A rathery scary vision of the future from "Trains" magazine

[Allegheny1600]

Hi All,

Browsing the always interesting "Trains" Magazine website, I came across this intriguing article on how the Union Pacific is planning to tackle train operating in the future: http://trn.trains.com/news/news-wire/2018/06/01-union-pacific-gears-up-for-technological-revolution

(subscriber only content, sorry), which prompted the following response from a reader called Michael Colby:

 

"Union Pacific has just made its first moves to become the railroad of the future. To quote “Positive train control will make one-person crews or fully autonomous operation viable sometime after 2020.” Having worked with and sold CNC machine tools for most of my career, I have come to understand one fact. When computers are adapted to a task such as machining they will always outperform humans. 

 

For a long time, I have been considering the long-term survival of railroads and whether they will be eclipsed by technology and all but disappear from the evolving world of transportation systems of tomorrow. Or can they reinvent themselves into a competitive transportation system. 

 

Imagine it is Saturday, August 4, 2040, and thousands of people have gathered along the Global II, Proviso Yard to Global III in Rochelle right of way of the Union Pacific railroad. This day will mark the last operation of a freight train with a human crew to be operated on a main line railroad in the United States. The locomotive is a GE ET44AC a 21-year old diesel electric that was the last new engine designed to be operated by a human crew that was bought by the Union Pacific railroad. The locomotive number 4026 will be pulling a train of 20 cars, a mix of boxcars, covered hoppers, a tank car, and a TTX flat car with 2 conventional truck trailers that have been borrowed from a museum. The end of the train will have 3 cabooses, also borrowed from a museum, to provide a place for the VIPs to ride. At 12:00 noon the Union Pacific, locomotive 4026 sounds its horn 2 times and the last human-crewed freight train departs Proviso to keep its date with history. On completion of this run, the train will be deadheaded back to the Illinois Railroad Museum and presented to them for preservation. 

 

Okay, so your first thought is this can’t happen, but why not? Over the years almost everything that I have known about railroads has changed or gone away. The only constant that I can count on is 4 feet 8 ½ inches. As long as the model of how to run a railroad is still to move the least time sensitive freight. Tie every car in the yard onto the fewest number of locomotives that will move the train, and re-sort all of the cars when you reach a major terminal, you will not be able to compete with the soon-to-be deployed robotrucks. 

 

The winner is always any enterprise that adapts the best technology to the problem. An article in Teckcrunch recently described: “A convoy of self-driving trucks that recently drove across Europe and arrived at the Port of Rotterdam.” They also stated that. “It is estimated that the cost of operating a self-driving truck from New York to Los Angeles will be cut by as much as 75% without having a driver.” It has been proven over and over again that once you start to adapt a computer to a given problem, the computer will prevail and do the task better, faster, and cheaper. 

 

With this kind of threat, how do railroads survive beyond a few main lines hauling the least desirable freight with the lowest return in profit? Only by rethinking how the railroads operate and making drastic changes can they hope to survive. 

 

The model for the future railroad needs to be fast, frequent, and efficient. Only by attaining all three can they expect to be relevant by the middle of the 21st century. Railroads of the future will have a major advantage in controlling their own right of ways, unencumbered by the traffic jams that the robotrucks will be facing. The train of the future will be able to move over long distances with predictable schedules and high speeds. Speeds of 100 to 120 mph will need to be attained to properly use the new model correctly. With speeds of 120 mph, markets that are out of the question today would be competitive for this new system. Overnight air cargo could be challenged, within a 2000 mile distance, with the new trains making the 2000 mile trip in about 17 hours. By noon of the next day, deliveries would begin to be made. 

 

Frequent service is required to match the robotrucks ability to get on the road as soon as they are loaded and keep driving around the clock until they get to their destinations. Departures for the new railroads need to be no more than 30 minutes apart, or as soon as you can load 15 cars all heading to the same location. Trains should be running like streetcars, one right after another, performing more like a conveyor belt than a barge. 

 

Efficiency is the most important of the requirements of the future railroads. The first item to be considered is the locomotive. The 4,500 HP engines of today would be overkill for the short, fast trains of the new railroads. The new engine at 2,500 HP would be a computer-controlled machine that uses positive train control, GPS, and a digital model of the railroad. This engine would be a hybrid genset that would be paired with a lithium battery tender. Being a hybrid, all dynamic braking or coasting would charge the batteries in the tender instead of wasting the power running it through resistance grids. The saved power would then be used to accelerate or maintain the speed of the train. Also being a genset, the engine would use only the required number of its 3 diesels to maintain the necessary speed. 

 

A key piece of the locomotive technology is the lithium battery tender. When the engine arrives at a terminal, the tender is uncoupled from the engine and is moved to a charging station. A fully charged tender is then coupled to the outgoing engine to provide power for the traction motors until it is depleted; then the diesels start to power the traction motors. The charging stations will receive their power from the solar cell arrays that cover the entire facility. Wind turbines could also be erected to generate electricity for charging the tenders. 

 

All freight will be moved in containers whether it is grain, car parts or anhydrous ammonia. The double-stack car will have couplers that are more like transit system cars than standard railroad cars. All connections will be made when the couplers come together, including air and electrical connections. The brakes will be air but applying them will be through an electronic system and not by reducing the air pressure of the train line. The double-stack cars themselves will be much lighter than they are today because the design requirement will be for a train that is never longer than 15 cars. Aluminium or carbon fibre construction would produce drastic reductions in the weight of this equipment. 

 

The terminal of the future would be similar to any container terminal of today. It would look like a big parking lot with tracks running through it. There would be two major differences. The first is the entire facility has a roof of solar panels collecting light and making electricity to charge the tenders and power the terminal. Where wind is abundant, wind turbines would supplement the power production. (Just think about how much power you could generate in Amarillo or Cheyenne.)

 

The second feature of the new terminal would be computers controlling all loading and unloading of the containers. This lights-out terminal would use bluetooth-retrieved data from the container to sort outbound loads for one destination and place them on to a train headed to that location. The time allowed to load a train will be limited to no more than 30 minutes. The train would not be longer than 15 cars. When the condition of 30 minutes or 15 cars is achieved the train departs. When a train arrives the containers are removed by computer operated cranes and placed on a waiting robotruck. A bluetooth device on the container identifies it and downloads the GPS coordinates of its destination to the robotruck, and it is on its way.

 

Like everything else on this future railroad, how it is operated will change to accommodate the new trains. Single track medium to high density main lines will be built with numerous 45 car length sidings. The computer will control the speed of an opposing move so both trains arrive at a given siding at the same time. Both trains continue to roll with one holding the main line and the other entering the siding. The trains pass each other, never stopping, and the train in the siding reenters the main and both trains accelerate back to track speed. High density double track lines will be able to run trains bumper to bumper. There would be no need for block signals to separate or restrict trains by 2 blocks. A group of 2, 3, or more trains can be moved closely together at high speed and operated as one train by the computer. All of the engines will receive identical commands at the same time and act as one. 

 

And yes, there will always be a place for a human in this new railroad. Programmers, administrators, inspectors, and technicians will be required to keep the railroad running. 

 

My views may seem like they came from Isaac Asimov, but to not rethink how railroads do their business is condemning them to join the horse-drawn freight wagon in history."

 

I think Mr Colby has a bold vision of the future that could quite feasibly turn into reality, maybe not by 2040 but possibly by 2050 or beyond.

I know that no-one can really predict the future, of course, but most things mentioned are currently possible and/or being developed.

Thoughts, anyone?

Cheers,

John.

Edited June 15, 2018 by Allegheny1600

[dave1905]

Caveat:  I don't have a subscription to Trains so I haven't read the article.

 

While unmanned trains are possible, and will probably come to pass eventually, a lot of what John is proposing will not.

 

Unmanned trains are not as important as they once were, back "in the day", when most of the expense of operating a train was crew costs.  Today, not so much.  Fuel and equipment are a bigger slice of the pie.  They are certainly NOT 75% of the railroad's cost.  Fuel is the big cost.

 

The statement that you won't need high horsepower engines is contradicted by the idea of running high speed trains.  HP = speed.  That's why passenger trains have a LOT of HP, that's why intermodal trains have a LOT of HP.  There is also no need to run most of the tonnage at 125 mph.  There is no economic reason to move 20,000 tons of frac sand at 125 mph.  A bulk train runs at about 1/2 hp/tt, while a 60 mph intermodal runs at about 3 hp/tt.  To get to 125 mph you would have to get to 5 or 6 hp.tt.  That means 10-12x the hp of a bulk train today. Horribly expensive from an engine or fuel standpoint.  And wasteful since there is no economic need for 125 mph service.  To support 125 mph, that means that everything has to operate at 125 mph, mixing 50 mph coal trains with 125 mph manifest and intermodal trains isn't feasible.

 

The other problem with higher speeds are that 125 mph will require new right of ways be constructed because current right of ways will not support 125 mph operation due to curvature, switch alignments, grades, terminal arrangements and junctions.  If a double stack is limited to 40 mph around a curve now, it doesn't matter whether its on a 10 car or 100 car train, what the HP is, it will be restricted on that same curve to 40 mph in the future.

 

Putting the tonnage of current trains in all container is very inefficient.  When you break a shipment into smaller containers, the tare weight of the containers increases as a percentage of the total weight of the shipment.  That means you are hauling more dead weight and less commodity.  Weight = HP = fuel.  Very expensive.  That's why railroads have for the last 175 years tried to minimize the weight of the container/car and the capacity in weight and volume has increased so much.  That's why container shipments are more efficient than TOFC.  

 

The other thing that I always marvel at is the insistence at running tiny trains.  The efficiency of a railroad is running big things.  In dispatching (and whether its by computer or by human the same rules apply) a short train operates by the same rules as a long train and in many ways takes up the same space as a long train.  Ten 20 cars trains take up waaaaay more space than one 200 car train.  With small trains each train has to have stopping distance between it and the train ahead.  So between the 3rd group of cars and the 4th group of cars there will be about a mile of space.  On the other hand, between the 3rd and 4th group of cars on the long train is about 3 ft (coupled) or 210 ft (set of mid train engines).  With Distributed Power (current state) or electronic braking (future state) the braking distance of a 20 car train and a 200 car train won't be that different.  The distance to stop ten 20 car trains is longer than the distance to stop one 200 car train.  The stopping distance now is controlled by how long it takes to propagate the signal through the train.  With one 200 car train with electronic brakes all 200 cars will need seconds to have the brake signal propagate, with ten 20 car trains the stop signal needs to be detected an communicated to the following train and then propagate 9 times.  We won't even get into the complexities of what has to be done at either end of the railroad to terminate 10 short trains as opposed to one long train.

 

A lot of this sounds way cool, but if you actually look at the costs to do it, it becomes horribly expensive, more than the market will bear.

 

The real savings in automated trains is in the long runs of "bulk" trains across the western half of the US.  Really long runs with minimal switching required, hub to hub operation where the time consuming yarding is minimized.  Yes, automated trains will happen at some point, just not the way it was described by John (unless somebody just has a couple trillion dollars to burn).

Edited June 15, 2018 by dave1905

[runs as required]

Turning the telescope 'wrang way roond' another future might be branches and byelines like the Windermere Branch being 'Community run' (like pubs) with whole brigades of RMweb enthusiasts staffing them - (under the standard wage/benefit proposals going the rounds of 'futures theorists' and I believe tried in Scandinavia)

 

dh

[Enterprisingwestern]

That John bloke must have the darkest of rose tinted glasses on, and I admire his confidence in the perfection of computers.

 

Mike.

[royaloak]

Its such a waste having one human in charge of a train carrying hundreds of containers and yet having one human in charge of a lorry carrying one or two containers is seen as acceptable.

[PaulRhB]

Why in a fully automated railway would you be swapping out fuel tenders? Surely a container dropped in from an automated crane would be faster? The main weakness in all these automated systems is the environment. A small 'sealed' system like the Docklands is relatively easy to automate. Automating a 2000 mile freight run is something else entirely.

The environment they operate in is extremely hostile and causes wear, coal dust, brake dust and whatever is in the atmosphere all contribute and there's no real advance if you  have to reduce service periods to cater for those for reliability. I can think of several trials on UK railways where heavily researched new solutions faired badly when exposed to the real weather and in the case of some point lubrication human excrement!

What happens when a flexible hose or coupling bar breaks? What happens when it actually hits something, the dynamics of stopping thousands of tons is no different when a large animal, tree or vehicle is across the track regardless of who or what puts the brake in. You'd need someone or something capable of assessing and fixing the fault as you can't wait 8 hours for a helicopter to bring in a repair crew who may or may not have the right spares to hand

 

While computers may well significantly assist with fatigue and provide extra levels of safety I can't see it going past semi automation for a very long time as nothing has yet matched humans for the dexterity in the field where they can clamber over, through and under to find and fix faults We are pretty useful at multi tasking to keep things moving.

[TheSignalEngineer]

In the early days of experiments with automatic route setting on BR some assessments were done on a busy section of panel comparing the performance of a human operating in the real world and the computer making decisions on a simulation of what was happening on the ground. For the first six hours of the shift the human performed better than the computer program as it then stood.

Prior to that I had a play with a system at Derby Research ( c1972) which was called Junction Optimisation or something similar. It was running a simulation of Weaver Junction to Crewe. It was never heard of again as far as i know.

In the 1990s Railtrack promised Beardy that the WCML would run like a dream with the new system they were going to put in at Saltley. They even signed a contract with US company Union Switch & Signal to install the system, despite it had never been successfully proved in service. We've still got Victorian signal boxes with levers pulled by humans  working some of the signalling that it was supposed to be supervising about 15 years ago.

Development of railway automation sure takes a long time.

[jjb1970]

I think increased automation and some level of autonomous operation will happen, but I don't see that as being synonymous with unmanned.

[DCB]

Why not use Mr FW Webb's patent liquid pick up system of a trough between the rails and a scoop for refilling the locos fuel tanks?

 

The big time saving will be not needing to stop when someone gets run over.. The big problem for UK designed motive power would be the need to carry an electrician, a mechanic and a computer guy to keep the thing running.  The alternative, breaking down 100 miles from civilisation because of a faulty sensor...

Edited June 15, 2018 by DavidCBroad

[dave1905]

Quite frankly with today's engines, if its not something somebody can fix by resetting a switch, they aren't going to fix it on line anyway.  Even with 2 man crews, if something goes wrong they have to call for help.

 

What would probably happen is that there would be roving "utility" trucks with one or two persons who are qualified as electricians or car men or train crew.  If something broke they could do the minor repairs, if really failed they could manually run it to the next terminal.  At the terminals, there would be two man crews that would use RCL to make any set outs and pick ups.  As things are now, if a train can see an obstruction, its pretty much too close to stop before it hits it.  Terminals are an issue because in terminals there are more manual switches and variations in routing that would be horribly complex to automate.  That's why over the road, long distance runs are the easiest.  You start the train at A and stop it 1000 miles away when it needs gas again.

 

Its a function of money.  PTC is only on about 30% of the US rail network.  Because of that engines would still have to have all the manual controls, seats, cabs, watercoolers, etc.  Any way you go its a huge up front cost.  Will the railroads want to invest trillions of dollars?

 

N American railroads invested billions of dollars to bullet proof the coal routes and then in 5 years the coal business vanished, leaving them with huge amounts of over capacity.  Will they risk that again?

[mdvle]

First, percentages need context.  Not to knock a 75% cost reduction, but without knowing the actual amounts for trucking and rail it is hard to know how it changes the relative economics of the two.

 

Waymo (self driving vehicle division of Alphabet, who also owns Google) has driven over 5 million miles in California with self-driving cars, and for a year now has had 400 volunteers using Waymo minivans as a taxi service with no one in the drivers seat.  Apparently current plans are to launch a driverless paid version in Phoenix by the end of the year.

 

So there is no question this is coming, and likely sooner than many would like.

 

This in turn means it is coming for trucking sooner than many would think, though I suspect governmental interference could slow things down.

 

As to where it all leads, interesting question.  A lot depends on how society adapts (or perhaps if it adapts).

 

One viewpoint is that all those workers put out of work become unemployable, at which point demand goes down and road congestion becomes a non-issue unless the government intervenes in some way.

 

Even if mass unemployment doesn't happen there is no way to know how things will happen.  Congestion primarily occurs in urban areas, and moving to rail doesn't solve that issue.  On the other hand, computer driven vehicles can help by eliminating many of the human behaviors that cause some of the congestion.

 

The biggest question really is can they program/teach the driverless vehicles to deal with snow.  If they can't then areas of the world that get more than the occasional snow are going to be in trouble as they become economically noncompetitive.

 

More specific to trains, it all depends on cost and continued demand and legislation.

 

The suggestion of high speed is a non-starter as explained above.

[Budgie]

Why not use Mr FW Webb's patent liquid pick up system of a trough between the rails and a scoop for refilling the locos fuel tanks?

Won't that spray the fuel all over the countryside within 50 feet of the trough?

[dave1905]

Sounds cool as long as you use a steel pan and a steel scoop.  That way one spark and the train gets a boost from "afterburners".

[Chris M]

Regarding driverless cars - can they cope with heavy urban traffic? This question especially applies where large numbers of potentially erratic and sometimes very aggressive city drivers will be mixing it with driverless cars. Will the driverless cars be able to cope with it all?

 

I’m wondering what will power heavy lorries in the future. With current technology you will need a lot of battery weight and space to shift 40 tons 500 miles or more. Technology will of course improve but it has a long way to go. We might see a move towards rail for longer distance in the UK if diesel lorries cease to exist.

 

I would have thought driverless passenger trains would be relativelyeasy in the UK. At the risk of being controversial I see a longer term future for a guard on every train than I do a driver.

 

There is a danger that driverless cars will cause a lot more congestion. If I had a driverless car I would ride in it to work then tell it to go home and collect me at the end of the day rather than pay to park all day in the town centre. This could potentially double the number of commuter journeys.

[newbryford]

Sounds cool as long as you use a steel pan and a steel scoop.  That way one spark and the train gets a boost from "afterburners".

 

There is a safer  way to pick up "fuel" on the go - overhead cables...…….

 

Cheers,

Mick

[dibber25]

Eliminate the human being from everything, and how is he/she going to earn the money to buy all the things the driverless trains and trucks are distributing? (CJL)

[scouser]

Eliminate the human being from everything, and how is he/she going to earn the money to buy all the things the driverless trains and trucks are distributing? (CJL)

Ahh, the 64 million dollar question!

Not just the transport industry either. It is already happening in other industries and a lot of people are not aware of it.

Good luck coming up with the answer.

[Ohmisterporter]

How are the unions reacting to these proposals?

[dave1905]

Eliminate the human being from everything, and how is he/she going to earn the money to buy all the things the driverless trains and trucks are distributing? (CJL)

 

Evidently you have missed the last 150 years of history.  Trains used to have crews of 5 or 6 , now they are 2, airplanes had flight crews of 4, now have crews of 2.  It used to take an aircraft carrier to retrieve a space capsule.  Now it can land itself without a human around for miles.

 

People will do other things.

 

Back in college I worked for Green, Tweed & Co.  Their claim to fame was they were the worlds largest manufacturer of surrey fringe.  The surrey in the song, "The Surrey with the Fringe on Top" probably had Green, Tweed & Co. fringe.  Guess how many people they had making surrey fringe in 1900. Dozens.  Guess how many people they had making surrey fringe in 1977.  Zero.  They were making hydraulic fluid seals for excavators and aircraft brake systems.  Guess how many people they had making Gulfstream jet brake seals in 1900.  Zero. Guess how many people they had making seals in 1977.  Dozens. 

 

Mutate, migrate, adapt or die.

[scouser]

I believe rose tinted glasses were mentioned relief in this topic.

The idea that other jobs will be created sounds nice. Try travelling through Spain, Portugal or Greece and you will find the reality very different.

Oh, and large parts of the UK!

[APOLLO]

Its called rampant unfettered capitalism / globalism.

 

We (well actually THEM - those who control) can do things today unimagineable just a decade or so ago. The pace of change is also ever accelerating.

 

Not a fan of it all myself, and I don't know many answers, but like dibber25 wrote I can foresee a whole new set of problems arising year on year.

 

Brit15

[dave1905]

I never said there would be a one-one transfer of jobs. 

 

The reduction of the need for human labor has been going on for hundreds of years.  Those that adapt, survive.  Those that don't, don't.  The will, resources and environment to adapt is not evenly distributed.  As far as technology goes, once its possible, its probable.  You can't put the poop back in the donkey.

[locoholic]

I never said there would be a one-one transfer of jobs. 

 

The reduction of the need for human labor has been going on for hundreds of years.  Those that adapt, survive.  Those that don't, don't.  The will, resources and environment to adapt is not evenly distributed.  As far as technology goes, once its possible, its probable.  You can't put the poop back in the donkey.

The poop goes into the ground, makes the grass grow better, and the donkey eats the grass. So the poop does go back into the donkey.

 

Just because humans have adapted to change in the past doesn't mean they will in the future. As has been pointed out, the pace of change is very fast these days, and still accelerating. Everything uses the internet now, and everyone assumes that the internet is unbreakable. We now have the most complex socio-economic system in our history, which has new and exciting modes of failure that we have no idea how to cope with. And then there's climate change and pollution...

 

What I'm trying to say is that your Micawberish belief that "something will turn up" is rather complacent.

[dave1905]

History has generally shown there are 4 choices:

 

1.  Supress or prevent the technology from being pursued, or shared.

2.  Prevent the technology from being implemented.

3.  Require the jobs to be preserved regardless of the technology.

 

Or …..

 

4.  Adapt or change to figure out what to do if the technology is implemented.

 

I am suggesting that the last several hundred years of history shows that #1 has never worked, #2 will work for a while, #3 will work for a while, but #4 is ultimately what happens.  Those that choose #4 (which is the hardest choice) early are more successful in the long run.

 

Which option are you proposing?

Edited June 27, 2018 by dave1905

[APOLLO]

The last several hundred years have not had the rate of technological / social change we have today.

 

It's more of a social problem than a technological one. How can quality of life be maintained / improve for the average citizen whilst implementing these changes ?

 

The short answer is it can't possibly be - so a race to the bottom (for most) or towards the top (a very small minority) is what we currently see (at least the start thereof).

 

Maintain the status quo ? - I honestly don't know.

 

Brit15