Former SpaceX engineers raised $50M to build a Tesla for freight trains | Hacker News Reader

Former SpaceX engineers raised $50M to build a Tesla for freight trains

fastcompany.com

41 pointsjpm_sd24 comments

Unfortunately Rio Tinto already beat them to the autonomy market. AutoHaul, their self-driving solution, has been in use for the past few years operating on around 1800km of track. Around 50 trains at any given time are currently driverless throughout Australia, including navigating level crossings without issue.

https://www.railjournal.com/in_depth/rise-machines-rio-tinto...

Those are huge long trains in Western Australia, and do not match the use case envisioned by Parallel. The use case of Parallel is to replace trucks on the last mile, and to be more interoperable with the five train networks in North America.

https://moveparallel.com/product/

Can I buy shares please?

I’m not a domain expert in electric propulsion or trains. My immediate questions:

1) It appears these are “trucks” and there is no actual “car” (frame). How much pushing/pulling force can a shipping container withstand?

2) The weight capacity is listed as 128,000lb per car, which I interpret as a pair of these trucks. This is compared to a road going semi truck but how does it compare to a conventional container car? Can still containers be stacked?

3) Is decoupling a benefit? It seems to be trains gain some benefit of shared braking ability by being connected, how do these cars work at the top of a steep grade? Is the speed and range limited by the heaviest car in a consist? Can the cars share power?

4) Can a regular car be retrofit to provide an electric “boost” but maintain the same “api” (coupler, beakes) to the rest of the train so they can be mixed in with conventional trains?

I think this is a really interesting idea but seems like it could be more successful if it was less radical.

Is it really better to have batteries on them than to run powerlines?

I presume the costs to convert enough of our existing rail networks to support electric trains (in the US) would be extremely high, especially considering so much of the freight rail network is owned by a few companies.

Freight rail seems like a great use case for large scale wireless charging in certain areas... have "rest stops" every X miles on key lines where trains can automatically charge up in the middle of a haul.

Im sure they could also take advantage of induction braking on the downhill section of mountain passes.

And because the route/topology would already be known the energy budget could be rather easy to calculate, e.g., enough to get it up and over the mountain.

> especially considering so much of the freight rail network is owned by a few companies.

Why does it makes cost high? I thought easier than what if many companies have rails.

Can they even fit the batteries on the trains?

I know there are a lot of variables involved, by all my napkin math gets me to somewhere around half a megawatt-hour of battery capacity required to move a monster train (e.g. ~18 kilotons) 1 mile.

Trains don’t really have a hard weight or length limit. Just stick another few battery cars on it.

but they can't have exactly 256 axles !

> Trains in Switzerland are not allowed to have 256 axles

https://twitter.com/standupmaths/status/1185544321933758467

Read the article. The idea is about batteries and motors on each car.

Apparently this attacks the inefficiencies in joining/splitting up cars at their destinations.

Right, but will this leave any room in the car for actual cargo? I'd like to see actual numbers (kg of battery / mile) here, my gut feeling is that lithium ion batteries are not good enough for this application, and they'll just end up using mostly diesel.

From the article:

> Parallel Systems cars can carry 128,000 pounds, or 2.8 times more than a semitruck. They also recharge in an hour: Without being plugged in or pulled off the track, the wheels can be recharged to drive another 500 miles.

Is this saying that a single charge can move a full load of cargo 500 miles?

I dont think it needs to pull all of it by itself.

I can imagine a system where in the areas the Train needs to stop/reload/whatever being surrounded by electrical infrastructure that at all moments is shoving electricity into the battery as fast as possible.

Then during departure, for the first few miles, electrical lines follow the track continuing to recharge the train while it gets up to speed. Afterwards, the battery is only really needed to overcome air resistance and friction, since so much kinetic energy has been given to the train for "free"

This is all just a random thought really, no idea if this is anything like what they are going for.

EDIT: should have used rolling resistance, not friction, so the final figure here is at least an order of magnitude too big, maybe two.

To me, the numbers just don't add up. Let's assume

- 0.2 as the coefficient of kinetic friction between the wheels and the rail [1]

- the car holds 100 tons [2]

- batteries can store 250 Wh/kg [3]

How much battery is needed to combat friction for one mile? (normal force times friction coefficient times distance, then scaled to battery capacity)

> (100 tons * 10 m/s^2 * 0.2 * 1 mile) / (250 Wh/kg)

> = 324 kg

So, 324 kg of lithium ion battery is needed per mile? Am I doing something wrong here because this looks completely untenable.

[1]: https://the-contact-patch.com/book/rail/r1717-friction-betwe...

[2]: https://www.csx.com/index.cfm/customers/resources/equipment/...

[3]: https://www.cei.washington.edu/education/science-of-solar/ba...

I don't see how the limit of friction is relevant here. Trains don't drag their wheels along the rail; the whole point of a wheel is that it rolls. A useful calculation would be based on rolling resistance.

totally forgot about that part. Rolling resistance seems to be at least an order of magnitude lower than friction so that does bring the numbers into a reasonable range.

wikipedia indeed says the rolling resistance for rail cars is indeed 0.002, so you are off by a factor of a round 100. 3.24 kg/mile is not that bad. Actually even 10kg/mile sounds very acceptable to me, so they may end up being able to use cheaper batteries like NiMH.

Rolling resistance, for trains, is just a small part of the equation. Air resistance and changes in altitude are also quite significant.

The article, with pictures, has a better description. This isn't about replacing diesel train engines with battery engines. Instead, it gives each individual car a battery and motors to propel that one car.

They're not mutually exclusive. Batteries could be good in combination with a partially-electrified train network. Recharge in places where installing a third rail is cheap, discharge in places where installing a third rail is expensive.

Power lines are way worse than batteries IMO. You need literally thousands of miles of them for the freight network, which is costly. Much of which is extremely far from population centers (and the ability to repair them). Battery powered freight trains — swapped out at places near cheap energy, like random wind farms and hydro plants — could potentially be a leap in reducing the cost of freight.

The US currently has only 1% of the rail network electrified, so running power lines is going to be a huge project. Europe went the other way and has on average 54% electrified. With the Benelux / DACH / France region even around 70% or so.

At only 1% electrified network is makes a lot more sense to put batteries on the train. At 70% it makes more sense to continue electrification where needed.

I think it depends on the country. The US has very long distances and the margin on trains is very small and competition fierce. They're also not subsidized. The trains need to be as cheap to run as existing trains and they can't afford a massive electrification effort.

This is scaling Kiva robots from warehouse to country level. Pretty obvious once you have the funds secured.

Reply on news.ycombinator.com