Musk's Hyperloop math doesn't add up
greatergreaterwashington.org
greatergreaterwashington.org
>That's a pretty significant issue because safe vehicle operation means never getting closer to the vehicle ahead than the distance it will take you to stop. [...] That means that the minimum separation between pods is probably closer to 80 seconds or more.
Two flawed assumptions in here:
1. 0.5 g is the maximum deceleration under normal conditions (see p. 58 in Musk's PDF). Emergency braking would obviously be capable of more. The linear accelerators themselves are spec'd to be capable of 1g, and there's an onboard emergency braking system on top of that.
2. A vehicle that crashes doesn't stop immediately, but continues on its direction of motion due to inertia. That's why the two second rule is plenty safe on highways [1], even though expected vehicle stopping times are six seconds or more [2].
Furthermore, the article is basically an attack on maximum capacity, but that's not a real issue. The system is economically viable if it can get passengers from point A to point B at comfort and cost levels (including recouping investment) that compare favorably with other means of transportation; Musk cites $20 per ticket, which the article doesn't dispute Maximum capacity can always be expanded if the economics work otherwise.
This whole thing reads to me as purest wind.
EDIT: At least on a conventional train, I can ride from somewhere that's not at one of the endpoints to somewhere else that isn't; and, of course, I can go to the bathroom.
I read a post last week that compared the situation to the Shinkansen, which was an expensive political show project until they figured out how to get it to population centers.
If your point is they started off with areas they still needed to work out like connection to population centers, then I think you're on the wrong site: every startup we discuss pivots in one way or another to overcome challenges, that's an accepted part of developing something useful. If you don't try you're never going to get to a point when you can find a solution to problems you don't know exist yet.
If there was a hyperloop between the edges of SF and LA, how many minutes do you think it would take before someone worked out how to get passengers the short distance from the station to the city center?
On your other point: how much incentive is there to get people from SF to Mountain View faster? HUGE incentive, is the answer to that question. What's the answer right now? 101 and 280, or a long slow train ride.
EDIT: Also, a Hyperloop that terminated where the PDF proposes just wouldn't -- and probably shouldn't -- get built.
I'm sure we'll see it in built in China in a few years time then. No doubt it will be plagued with problems, maybe a few deaths. It'll be labeled a white elephant for a few years. But in 25 years time when everyone in China is zipping about at over 1000km/h and people in SF and stuck in traffic shouting about the good old days and refusing to build anything new because good isn't perfect, you'll be lamenting the lack.
But I doubt it. And in any event, the idea that we should build a toy project because it offers some improvements to some parts of a current problem for some small number of people is a perfectly defensible one -- on Hacker News. As an actual spending cash sort of idea, it's fairy tale nonsense. You can see how seriously Musk takes it by how serious the proposal actually is w/r/t nuts and bolts -- not at all.
How can they know where the track would run? You want an actual train schedule already?
If it is unacceptable not to have a bathroom, then cars are also unacceptable. Somehow it works for many people anyway.
I wonder whether existing HSR plans had to meet requirements this stringent.
And pitching this fluff as SF -> LA without tackling the last mile -- perhaps the hardest part -- is all a reasonable person needs to know to know it's just waffle.
Amtrak is obviously not a model of efficiency, but it's certainly a valid data point about the scale of operating costs we're talking about.
The $20 ticket is a myth. The only way anyone pays $20 to ride this thing is if the taxpayers take up the other $140 of slack.
The only point I was actually trying to make is the (I hope) uncontroversial one that operating expenses for a transit system are high.
The overhead of those workers' wages would scale roughly with ridership. Amtrak ridership is up dramatically this year[0], but it is still 20x smaller than for domestic air travel[1]. They employ nearly 5000 people in mechanical and maintenance roles[2], which is a cost largely independent of ridership.
You can argue the case that Musk's cost estimates are overly optimistic, but Amtrak's operating costs are largely relevant.
[0] http://www.sacbee.com/2013/03/01/5227255/amtrak-ridership-up...
[1] http://www.transtats.bts.gov
[2] http://www.amtrak.com/ccurl/533/3/Amtrak-Mechanical-Services...
I think (if it gets built!) we'll see very few humans operating or maintaining the hyperloop.
It's never getting built, though. An insurmountable advantage of HSR is that it can serve population centers along its route, which makes ROW acquisition feasible.
I don't think the car company has much to do with things (you can make an argument that a better HSR infrastructure would help Tesla, not hurt it, by reducing its customers need for the long-haul drives that are the weak point of electric cars), but I do think the Hyperloop is best understood as a political maneuver to complicate CA HSR.
If capacity becomes an issue, you could just add more tracks. If the cost is wrong (and it no doubt is), why is nobody coming up with the real cost so we can compare this properly? Hand waving about 'too low!' is only useful for arguing, not for finding a solution. Are there other ways that the cost could be brought down?
HSR plans get around this by terminating in the middle of major metro areas, which they can do because they get to reuse existing rail lines.
Not necessarily. It depends on demand.
Think of travel time as latency, and volume as bandwidth. If the system is unsaturated, then the 30 minute travel time is great. You show up, get on, and ~30 minutes later you're at your destination.
But if the system is saturated (which seems likely at $20/ticket), then latency becomes more or less irrelevant because you're limited by bandwidth. Sure, once you get on the car, you'll arrive in half an hour. But you will have to wait in line, and that wait will be determined by the volume limitations of the hyperloop. If the hyperloop attempts to serve the same demand as the proposed HSR, a passenger's total travel time will be four times as long.
The solution is to charge more for a ticket. The actual price needs to be calculated in terms of a point where you don't end up with more than 3000 passengers per hour (at which point you break even).
However, even if you do that, it doesn't as effectively solve other problems that the HSR will help with, like total car traffic.
Web-based reservation systems are not hard. Last-minute tickets could be sold at higher price for those that need and can afford them as well.
You can't wave a web browser around as a magic wand to solve these problems.
This is what led me to believe he was talking about standing in line for a place in a capsule.
Why is the wait for security determined by the volume limitations of the Hyperloop? Isn't that a function of demand, security lanes, and space to fill with waiting passengers?
Passing through the checkpoint itself takes ~90 seconds. The other 20-90 minutes is waiting for the queue in front of me to be processed by 1-2 lanes working in parallel. If the airport had a wide enough hallway and a big enough payroll, why couldn't it run 10-20 lanes and cut down the wait by an order of magnitude?
Obviously the federal government doesn't choose to spend money that way, but it could - regardless of the seating capacity of the aircraft. In fact low-bandwidth aircraft make the problem easier because fewer people need to depart at once.
I trust that you know what you're talking about when it comes to security, but I don't see it. Why does Hyperloop's low bandwidth make the security line problem worse than HSR?
The larger problem is: if you can build a hyper loop track at 10% of the cost, you can build an ultra light conventional "train" on the same pylons and with lighter weight,cheaper, track. It would run at maybe 200 or 250pm and hence with reduced capacity. At best Musk's cost claims are a capacity/cost trade off. At worst, they are suspect. I think they are suspect, especially given the extreme tolerances of the tube
For example, I think the discussions of headway at the beginning of this article are very good, and he lays out both problems and potential solutions, albeit ones that make the system less useful.
However, by the time we reach the end, we get crap like "Clearly he's unfamiliar with the Cypress Street Viaduct." Yes, I'm sure Elon Musk has never heard of that, sure. I'm sure he didn't research earthquakes at all when making his statement. Come on, really?
This kind of nonsense weakens the article considerably, because readers will tend to concentrate on the weakest aspects of it, not the strongest.
To mount the argument that engineers and critics should be focusing on rehabilitating Musk's proposal rather than shooting it down, one of the following needs to be plausible:
(a) that door-door transit times on Musk's system will be significantly better than the HSR plan
(b) that construction and operation of Musk's plan will be significantly cheaper than HSR
(c) some X factor makes Musk's plan more attractive
There are compelling arguments against (a) and (b):
(a) because Musk's system can't reuse existing urban rights of way, it terminates far from either city center. It gets a "~30 minute" estimate, which message boards are eager to latch onto, by waving a magic wand and turning Hayward into San Francisco and Sylmar into Downtown LA. If that was a viable strategy, SWA would still be relying on Islip to service NYC.
(b) Musk's estimated tunneling and el track costs are so low that, were they viable, they'd revolutionize all of the rest of urban planning. For instance, Musk proposes to lower the cost of running viaduct by an entire order of magnitude. No comparable tunnel has been built for anything like what Musk proposes his tunneling costs will be. And the expenses he lays out make optimistic assumptions about the routes he'll be able to use, especially through the mountains.
It's gotta handily beat 2hrs 28min for the route people actually care about, OR be substantially cheaper. These aren't nits: they're the entire motivation for the proposal.
Facts, like tweets, have a size limit. "Hyperloop will connect areas relatively near to sf and la" is too long. "Hyperloop will connect sf and la" is a shorter, and therefore better, fact.
But yeah, screw those guys. Much more fun to solve the problems.
Since I lack the expertise to know myself whether the numbers are plausible or will work out, I need to rely on the expertise of others. And one of my ways of evaluating the expertise of others is if they appear to be guided by facts or by an agenda.
Debating motives is a waste of time. If his piece is all motive and no meat, the article should be easy for you to pick apart. So do it.
So, the most useful analysis would be:
1. What factors are preventing HSR from being fast, technically feasible, heavily used, and economically viable in the United States? (More concretely, why does California's HSR suck so much when it's been demonstrated that HSR doesn't necessarily have to suck?) Demographics (e.g. population density and layout)? Political and legal constraints? Terrain? Culture?
2. How does the Hyperloop mitigate or sidestep these problems, due to its superior specifications or design differences from HSR?
3. For additional insight, what would California's HSR have had to do differently to mitigate or bypass these issues, and would it even have been possible?
1) If America cannot understand the potential and purpose of innovative technology, it is doomed to slide into socio-economic, scientific, and technological poverty. The point is that building a slow and expensive HSR is stupid, and people have approved it solely because they are, similarly, too stupid to envision what is possible. The primary point of building the Hyperloop is not to move people physically, but to move them intellectually.
2) the Hyperloop, as designed by Elon, has zero emissions. We should do it for that reason alone, otherwise we are genuinely too stupid to exist, which fact says more about us than any debate. If we do not begin building zero-emissions systems, the debate will not matter because we will be gone.
3) Capacity problems are not an issue. Elon has intelligently anticipated a diminishing requirement to physically move people because of the improvement of telepresence technology. If you still need to travel as much as you did ten years ago, you are an idiot. This trend will only increase. Elon is too savvy to mention this explicitly, but he no doubt understands that the need to move people physically is declining. The HSR will be mostly empty if it opens in 2024 with the capacity required in 2014.
4) Our entrenched institutions have failed us, not because of anything specific that they have done wrong, but simply because they are entrenched. We should build the Hyperloop for no other reason than a little revolution being a good thing, now and again.
For people from the east coast, imagine a train system that claimed to connect New York and DC, but actually terminated in Manville, NJ and Bowie, MD.
EG: They fly you to Paris, not CDG or Orly, but to Paris Beauvais.
Where the author really gets it wrong though is in saying "Maybe he can resolve that by using larger pods. But of course, a larger pod will weigh more. And that will probably mean using stronger steel for the tubes, which means that the cost will go up". That would be true, but if you're happy to resolve the issue with larger pods then you should be equally happy to resolve it with groups of smaller pods travelling at smaller separations. A 737-800 seats around 150 people. Hyperloop passenger capsules seat 30. So no one should have any conceptual issue with running 5 hyperloop capsules in convoy with separation of a couple of seconds. This 'couple of seconds' still means almost 1km, so they're certainly not sharing the same segments of tube, so you don't need to scale the tube design at all. It's still dirt cheap, and you have plenty of capacity, and it's still MUCH safer than a plane because if one capsule has a fault the others following can still likely stop in time (given they're all decelerating at roughly the same speed).
I mean, simplifying of course, if the Hyperloop can carry ~3,000 from A to B in 30 minutes, and the High Speed Rail can carry ~12,000 from A to B in 2 hours, which is better? In this scenario wouldn't they be equal?
Probably not a concern for the given case - but if you're comparing transport options in an abstract sense it's something to be aware of.
I can't see a plan involving the idea that in 2029 Tesla would be so dependent on Californian sales rather than global, that it would seek to stop a local rail network from progressing. Especially a plan so convoluted as to involve actually taking time out from running a rocketship business, car company and solar power company, in order to design a rail network that according to the theory, you do not want built.
He says he was pulling all-nighters to get it ready for the press, and also said “I wish I had not mentioned it, I still have to run SpaceX and Tesla, and it’s fucking hard.”, which seems to indicate that it was squeezing his time.
Unless he has quite a lot of free time, which it doesn't sound like, I can't see how this could have taken no time away from Tesla, especially given what he has stated. It might not have taken that much, but it will have taken some.
All I am saying is that it seems likely that it must have taken some time out from his other activities, especially if he is talking about pulling all nighters just to work on the thing.
This seems so patently obvious that I am not entirely sure why you would even contest the point, let alone then resort to rhetoric above sound argument to try and win it.
The question of his motives --- which is a sideshow --- would be easier to dispense with if he had actually invested in the proposal.
He has to design the thing to then write a paper on the design. You are saying that he took time out to do that as though it is not what I am saying, however it is in complete agreement with what I stated in the first instance.
All I ever said was that he took time out to design the thing, and the paper on the design is obviously the public part of designing the thing.
You keep coming up with this build nonsense, which I never implied at any point and which you know is a non-point as anyone who has been following this in the slightest, obviously knows that it hasn't been built yet.
And he has invested in the thing, he has invested the time it took to design the thing to the level required to then write the published paper.