it our way; show us how your solution does it better without the pitfalls we know too well..."
But we have an army of consultants with fresh MBAs!!
zugmannIs that the other side of the coin that says "this is a new way - it HAS to be better!"?
Fred M Cain"Won't work. We've never done it that way".
Is that the other side of the coin that says "this is a new way - it HAS to be better!"?
[quote user="Fred M Cain"]
[quote user="greyhounds"]
This does have possibilities. Think of how to use it and make it work. Not of reasons it won’t work. Problems are something to be solved.
greyhounds "Greyhounds", I agree with this point you made at the end of your post 100%. I believe you are spot on. It seems that when someone posts a potentially good idea on a forum - any forum, really, other people try to do all they possibly can to point out the potential pitfalls and end up trying to shoot the idea full of holes. I think we would be better off to try and think other people's ideas over a bit more and look for advantages and how the difficulties and pitfalls might be overcome. This makes me think a bit of traditional railroad mentality where a new manager would come in with new ideas of ways things might be made to work better. Then the older, entrenched management shoots the guy down. "Won't work. We've never done it that way".
"Greyhounds",
I agree with this point you made at the end of your post 100%. I believe you are spot on.
It seems that when someone posts a potentially good idea on a forum - any forum, really, other people try to do all they possibly can to point out the potential pitfalls and end up trying to shoot the idea full of holes.
I think we would be better off to try and think other people's ideas over a bit more and look for advantages and how the difficulties and pitfalls might be overcome.
This makes me think a bit of traditional railroad mentality where a new manager would come in with new ideas of ways things might be made to work better. Then the older, entrenched management shoots the guy down. "Won't work. We've never done it that way".
Picking apart 'new' ideas is a part of the 'scientific method'. Every change from the orthodox has to be peer tested - be that some change in scientific understandings of the world or a changed in human activities.
If the idea can withstand the 'nit picking' of 'peers' and come out the other side as something that can draw investment and implementation, then you have something. If the idea can't make it through the process it was a less than practical idea. It does take a lot to move the inertia of 'we have always done it that way' to doing something different. aka 'If it is not broken, don't break it'.
greyhoundsWell, if the automation technology of Parallel Systems was to be placed on a small battery electric locomotive, operated with no crew, small unit trains of maybe 6 T’d up containers could be economically viable. This would reduce the rail’s terminal cost disadvantage. Loading/unloading of the railcars could be done with autonomous electric yard tractors. Think of the CP’s Exxpressway terminals writ small. This does have possibilities. Think of how to use it and make it work. Not of reasons it won’t work. Problems are something to be solved.
The big "cost reduction" idea here isn't just 'autonomous operation' of the train moving from one place to another; it's getting rid of all the costs involving switching and 'handling' of individual cars. Consequently one of their original working ideas, which made great sense if you were an SPV2000 kind of engineer, was to put the motive power and guidance entirely in the separate bogies, which were separate bidirectional entities, and coordinate autonomous container-handling equipment with 'hostling' of the autonomous bogies so that only one size of unit handled every use case.
Meanwhile they had some rather vaguely expressed idea of motorizing generally low-cost loose railroad cars with a version of the autonomous bogie that would stay with the car. I saw them kinda faking having three-piece trucks (for compatibility when operating in conventional trains unpowered) but little more than handwavium regarding actually powering (and passive dynamic-braking) existing geometry. Could that be done? Clearly I think so, but whether it would be cost-effective, even before all the new car-inspection and documentation rules from the Government come to be applied to that sort of thing, is quite a different thing.
Even a quick and cursory update of Kneiling's integral train to use battery power is superior to some autonomous transfer locomotive. That might best be made as three-unit wells (an update from his skeleton flat chassis) with some kind of deck and hold-down for TOFC sitting at the bottom of the well for easy conversion. Any number of these could be platooned (which, remember dumb cars towed around by the autonomous locomotive won't be) and inherently handle crossing issues (which, again, requires a fairly intricate sensor suite on 'the point of a shove move' even if the transfer locomotive is trying its best to be vigilant from the other end.
What you're saying is starting to sound to me like a battery/autonomous 'update' of that Canadian proposal that was going to have the little power modules spaced throughout its consist. Do you think that's going to work with a couple of orders of magnitude higher first cost, with interesting required security methodology?
There is still the last-mile dock problem. Essentially all the containers that are going to be used have end doors, like a van trailer but without onboard 'landing gear'. There have been a tireless series of wacky hydraulic unloading arm devices, extendable leg frames, and expensive and relatively slow and fragile container lifting and handling devices over the years. I have yet to see one that would go with a distributed move, lift and shift the container, and 'intermodally' position it either by road or onto some sort of chassis arrangement to get the end doors pointing where the customer can strip or stuff what's inside.
You'll get around some of that, I think, if the rail mode is involved with strategically centrally-located crossdock facilities (like Rotterdam in the UP model, but more 'granular' in location) where the autonomous mode transfer happens under tight and professional authority and control, and what amounts to LCL now schedulable point-to-point is transloaded to more purpose-bulit electric, level 4 with by-then-costed-down technology, last-mile delivery vehicles. (For example those of an outside provider that doesn't want their rail 'value chain' to have to be accounted in-house...)
There was an interesting proposal, probably a couple of decades old at this point, that involved using equipment like the Adtranz CargoSpeed (in two-unit sets, IIRC) to cruise to all sorts of facilities in the general northern New Jersey region, pick up containers or trailers, and move them over the existing rail system to keep them off congested roads even for short moves. They worked out very carefully where the bounds of the service would be, and what corporate entities (like Conrail Shared Assets on the 'chemical coast') would have to be involved. This was in concept a bit like the contrapositive of systems that provide rapid end-to-end execution (that few shippers are willing to pay extra for given the likely issues at one point or another with achievable QoS) -- it would give an as-available pickup and then meandering trip down the back tracks to get close enough to where it was going, with at least in principle some guarantee that both delays and intermodal handover could occur with precisely-scheduled 'precision'. There might just be enough perceived value in such a thing -- if all the tracks are still there and as uncongested -- cutting out the need for human presence and foreground attention to run.
OvermodThat was, amusingly, the original Parallel Systems concept, if you can truthfully call it that. The 'container' was carried as a stressed member between two little autonomous electric modules, a bit like having a FlexiVan supported on a bogie at each end. Those would automagically run themselves to the right distance apart for 'any' container, large or small, and (with some handwaving to be solved 'in development' would lock the container against displacement or overturning. Then the two of them would happily 'handshake' the necessary differential drive and braking information to fake the 'buff' and 'draft' at either end when you had the thing in a train. Note the additional step of 'reality' involved in their current approach, which appears to have a contact probe acting a bit like a British 'buffer' at each end that does physical connection instead of sensed distance for the critical distance between units while platooning. I presume they have enough college-trained engineering talent to figure out how to sense incipient node movement and start compensating before the resultant gets there. Let's really, really hope so.
I think some people are missing some important points here.
Let's start with the reasonable belief that the folks at Parallel Systems are quite knowledgeable with things such as AI, computers in general, engineering, and sensors. Let us also acknowledge that they probably aren’t too familiar with how to move freight.
Everyone also needs to realize that rail movement, as it’s done today, is noncompetitive for most freight revenue in the U.S. That’s because most freight moves less than 500-600 miles. Except for unit trains, current rail systems incur terminal costs and/or drayage expenses that are largely nonexistent for truckload movements. Rail generally does have a per mile linehaul cost advantage, but it takes a long haul for this cost advantage to offset the extra terminal/dray costs necessary for rail movement as it’s done today.
This misses most of the market because most of the market is shorter haul.
How can this be fixed?
Well, if the automation technology of Parallel Systems was to be placed on a small battery electric locomotive, operated with no crew, small unit trains of maybe 6 T’d up containers could be economically viable. This would reduce the rail’s terminal cost disadvantage. Loading/unloading of the railcars could be done with autonomous electric yard tractors. Think of the CP’s Exxpressway terminals writ small.
zugmannIf we're doing self-propelled why not just have the containers go directly to the customer?
Note the additional step of 'reality' involved in their current approach, which appears to have a contact probe acting a bit like a British 'buffer' at each end that does physical connection instead of sensed distance for the critical distance between units while platooning. I presume they have enough college-trained engineering talent to figure out how to sense incipient node movement and start compensating before the resultant gets there. Let's really, really hope so.
Fred M CainI have wondered if such an arrangement could be possible and maybe even desirable.
You may remember the original Parallel Systems scheme, as used to raise $37M of venture capital from people who, well, don't understand railroading too well. This was to have the 'trucks' as the autonomous modules, each with its own guidance and standoff system, and use the container itself as the 'carbody' between them in operation. I think the idea was that you'd use the same two units for any length of container, and effectively make it 'autonomous' without having to account for 40' functional equivalence, losses due to air resistance between containers in a platoon, etc.
Rather obviously, at least to me, once they actually starting having to consider something that would run, they abandoned this and now have a stout 'perimeter frame' with twistlocks, etc. (It still requires the stuff to handle intermodal containers on and off underframes, and it will be fun how they start getting around the obvious things that are going to come up with typical end-loading marine containers in customer end delivery and demurrage of the very expensive tech in the modules, but that happens long after the question of combining the loaded modules into a passive, presumably one-pipe braked and block-switched by baboons PSR monstrain consist.
Look for them to expensively and painfully discover why college AMS kids have trouble motoring three-piece trucks effectively, and then safely. Then look for the sensor and BITE follies as they incorporate active drawbar control into the consist with buff and draft nodes from "typical" train handling...
You should concentrate for the moment on the security issues for autonomously-dispatched, crewless container traffic in our current state of law enforcement. You'll need to have that one wholly addressed before you're likely to get any real long-term repeat business with insurance coverage you can afford...
daveklepper Yes, absolutely. 1. The one-car or group of automous cars train is not crewed. Automatic train oprtation is simply never going to be legal and safe on any railroad with infenced and at-grade right-of way. Complete isolation as in rapid transit systems and within land owned by the coirpration that is the customer and/or owner of the railroad are possible, and CN, former IC, western line c definitely fails the separation requirement.
Yes, absolutely.
1. The one-car or group of automous cars train is not crewed. Automatic train oprtation is simply never going to be legal and safe on any railroad with infenced and at-grade right-of way. Complete isolation as in rapid transit systems and within land owned by the coirpration that is the customer and/or owner of the railroad are possible, and CN, former IC, western line c definitely fails the separation requirement.
The autonamous idea that I've toyed with for years, would be for autonomous cars to only make the short part of the trip from the shipper's dock to the yard. Provided that those distances were fairly short, then a completely fenced-in and grade crossing free right of way would be doable.
Once in the yard, the autonamous cars could be marshalled into a long train with other cars having the same hub/terminal destination. The hub-to-hub portion of the trip would represent the long haul portion and could then use convential operation with a one or two person crew.
Upon arrival at the destination hub yard, the train could be broken up with the autonomus cars once again delivering themselves to the receivers' docks (once again that distance would have to be short by necessity).
I have wondered if such an arrangement could be possible and maybe even desirable. For one thing, historically, many shippers were already within a close proximity of rail yards to begin with. This could have the potential to revive so-called "loose car" freight. That would not only benefit shippers and the railroads but the public as well.
The decline and near total demise of loose car railroading - especially boxcars - has resulted in too many dangerouse trucks on the roads. Some of our highways are just about too dangerous as a result. I can see the time coming when we will need to try something different.
While it might have some utility in terminal areas, where vehicles can be sent out from a central yard to various industries/customers, I'm not sure it will catch on for any long distances. Long being a relative term, say 50 or 100 miles instead of 500 or 1000 miles.
Even then one must take into account what facilities exist at the shipper/receivers. Everyone talks about throwing the switch to get in or out of the customer's facility. In reality, you're lucky if there even is still a switch. Or if there is, that the track hasn't been lifted beyond the gate or property line. Even if there is a track, most rail doors have been blanked out. (Which may not matter since all the designs for this equipment looks to be for transporting standard containers. Very few have side doors, but that could change if needed.) Most customers are now truck oriented. Most will have to have the container taken off the rail vehicle and put on road wheels so it can be backed into a dock door.
When talking about rail routes for intercity moves, you can lose flexibility when the rail line only goes one direction. A load may have to go way out of it's route on rail when a truck usually has many route options available. In short, I just am not sure that trying to operate like a truck on a limited and fixed route is going to get much traffic. If a company prefers trucks over rail cars, I don't see much of a rush to convert to an autonomous container transporter.
The automating from point A to point Z is the easy part. It's all that goes on that isn't easily seen that becomes road blocks. If you can't get a customer to use rail, or a railroad to want to serve a customer, it doesn't matter what kind of equipment you have or can come up with. It's not the equipment that's the problem to overcome.
BaltACD Can't see this operating on anything BUT a line whose sole traffic is between a single on line shipper and on line consignee; with no other traffic using the line.
Can't see this operating on anything BUT a line whose sole traffic is between a single on line shipper and on line consignee; with no other traffic using the line.
Just like your toy train layout that has a Lionel hand car going back and forth?
Railroading in one's and two's is simple. Railroading in the thousands of carloads and hundreds of origin/destination pairs within the serving area is infinately more complex.
greyhoundsIf these cars operate over a line with PTC I'd say there would need to be PTC on one of the cars.
On a non-PTC line, it seems there would also need to be a positive way to stop an automated car approaching an occupied at-grade crossing with another railroad.
Reminds me of all those humpyards that were ripped out in the pursuit of PSR.
Now the cars are flat switched or there's block swapping (which just means crews further up the pipeline flat switched them out).
MidlandMike Who will align the switch when the autonomous car(s) leave the industry siding, or out on the mainline when it needs to take a manual switch siding? I don't remember if PTC was brought up, but would each auto-car need to be equiped with PTC?
Who will align the switch when the autonomous car(s) leave the industry siding, or out on the mainline when it needs to take a manual switch siding?
I don't remember if PTC was brought up, but would each auto-car need to be equiped with PTC?
I would imagine they could use something like this:
https://tracsis-us.com/solutions/yard-automation/remote-control-routing/
If a person can push a button to send a signal to line a switch, a computer can cause the signal to be sent. There would seem to be a need to equip yard tracks and passing sidings potentially used by the cars with RC switches.
If these cars operate over a line with PTC I'd say there would need to be PTC on one of the cars.
The signal activation issue would have to be handled proactively, probably with a defined modulated voltage across the wheelsets and monitoring of the actual shunting of the signal circuits. It would be easy to implement this at the same time as the braking and motoring arrangements for the "trucks" if done coherently.
I don't see any reason why either braking or acceleration of an eight-wheel loaded car differs any from the counterpart in a MU-controlled consist. That is especially so if traction and antilock braking control are adapted from OTS automotive sources, which are far superior to anything likely to be kludged up to be hose-compatible with one-pipe Westinghous pneumatic.
Assuredly, if manual switching or staging is required, the schedule that platooned and operated the cars will indicate when, and where, someone in a road or hi-rail vehicle would have to be at the time the car gets there. In the perfect world of the kind of engineers who did the SPV2000, the folks touting the autonomous service would be arranging for power switches (perhaps as simple-minded as trolley switches operating off controller position) for all those manual switches that have had to be equipped with PTC position indication and, presumably, some sort of safety interlock. That's another round of VC financing, but it need not be showstopping as such...
The thing that keeps cropping up since I started looking at level 4 autonomy as a practicable rail technology is all the overlap with Kneiling's integral-train construction, modified for battery instead of gas-turbine power and equipped with suitable decking and guideways for 'autonomous' yard tractors to load and unload. Those as I recall were to be built in three-car 'rakes' -- just the size most desirable for articulated well cars -- and it would be easy enough to increase the length to get single-level TOFC capability "per guidance system" that would equal the (as-yet-undeveloped-by-Parallel) stack-block equivalent.
Note that where the great pending foundering of their business model looms is the issue of operating these things in conventional trains, whether dedicated-equipment (as in RoadRailer) or the original dotty Parallel proposal of loose cars fitted with their conversion pacs. Much of the potential attractiveness of this kind of autonomous-car operation would hinge on actual carload delivery, rather than break-bulk LCL which is an interesting model but somewhat difficult to justify for the proposed technical methodology. Much of the stuff I've been playing with over the past couple of decades involves effective 'motoring' and control of loose car bodies sitting on three-piece trucks that can be towed at high road speed without dynamic instability, using as many OTS parts from the existing rail industry as possible. One at least potential "PSR" approach is then to allow the equivalent of slip-coaches from a point-to-point train appropriately blocked, with the dropped cars or blocks then decelerating under power and control to enter appropriate sidings facing either way.
Our community is FREE to join. To participate you must either login or register for an account.