Hacker News Reader

Comment by JumpCrisscross | Hacker News Reader

> The Hyperloop alpha paper explicitly contradicts this assumption on page 18, where the flow diagram gives the compressor input at 99 Pa and the output at 2.1 kPa (a 21:1 pressure ratio)

One standard atmosphere (atm) is defined at over 100,000 Pa [1]. The 21:1 pressure ratio in Figure 11 on page 18 is for the "passenger plus vehicle capsule" [2]. I'm talking about the tube.

Going from 99 to 2.1 isn't hard. Going from 100,000 to 100 and then keeping it there is.

[1] https://en.wikipedia.org/wiki/Atmosphere_(unit)

[2] http://www.spacex.com/sites/spacex/files/hyperloop_alpha-201...

View on HN

But that's no challenge at all. You're just describing a usual vacuum pump. And just a single-stage roughing pump (Harbor Freight, anyone?) can get down to less than 1 Pascal, so 100Pa is a piece of cake. That's only needed for slight leaks in the tube, not for the pod itself.

Keep in mind that the absolute pressure difference is far less for Hyperloop than for a typical natural gas pipeline, and yet the latter can keep leaks to an absolute minimum in spite of thermal expansion, etc.

> You're just describing a usual vacuum pump

Scaling is hard. The largest vacuum chamber we've built is a fraction of the size of the proposed Hyperloop. It was a very, very hard problem [1]. It's expensive, sucks up loads and loads of power and needs lots of thermal management, structural reinforcement and vacuum-off maintenance.

> Keep in mind that the absolute pressure difference is far less for Hyperloop than for a typical natural gas pipeline

Pipes are pressurized from the inside. See my comment from elsewhere in the thread on why vacuums are different [2].

All this said, I generally agree with you. I don't think building a giant vacuum is beyond current technological capability. I do think building one structurally sound and thermal-expansionwise stable enough to carry passengers at high speeds is.

[1] http://teamcorporation.com/images/technical_documents/Space-...

[2] https://news.ycombinator.com/item?id=15459873

>The largest vacuum chamber we've built is a fraction of the size of the proposed Hyperloop. It was a very, very hard problem [1].

Do we really have to say that a long extruded tube that only has to maintain low vacuum is a lot simpler than the SPS?

SPS: big thing made up of many custom one-off parts that runs at 1/380,000,000ths of an atmosphere and can blast test articles with intense simulated sunlight.

Hyperloop tunnel: big thing made up of many identical parts that runs at 1/1,000ths of an atmosphere.

>It's expensive, sucks up loads and loads of power

I think they know that, don't you? There's no line-by-line breakdown, but the paper allocates $260 million for the station+pumps and 21 MW for electricity consumption (pumping, accelerating pods, and charging pod batteries).

>needs lots of thermal management

I agree that you need it, but what specific problem do you see?

>Pipes are pressurized from the inside. See my comment from elsewhere in the thread on why vacuums are different [2].

You said a soda can is stronger when pressurized from the inside. That's true, but the obvious solution is to make it thicker than a soda can. :) The paper calls for 0.8-1.0 inch thick steel, which my math says is more than adequate.

> I think they know that, don't you?

Who's they? Hyperloop One? No--they've just punted the "hard" stuff to the end. (Kind of like the Wright Brothers designing the seats of their plane before getting it flying. Oh look! [1])

Elon Musk? Yes--I do. That's why he's waiting.

> what specific problem do you see?

The top heats up and the bottom doesn't.

> The paper calls for 0.8-1.0 inch thick steel, which my math says is more than adequate

Adding mass adds strength while increasing the time the structure takes to reach thermal equilibrium. The thermal gradient isn't itself a problem. But if you look at the forces necessary for the materials in question to tear themselves apart, and then consider their thermal coefficients of linear thermal expansion, you can derive a maximum tolerable thermal gradient given the size of each tube segment (we'll assume the problem of reticulating vacuum seals is solved).

When you solve for strength, you get too much material for the system to reach equilibrium before inclement weather either causes (a) the structure to buckle, laterally or (b) the outside of the tube to start shearing itself from the cooler inside.

When you solve for thermal stresses, you lose your strength. Microbuckling and microfracturing may not seem like a big deal, but it is when you're talking about 1 standard atmosphere bearing down from the outside with a capsule swinging about on the inside.

We need a strong material that either (a) conducts heat really well or (b) doesn't change shape when asymmetrically heated. We don't have something that meets those requirements yet that we can manufacture at scale.

[1] http://www.stacksmag.net/2013/10/space-virgin-galactics-rich...

>No--they've just punted the "hard" stuff to the end.

Citation needed. The link provided gives no support.

I would be interested in seeing your calculations on the thermal side. An ANSYS multiphysics simulation (the software SpaceX uses) would show the problem, no?

SpaceX has already built and used a mile-long, 11' diameter vacuum tube in Southern California, which would seem to put this matter to rest (the proof is in the pudding, after all).

> An ANSYS multiphysics simulation (the software SpaceX uses) would show the problem, no?

Yes. But SpaceX != Hyperloop One.

> SpaceX has already built and used a mile-long, 11' diameter vacuum tube in Southern California, which would seem to put this matter to rest (the proof is in the pudding, after all)

I'll go ahead and predict that the strength of that tube will have materially decreased after 1 year in the elements. I'll even posit that will occur independent of rusting, which appears to have unfortunately taken place, due to the formation of microfractures within the metal due to repeated lateral thermal flexing. That said, this was a demo track. It wasn't designed to withstand one standard atmosphere while whizzing fast, heavy capsules inside it for years on end.

>I'll go ahead and predict that the strength of that tube will have materially decreased after 1 year in the elements.

With steel, the real question is not the weakening from Year 0 -> Year 1, but from Year 4 -> Year 5. After an initial period of weakening the strength of steel levels off.[1]

Anyways, another user pointed out that a cheap foam layer can reduce the daily flexing to arbitrarily low amounts, and give an arbitrarily amount of time for the tube temperature to equalize circumferentially.[2] Cheaper than either a temperature equalizing water jacket or an underground tunnel. I know you disagree in that other thread, but I now consider this problem solved.

[1] https://en.wikipedia.org/wiki/Fatigue_limit

[2] https://news.ycombinator.com/item?id=15460535

That vacuum chamber has two sets of fifty foot tall doors. The hyperloop is an inch-thick welded steel tube. Pinholes are trivially detected by ultrasonic testers. There are no leaks. Even hydrogen and helium would take centuries to diffuse inside. The only problems are sealing the ends, which are a couple meters wide.

The hyperloop as written was very challenging, but the engineering aspects are not. The practical building (grinding the inside, transporting the tubes, etc.) are significant. The financing is insane. The theory is not even particularly hard, much less at the level of a space elevator.

Axial compressors aren't the only compressor design. Rotary vane compressors have no problem with those pressures.

The internal pressure was chosen explicitly because it was easy to maintain with simple, single stage mechanical vacuum pumps (see Figure 13 on page 22). Heck, the first result on Amazon for "vacuum pump" goes down to 5 Pa. https://www.amazon.com/dp/B012CFTYX4/

Compare the Hyperloop to true "vacuum train" designs, which need to run at 1/1,000,000th of an atmosphere to mitigate sonic booms (an alternate way to get around the Kantrowitz limit). This requires multi-stage pumping with mechanical roughing, followed by turbomolecular pumps and cryopumping. That 1000x harder vacuum takes 1000x as much pumping power (not because the differential pressure is meaningfully different, but because you expel 1000x less air per stroke).

Reply on news.ycombinator.com