Monday, August 12, 2013

Hyperloop proposal: Bad joke or attempt to sabotage California HSR project?

Was Elon Musk's s mega-announcement really just a last-ditch attempt to sabotage the California High Speed Rail (HSR) project, rather than a serious proposal to revolution travel? Something smells very fishy, so let's take a look....


By now you've probably heard about Elon Musk's widely publicized proposal to build a tube transit system that can get you from LA to San Francisco in 35 minutes.

I was excited to hear about the proposal, as there had been some hype attached to it.  Elon Musk is a serious guy - founder of Tesla, SpaceX and Pay-Pal - so when he says he has something big, it makes sense to listen.

The headlines note the following facts:

  • SF to LA in 35 minutes
  • Cost under $6 billion
  • Something that could be built within the next decade

Fantastic right? The future is here!

Problem is, taking a look at the documents that came with the announcement, it seems to be a fantastic joke. Those claims do not appear to be true - his own proposal doesn't even get close to supporting them.

Now before I get into some details on why the proposal is so wrong, let me ask a question:

Was the headline an attempt to derail the California High Speed Rail project instead of an actual proposal for a real project?

Sadly, we live in a world where people read headlines...and then stop. We also live with a media that isn't so great at fact-checking, and letting us know when that scandalous headline was actually not true.

Remember "climategate", when some hackers discovered that the worlds climate scientists had conspired to cook the books, and climate change was fake? That made quite the headlines. When a few months later it turned out no books were actually cooked....well, not so many headlines. For years people refereed to the incident, based on the initial headlines and not the fact-checking.

Or try this; ask around. Were WOMDs found in Iraq? I think you'll find a whole lot of "yes" or "maybe" or "I think so" rather than what happened when the headlines were gone, and none were found. Funny how the memory works sometimes.

The point is, headlines can have a hell of an effect.

What I am wondering, was the announcement of this project the transportation equivalent of "Have you stopped beating your wife?"

That is, once people see the bullet points above, the immediate reaction is  STOP THE HIGH SPEED RAIL PROJECT! One CNET article even noted that the proposal came just in time to save us from sinking money into HSR!

Why are we building something that is more expensive, slower and will take more time? That makes no sense! It's madness! Damn the government!

.....except of course, what if the proposal is NOT cheaper, does NOT save as much time, and will NOT be faster to build?

Should we trust that man who claimed that California building the world's slowest bullet train (false) and the world's most expensive rail line (also false) as his inspiration? 



Taking a closer look at the proposal, I am going to ignore the whole technical side of things - the pods, the power source, the air pressure....the guy is an engineer (I am not), so I will assume all that stuff is right. After all, that's his area of expertise. He builds cars and spaceships, I assume he can build a pod.

We'll take a look at the other side of things, the infrastructure, political and budgeting side. The side he doesn't have experience in.

This is the PDF

Claim one: The project links San Francisco to LA.

It's the most simple claim of them all. Surprisingly, it's not true.
The system consists of capsules that travel between Los Angeles, California and San Francisco, California. The total trip time is approximately half an hour, with capsules departing as often as every 30 seconds from each terminal and carrying 28 people each
What do the included maps show?

For LA, we get this:

 photo loop01_zps502c9fa6.jpg

Look closely. The line terminates nowhere near LA, but in Sylmar. By conventional rail, you've still got over an hour to go on the Antelope Valley Metrolink Line to reach the city. 

What of the other end, in San Francisco?

The smaller map seems to indicate this end does actually serve SF...

 photo loop02_zpsd21888b5.jpg

With one tiny little problem.

The cost analysis conveniently forget one little detail: The SF bay. A new bridge? A new tunnel? How do you get across it! Who knows - the project team certainly doesn't.

That cost is far from trivial. The brand new eastern span of the Bay Bridge - ie, only half the bay - has now exceeded $6 billion in costs.

How about the Transbay Tube, what BART uses to cross the sea? Built in 1970, it cost the equivalent of over $1 billion today.


So now we're looking at a a system that doesn't actually reach LA, and while the map suggests it reaches SF, the costs don't include it....


And that's not all. Stations are sort of important. The document provides us no indication of the size a system like this needs, but if it were to be built in downtown SF or downtown LA, a good amount of land would be needed just for your basic station services. If you have a pod leaving every 30 seconds, you also need quite the stacking area.

High Speed rail will be pulling into the under-construction Transbay Transit Center - which has a price tag of $4 billion.

Hyperloop? No costs included in proposal. Woops.


Claim two: The project can be built for $6 billion

Within the next few days, you will surely see some reputable websites destroy the pricing estimates for the viaducts and the tunnel as being way too low. For example, the "silver bullet" seems to be that the thing will be built on viaducts. Yeah, well, that was the plan with HSR as well. That's why the cost blew up. Viaducts don't come cheap.

As I noted above, the project also doesn't even attempt to price the connection into LA or SF.

The thing about that.....that's sort of what matters. That's where the high costs are. The big breakthrough of this project is where they claim they save money:

The key advantages of a tube vs. a railway track are that it can be built above the ground on pylons and it can be built in prefabricated sections that are dropped in place and joined with an orbital seam welder. By building it on pylons, you can almost entirely avoid the need to buy land by following alongside the mostly very straight California Interstate 5 highway, with only minor deviations when the highway makes a sharp turn


Ok, Building in the Central Valley is cheap. We know that. That's why the HSR project will be built there first. It's flat, there's plenty of land, and the ROW is relatively easy to obtain.

Amusingly enough, the California HSR budget for the Central Valley is under $10 billion. Ie, in the same ball-park as this proposal.  The reason the HSR project is going to cost $60 billion is because it has to face an uncomfortable truth; actually getting to LA and SF is expensive. Very expensive. That's where there's no free land. That's where you have years of property acquisition. In the shorter term, the plan for HSR is to simply share existing tracks, which the Hyperloop can't do.

So either the budget explodes, or the project doesn't actually serve the main cities. You can't have it both ways.

Is it a big deal?

Claim three: The project gets you from Sf to LA in 35 minutes

What's the big deal if the project terminates in Sylmar and, say, Dublin? Still, 35 minutes is remarkable right? HSR doesn't even come close. 700 miles per hour? WOW!

One little problem...

HSR between downtown LA and downtown SF: 2 hours, 28 minutes

Hyperloop trip between downtown LA and downtown SF:
1 hour from LA to Sylmar via Metrolink
20 minute transfer
35 minutes to Dublin
20 minute transfer
1 hour 10 minutes from Dublin to SF via BART

Total: 3 hours 25 minutes

An entire hour more than traditional HSR! It turns out that stopping at the edge of the metropolitan area, where nobody is actually going, sort of kills your time advantage. You see the same when comparing rail vs air. Sure, planes fly at 550mph....but the airports aren't your destination.

So the big time advantage? Suddenly gone. Unless you pour tens of billions into reaching the downtown destinations. You can't have it both ways.

Claim four: Land isn't an issue

Isn't the hypothetical world great? It's amazing how assumptions always go your way!
Even when the Hyperloop path deviates from the highway, it will cause minimal disruption to farmland roughly comparable to a tree or telephone pole, which farmers deal with all the time. A ground based high speed rail system by comparison needs up to a 100 ft wide swath of dedicated land to build up foundations for both directions, forcing people to travel for several miles just to get to the other side of their property. It is also noisy, with nothing to contain the sound, and needs unsightly protective fencing to prevent animals, people or vehicles from getting on to the track. Risk of derailment is also not to be taken lightly, as demonstrated by several recent fatal train accidents

Assumption 1: HSR doesn't use viaducts so people are forced to travel several miles to cross their property. We know that's false. That's what made HSR pricey. Tunnels and viaducts.

Assumption 2; Property owners don't care if you build your enormous aerial system through their property. We'll look at this in a second.

Assumption 3: Building a support structure every 100 feet does not disrupt farmland. They can't even be serious here.

For assumption 2, the project team made the biggest mistake in the world of infrastructure:

They ignored the NIMBY. (not in my backyard).

Why could we send a man to the moon in a decade, but California has taken decades to not even start building their HSR system?

Because there are no NIMBYs on the moon. 

The NIMBY is not a rational creature. The NIMBY does not want to help you. The NIMBY doesn't just want you off their property, they want you out of their city.

To assume that people will willingly grant your line of support columns an easement is an exercise in the absurd.  Worse is the assumption that an aerial structure is popular.

Remember Cape Wind? It was a Massachusetts proposal to build an off-shore wind farm. Far away from homes and property, way out in the ocean. It got held up for years and years and years by lawsuit after lawsuit.

You know what the problem was? Views. Aesthetics. People didn't want to look at these things way out in the ocean.

People love their views. Farmers love their views. To assume that an aerial structure is your golden ticket out of years in the courtroom is plain idiocy.

Cape Wind eventually won in court. It still hasn't been built.

Why will HSR cost $60 billion? Because their plan actually acknowledges the NIMBY, the court fees, and the settlement sums.

And again, all this assume they don't even try to actually get into a city. Let's not even get started on the Grapevine. Most of the land there is owned by one giant company - Tejon Ranch. No, they do not want your project running through their land. Just like everybody else.


Claim (assumption?) five: Politics aren't an issue

Let's pretend for a second this is a serious proposal. let's pretend we have a $50 billion project to link Oakland (not quite SF) with North Hollywood (not quite LA). Let's pretend the travel time is equivalent to HSR. Now we have a project that takes the same amount of time, but is cheaper.

What's not to love?

One major problem: It's a point to point system. LA to SF. No other stops. Why is this a problem?

Politics. HSR was initially funded via proposition 1a. If this project were real, it would be funded publicly as well - Musk does't propose to spend a dime. Take a look at the map of support, green being yes of course. It passed with 52.62% support.

 photo loop03_zps3d01178d.jpg
Notice something? The counties which votes yes have a stop on the HSR line. Those not served, voted against it. Essentially, "if I can't have it, nobody can!"

The Hyperloop proposal ignores the Central Valley. Goodbye Fresno, Modesto, and Bakersfield votes. You know how HSR loops over LA to serve Palmdale? Goodbye desert vote. You know how HSR continues to Anaheim? Goodbye Disney. How about San Jose? San Diego? Sacramento?

There's a reason HSR stops at every major city, and it isn't just because it makes sense for travel planning: It's politics. You want the vote, you need to serve the people. Good luck skipping most of them.

Claim six: It can be built quickly, within a decade

Perhaps the viaduct system, pods, and stations etc. can be assembled quickly. Not an issue.

What is, is everything that comes before it. For one, it's a new technology. You need to build a test system, and every safety agency needs to get their grubby little hands on it. Add a couple of years. Then you need to get your financing in order. All private money? Yeah right. Add a five more years for public money - triple that depending who's controlling the purse strings. And then you have your land battle. Your property acquisitions. Your court cases. Your injunctions.

Elon Musk, you aren't building some space toys in the middle of the desert. You want to build over, across, and next to people's homes. This isn't the 1960's anymore, you can't urban renewal your way to your goal. HSR began serious planning in the 1980's. It's using proven, off-the-shelf technology. You really think you won't hit the same road-blocks?



So what's going on here? This proposal is so off base it feels as if it was put together during a weekend of drinking.

Is Elon Musk broadcasting an incredible display of optimistic naivete, or is there a much more sinister goal here?

If you follow transit projects, you'll know that a VERY common strategy of opponents is to say they support your idea, just not your implementation.

"Oh we LOVE the idea of transit between x and y. We just don't like that it's expensive and unsightly light rail, we will TOTALLY support the BRT project that is cheaper, can be built faster, and is better for riders!'   .....and will not receive popular support, thus killing the project at the ballot box

You see the opposite as well in areas that are more pro-transit ."Oh we LOVE the idea of transit between x and y. We just don't like that it's ground-running light-rail , we will TOTALLY support underground subway that is safer and faster!"  ....and will be so expensive the project will be cancelled before the first shovel hits the ground

The goal isn't to build a better system. It's to destroy the process by presenting a false choice.

To me, this seems to be the mother of all false choices.

A remarkably attractive headline by a wildly respected man. What better way to pull support from HSR than by creating an alternative proposal that is better in every way? That's the thing about fiction though, it can be anything you want it to be.

Will the people fall for it? If they do, I guess it's time to put down a deposit for a new Tesla.

74 comments:

  1. Don't forget there's security theater proposed for the system. So add an hour to the door-to-door run time. Suddenly it's slower than HSR. No wonder the capacity is so low - it can't get that much ridership!

    ReplyDelete
    Replies
    1. I very much enjoyed the throwaway TSA line.

      Delete
  2. Don't worry, though-- you can take your car with you on the capsule, so you don't have to care about that silly transit stuff to downtown! *barf*

    ReplyDelete
    Replies
    1. I hadn't even thought of that. Although as someone who runs a car company, it makes sense that local transit isn't part of his math.

      Delete
    2. This is sort of late and I really like your article, but there is one part that I think is not entirely accurate, which is that the time estimations for the LA to SF transport via the tube takes about 3.5 hours. I live in Dublin and it takes about 30-40 mins to get to SF via Bart. That makes me wonder if what you said about Sylmar to LA trip takes that long. I'll assume that it realistically takes 30-40 mins. (granted you estimated both at an hour) The transfer times are realistically probably only 10 minutes long at the most unless you miss the train. (assuming the train runs by every 10 minutes which I'm under the impression that most do) Adding all of it up you could probably make the whole trip in 2 hours average. I might be a little biased for the tube so correct me anywhere if I'm wrong.

      Delete
    3. @Alex, I used the times quoted by google maps, which partners with the transit agencies.

      Delete
  3. Setting aside all the other comments' concerns: People are awful starry-eyed about this just because it's Elon Musk. We've seen many, MANY novel mass transit ideas (even ones that look like good ideas on paper) that run into massive technical problems in the development phase. Evacuated tube transit isn't anything remotely like a brand new idea - nobody's developed it because it's been darn-near certain to turn out to be ridiculously expensive and hard to work with.

    ReplyDelete
    Replies
    1. It was pretty amazing how many websites bought the proposal at face value. You see that a lot in the tech sector though. There's not much investigating, so much of it is just repeating press releases.

      Delete
    2. That's been annoying me as well. I just don't think there's a huge overlap between techies and transportation planners.

      Delete
  4. Amir0x is a fine fellow.

    ReplyDelete
  5. I noticed a typo in your article:

    1 hour 10 minutes from Dublin to LA via BART

    I think you meant Dublin to SF via BART.

    Great article!

    - Ben

    ReplyDelete
  6. I'm pretty sure Sylmar is technically part of Los Angeles. At least, it's within a mile or two of places that are. Though of course the main point is right - no one who's traveling to Los Angeles is going there, unless they happen to have relatives there.

    And I must say that I was certain his proposal was just this until I saw the documents - the idea that you can slow down and use some of the existing freeway infrastructure even with its bends starts to make it a little plausible that it can be done with much less land acquisition than rail.

    But I still don't see how $600 million for tunnels and $1 billion for land acquisition can possibly be enough unless you get extremely lucky at every single turn. And it can give just enough uncertainty for a few more legislators to want to pull the plug on HSR.

    ReplyDelete
    Replies
    1. It looks like you're right about Sylmar. LA has really funky boundaries. Looks like the Metrolink station is right on the border of the tiny town of San Fernando - not part of LA. But yeah, the point still stands, that's not the destination for most people, or what anyone imagines when they hear "LA".

      I guarantee some central valley anti-HSR legislators will be touting this thing for years as "what we should have done".

      Delete
    2. And since HSR goes to downtown and will easily connect with service to San Diego and other locations, Sylmar looks like a major distortion.

      I took a little closer look at what it would cost to bore a twin BART tube tunnel under the SF bay. I found an interesting 2002 study that puts the cost at $2-3 billion. Musk (nearly complete absence of details) seems to put a cost at 3.5 miles at $50 million = $175 million. A factor of 14X difference. Yet the width of his tube is implied to be either 8 feet or 10 feet depending on the option.

      Check on the PDF on the line saying "Cost Report 6-20-02" right here:

      http://www.mtc.ca.gov/library/bay_crossing/


      And it doesn't look to me like the size of the tubes is going to save that much. The design study I looked at includes a 15 foot wide space for a BART train, a gallery, drainage and electrical making the bore 30 feet in diameter. Musk of course doesn't talk about a gallery and drainage for his Bay tunnel - cause he doesn't give any details - but it is hard to imagine his tube without those elements.

      Nice writeup James! Thanks.

      Jeff L.

      Delete
    3. "I still don't see how $600 million for tunnels and $1 billion for land acquisition can possibly "

      With regard to land and ROW, the evidence is in Hyperloop's favor if the elevated path ROW model is valid. We *know* the cost of HV transmission lines from the like of AEP and others is $1 to $4 million per mile [1], to include land, ROW (and the transmission which we don't care about here). Thus over the 350 mile Hyperloop path land/ROW cost would be, at most, $350 to $1400 million. One can argue about the cost of the elevated path material, but I've seen no good argument about why the land costs should not be any more than that of elevated HV cable, and likely less.

      [1] http://www.aep.com/about/transmission/docs/transmission-facts.pdf

      As for the tubes, the proposal calls for 2.3 m ID, 22 mm wall steel pipe, which gives me 1.3 metric ton per meter. With steel at $650/ton, 350 miles (564 km) of tube is $480 million, each way, or $960 million both ways.

      Delete
    4. > Musk of course doesn't talk about a gallery and drainage for his Bay tunnel - cause he
      > doesn't give any details - but it is hard to imagine his tube without those elements.
      Why would you need a gallery or drainage for a solid steel evacuated tube? The reason for the elements in a BART tunnel is that that train itself is not in a sealed environment, but in the case of Musk's proposal, being underground is (almost) business as usual.

      Delete
  7. I for one was very happy to see Elon's proposal. I just can't see spending $60 Billion on a "me too" HSR system using technology purchased mostly from other countries. I don't mind funding a system which is faster and cheaper, even if it is more expensive and slower than is being advertised. I will take 2X the speed of HSR and half the cost. We have to think differently. Of course there are lots of challenges. We never would have built the Mars Rover, space shuttle, or planted a flag on the moon without challenges. In the end we can be a leader in technology and export it, rather than send our tax dollars overseas for old technology.

    ReplyDelete
    Replies
    1. Building the Mars Rover, space shuttle and planting a flag on the moon required overcoming engineering challenges. Not political challenges.

      There is a big, big difference.

      Political challenges don't just magically yield to the will of engineers, even Elon Musk.

      Delete
    2. And how do you suppose to overcome political challenges?

      Has anyone really tried? How's Obama doing with 'change' and the economy? It takes an Elon Musk to take that risk.

      He understands the political challenges quite well, the whole country needs a change in attitude and this a gigantic start.

      Delete
    3. So what are you advocating? A military coup where Elon Musk becomes dictator? No thanks.

      Delete
  8. Musk's position backed by his reputation could put a huge damper on the political will to continue with HSR. It would be awesome to use this opportunity to get him in a live debate on this issue. Somebody make it happen!!!! -Davey

    ReplyDelete
  9. If ROW is the main issue, and speed is not, why not take this out to sea? You only have one problem there, keeping things safe, but that's an engineering problem and we can solve that.

    Author should also keep in mind that people are more willing to cooperate with ROW issues when the result is very beneficial to them. Certainly some would be more stubborn, but if the idea is generally excepted then everyone is under pressure to support it instead of derailing it.

    My 2 cents.

    ReplyDelete
    Replies
    1. It's politics again. California has a fiercely guarded cost. Very protective of the ocean environment.

      Delete
    2. ??? cost = coast?

      Delete
    3. Also, offshore structures are definitely, definitely a lot more expensive than onshore structures.
      Comparing the investment cost of offshore wind power to onshore wind power will give you a good indication (rule of thumb is a factor 3, but it depends on conditions). The environment is corrosive, you're dealing with huge wave loads during storms, your pylons become very long if they have to reach down to the seabed. Maintenance is an issue (wind farms in the North Sea cannot be accessed for maintenance for longer periods during the stormy winter season)
      As you mentioned, safety would be an issue. How do you evacuate people from an airtight tube suspended over the sea, let's say because the two tons of lithium batteries in their pod went into thermal runaway?

      Delete
  10. A few things in the proposal really got to me:

    (1) He proposes I-5 as the corridor. However, north of Kettleman City, I-5 is actually fairly hilly. Plus it misses obvious stations in Fresno and Bakersfield. CA-99, on the other hand, is flat. And it passes through those cities.

    (2) So, what if your capsule, for whatever reason, stops in the tunnel and CANNOT start up again? How far are you from the nearest exit? 1 mile? 10 miles? 100 miles? I find it difficult to believe that there would be exits every couple hundred feet, especially if you're trying to maintain a near-vacuum. Can you even exit the capsule inside the tunnel? The proposal seems to indicate that the capsule exits are on the side. That would indicate to me that, even if the tunnel was brought back to atmospheric pressure, the only way out would be to disassemble the capsule from either the front or the back.

    (3) If an HSR train is stopped, or if a section of track is unusable (for maintenance, etc), crossover tracks would permit trains to get by a closed section, albeit with some delays, train slowing, etc. But at least the system would still be usable. If a section of HyperLoop is closed, the entire system shuts down; there's no way around it, unless you build a second, redundant loop (at twice the price)

    ReplyDelete
    Replies
    1. 1. The proposal shows a potential option for spurs to Fresno and Sacramento.

      2. They cover this in the proposal. The capsules would have a set of wheels for emergencies. It would be able to drive to the nearest emergency exit (I don't recall if they propose a distance for those) There is sufficient space in the tunnel for capsule doors to open.

      3. Seems true. Though this is true of some other types of trains (such as maglev) as well.

      Delete
  11. Great post! There's is a small mistake, however. BART from Dublin to Downtown SF (Embarcadero/Montgomery/Powell) takes about 50-55 minutes not 1 hr 10 mins as stated in the post.

    ReplyDelete
    Replies
    1. I went based off the google maps estimate.

      Delete
    2. BART.gov has the official schedule and it shows travel times of just under 50 minutes from the Dublin/Pleasanton BART station to the Embarcadero BART station.
      http://bart.gov/schedules/quickplanner.aspx?orig=DUBL&addr1=&dest=EMBR&addr2=&type=departure&date=8%2F14%2F2013&time=9%3A00+AM

      Delete
  12. I like the idea of course, who wouldn't, faster, cheaper, safer travel. Of course all of these are claims, assumptions may be inaccurate but that is not to deter an idea which has merit. So what if it isn't a completely new idea, that is not detrement to an idea especially when the guy isn't claiming it completely as his own - Edison technically didn't invent the light bulb but he takes a lot of credit becuase he made it mainstream.
    The way forward is for someone to take the idea and build prototypes and test on a smaller scale. Perhaps the economics don't work in some of the most expensive land areas in the country but with other city pairs the economics may be more favourable.
    Debate on these things is positive but when things have scientific merit they shouldn't be vetoed purely on economic concerns. The cost of new technology comes down with adoption...

    ReplyDelete
  13. Very picky. Musk has said this is a general outline, simply stating what current technologies can be used in the creation of this innovative transport solution, rather than putting together a detailed blueprint to follow.

    Which leads me to your first point. Very unlikely that Musk would take ten minutes tracing a map for an exact route, when there aren't even plans for a prototype. Also, why would there need to be a bridge? Just build along route 101.

    2. Yes, probably, but it will still be at least half the price of HSR.

    3. Read point 1.

    4. May have a point.

    5. Possible.

    6. Most of the time in the research phase will come from actually building a working prototype, not researching new technology.

    Okay, and after all that, I'd like to point something ELSE out. HSR2 is being built. Hyperloop WON'T replace HSR2, simply because. Instead, Hyperloop will be something that the government can then consider for future projects, for example, linking NY and Chicago, or for introducing elsewhere in the world. A good example would be Paris-Berlin, London-Edinburgh, Madrid-Paris, Guangzhou-Shanghai, Brisbane-Sydney, Adelaide-Sydney, etc. Simply because it might not fit this scenario doesn't mean it won't be useful elsewhere.

    Musk would have used SF-LA simply because it is the best and closest scenario to compare, as HSR2 is being built.

    Don't nitpick on details, focus on the bigger picture.

    ReplyDelete
    Replies
    1. "Don't nitpick on details, focus on the bigger picture."

      Cost overruns in innovative engineering projects generally arise because consideration of details tends to make expose the original big-picture scenario as being technologically naive and far too overly optimistic.

      Before ditching HS2 for new technology which doesn't even exist in prototype form (!) because of claims it's 10x cheaper and just as quick to build as the existing project, the details need to be nitpicked to the point where the actual costs can be more accurately estimated.

      Delete
    2. My seat of the pants estimate is that he underestimates the cost of the infrastructure by at least a factor of 10. I have no idea how much he underestimates the tech costs of an untested technology- there can be a lot of gotchas. But I would think a more realistic cost estimate is 100 billion. Not insignificant.

      Delete
    3. Ah, I remember hearing all of these same arguments about SpaceX.... and then about Tesla Motors... about how naive Musk is, thinking he can get into space and beat NASA at their own game at orders-of-magnitude less cost, and then about how naive he is, to thinking *any* new car company stands a chance of reaching mass production and competing with the big car companies, let alone building a radical new type of electric car running on laptop batteries.

      As it stands, NASA's Ares rocket was a total failure, SpaceX is now on their second generation and has 16 ISS launch contracts as well as private satellite launch contracts, and Tesla has produced over 20,000 Model S cars (more than all of my car, a Honda Insight, were produced globally in their entire run) and has gotten some of the best reviews from car magazines ever given.

      To reiterate what was written down, the reason the suggested costs are so much lower isn't magic. It's about size and weight. It's only 7 feet across (so, for example, think tunnels: a 7' tunnel has only 12% as much rock to bore out as a 20' tunnel). Its "trains" are not many thousands, even tens of thousands of tons, but only 60 tons fully loaded. So it makes things a *lot* easier to build.

      Honestly, I think the *engineering* solution is the right one. I think the *particular details* of the proposed layout are wrong. As it stands, it's more a replacement for a couple airports than for HSR. I think he wanted to keep the cost around something that a private company or wealthy individual could afford instead of something that would require state/national funding.

      Delete
    4. Certainly. I won't nitpick.

      I'll just point out that "Hyperloop" will cost, at a minimum, TEN TIMES what Elon Musk thinks it will cost, and will therefore be more expensive than HSR. It will also have lower capacity.

      I have analyzed the business model for SpaceX and Tesla and invested in Tesla (SpaceX, I figured out the business model a bit late to invest.) They are sensible business models.

      The Hyperloop isn't.

      The low suggested costs are simply pure magical thinking. The lateral forces are higher than for HSR (meaning, stronger pylons) and the weight of the tubes is higher than the weight of trains & tracks (this is needed to handle air pressure differences).

      Don't believe the hype.

      Delete
  14. Thank you, thank you, thank you.

    I was waiting for a no-holds-barred takedown of this idiotic idea.

    Like the author said, either Elon Musk, despite his brilliance, is breathtakingly ignorant about politics and public infrastructure, or he would like to destroy HSR so that he can sell more electric cars.

    So he's either shockingly naive, or he's devious and the opposite of public-spirited.

    Maybe it's time for a damper on some of the breathless positive publicity about him.

    I also think that we're running into the limits of the Silicon Valley/techie/utopianist mindset.

    Politics: there is most assuredly NOT an app for that. We are no closer to a solution for it than the Incans or Mayans or ancient Romans were.

    ReplyDelete
  15. Another thing... it appears that he severely underestimates the obstacle that the Grapevine presents. Those mountains weren't just placed there by magic; it's a very active seismic zone. And he appears to want to tunnel directly through the San Andreas Fault, near the epicenter of the largest California earthquake in recorded history: Fort Tejon, 1857.

    ReplyDelete
    Replies
    1. I fully expect some other websites/blogs will dig into the tunnel stuff better than I can.

      My go-to blog for tunnel pricing analysis is
      http://pedestrianobservations.wordpress.com/

      Delete
    2. On MeFi, people actually figured that the tunnel cost estimate was bullshit (link).

      Delete
    3. I think the seismic issue goes nowhere. Seismic issues affect all rail construction, not just Musk's proposal, and Japan builds high speed rail over fault lines all over the place. And towers often deal better with seismic issues than surfaces on the ground, the tower acts as a damper. Of course, anything can be built wrong and fail.

      As for that "bullshit" link, I really don't get their logic. Because the pylons are denser, that means they're a lot more expensive? Sorry, but things don't scale remotely linearly like that. The further apart they are, they stronger they have to be, and the stronger your connecting span has to be. There's a reason that suspended infrastructure isn't built with pillars great distances apart unless necessary (canyon, deep water, etc). And using figures from power transmission, I can't even see how that's remotely comparable to bridging. Giant towers with insulation requirements with lower weight, non-rigid spans. Not at all what's being talked about here. And comparing tunnelling to the Delta Water tunnels, really? The delta water tunnels would have over *22 times* the cross sectional area. And be built in an environmentally sensitive area. I mean, come on people...

      The reality is, his figures are consistent with actual construction cost figures. I've run the numbers myself, and while I was initially skeptical, it seems to check out. Primarily, the reasons his system is cheap is because it's light and small (at least, by rail standards - dozens of tons instead of tens of thousands of tons). Where I don't think he's been realistic is in accounting for "human factors" in resisting the line construction, and in treating his stations like airports rather than train stations and passing the buck to localities to connect them to the people.

      Delete
    4. The reality is, his figures are TOTALLY INCONSISTENT with actual construction cost figures. I don't believe that you've run the numbers correctly. Have you figured out how heavy the tubes have to be to maintain the near-vacuum, or what the lateral forces on them are? Mmmm.

      Power transmission can have extremely sharp vertical rises and falls and has to carry very lightweight cables. And it's still ferociously expensive when you are significant distances off the ground. Power transmission routes hug the ground.

      "Hyperloop" would REQUIRE shallow vertical curves, meaning that it would have to be huge distances off the ground in many places. It'll cost a fortune, just like any elevated train does.

      That's why you can't compare powerline costs to train costs: vertical curves.

      Civil engineering dominates the price. Musk has grossly wrong civils cost estimates.

      Delete
  16. Note- Musk corrected his statement about it being the "slowest" and "most expensive per mile" after the head of the HSR called him to correct him.

    Correctly put: it is "one of" the slowest and "one of" the most expensive per mile.

    It's basically a disaster that everyone should be ashamed of in California. I voted for it and I know cringe when I imagine how high the ticket prices will have to be to service that massive debt load.

    ReplyDelete
    Replies
    1. Transportation isn't supposed to turn a profit. Highways certainly don't The gas tax doesn't even cover 50% of highway maintenance let alone capital construction. CAHSR should have no problem cover it's operating costs which is more than can be said for highways and other public transit systems.

      Delete
    2. "One of the slowest" is still false. The CAHSR proposal is for average end-to-end speeds (LA-SF) higher than anywhere else in the world outside China. Certain critics have been complaining the speed target is if anything too ambitious.

      Delete
  17. Others have covered a lot of the details I wanted to bring up, so I'll leave it at this: take the cost all the to $60 billion (for the sake of argument) and connect it downtown-to-downtown. Now guess what? You have something 5x as fast which is also safer and more eco-friendly. And less land acquisition, even if it is substantially more thank Musk laid out.

    Of course, that doesn't mean it has a prayer of happening, for that particular city pair at leas. The nature of politics and government.

    ReplyDelete
    Replies
    1. edit:

      "take cost all the *way* to"

      "substantially more *than* Musk laid out"

      "particular city pair at *least*"

      Delete
    2. If you want to make it downtown-to-downtown, a decent cost estimate for the Hyperloop would be $600 billion. James was being generous.

      Physicists have noted that there's an unsolvable heat-dissipation problem in the design, too. It really is a joke.

      Delete
  18. 1) Musk has put this out there as an early-stage idea, not a polished proposal, 2) others will pick this up and refine it, unless 3) those with vested interests in the CAHSR project are successful in dismissing this idea. So, author, what is your stake in CAHSR?

    ReplyDelete
    Replies
    1. Let me refine it: the cost is going to be at least ten times what Musk says it is, just by looking at how much viaducts actually cost versus how much Musk fantasizes that they cost. Without the LA and SF connectors, mind you.

      So, anonymous commenter, what is your stake in Tesla Motors?

      Delete
    2. (I'm not the same Anonymous, and no I do not invest in Tesla)

      Let's add the cost of connecting downtown LA and SF:
      $60 (10 times the cost of proposal) + $3 (connecting SF per post above) + 10 (connecting LA - just a guess) = ~$73 billion

      HSR = $68 billion

      I agree that the proposal math is funny, but can you really argue that something that *might* cost roughly the same as the HSR and...
      - Travels three times as fast (618 mph vs 164 mph)
      - Saves more energy (I would doubt Elon here too, but his Model S is basically an amazing feat of engineering)
      - Has less noise pollution
      ...is a bad idea?

      With starts ups like Surfair coming around, cutting down the travel time between SF and LA to 1.5 hours, it's obvious that we need a more innovative and long lasting solution here.




      Delete
    3. "Just a guess."

      The Hyperloop system has shit capacity, and any attempt to raise its capacity to HSR levels would blow out costs because of the need for complex junctions near the termini and more expensive stations. Being able to support short headways on running track is not the same as being able to get into terminals.

      All would have to be done on brand-new urban viaducts and in tunnels.

      And the speed gain versus HSR would still be not much because of the security theater.

      Delete
    4. Its capacity is 3 times that of the SF/LA air link, but one third that HSR, hence it's an intermediary proposal. The proposed cost is less than one tenth of HSR, so scaling up is not an issue, you could build whole new systems if the junctions proved to be a problem (which I really doubt, plenty of room in the valley for a big arc). Honestly, it should have been proposed bigger - more destinations and capacity; it would have made it more credible while still having far lower costs. Would also have made the problem of line outages less extreme, if there's multiple lines.

      The inner-city connection problem is a real issue, but this is not.

      Delete
    5. The key problem is that the proposed cost is fictional.

      Muskl never trained as a civil engineer, and it shows.

      Delete
  19. I feel you have some very strong points, and I also have some very big disagreements. If I may?

    First off, I think the conspiratorial aspects take strongly away from your (actually good) points. Since when is Musk famous for hating the environment or being a pawn of big industry? He's always been an upstart, proposing radical concepts and following through on them. He's less likely to follow through on this, of course, all his resources are tied up in other ventures. But I know the feeling. I too am geek and always coming up with ideas. And if it sticks in your mind, you start doing the engineering, running the calculations, doing the cost estimates... the only difference is he has platform and a lot more resources!

    Probably your best point is the "from" and "to". And I think you can easily simply trace this back to the creator - Elon Musk. I honestly think he knows little about trains. This is a billionaire who travels in luxury sports cars and private planes. And so what does he design? A system that's direct point-to-point with hubs on the outskirts of towns, like an airport, except that it can take one of his luxury cars with too! He talks about it as a replacement for high speed rail, but really, it's like an air link. Which you can bring a car with on.

    The concept *could* become more rail-like, and later on, he mentioned branching and stuff. It needs a lot more, of course, to be a true rail replacement. And that would, indeed, mean major cost increases. But that hardly invalidates the concept from an *engineering* standpoint. From an engineering standpoint, I think he's really hit on the right way to optimize moving passengers along a several-hundred-mile corridor with little energy and minimizing the needed infrastructure.

    You're absolutely right that he shaves a lot of costs off by staying out of the cities (btw, while the long-distance graph shows it crossing the SF bay, the station graph shows it near Oakland). But some of the reason for the low cost, honestly, is simply due to the technological choices. The key issues are that his track is rather small and light and never carries that much weight at any given point, or for that long of a period of time. So, for example, when I first saw his tunnel boring costs, I was immediately skeptical, they looked too low. And then I remembered, wait a minute, the tube is only *7 feet wide*. This isn't a full sized rail tunnel, it's more like a big water main. And thinking about the difference, I think his costs are reasonable. Likewise, the column costs hit me as too low, until I realized, they're not supporting all that much for something that you only need to build 15 of per mile - about 60 tons with no train, and for the brief split second a train goes over, about 120 tons. I did the math on the steel costs, they should be less than a third of the total cost of the tube. And on and on...

    So yeah, from an engineering standpoint, I totally agree. From a practical standpoint, I think he's got a long way to go. Namely, he's taken an approach of, "Local transportation to the station? Not my problem, that's an issue for the city." But it really *does* need to be his problem if he wants to argue practicality. Even airports, which is the approach he's emulating, have to think about those things.

    ReplyDelete
    Replies
    1. Hmm, looking at the map better, it looks more like the station is in not Oakland but Fremont (certainly no further than Hayward). And hey, more from the Musk Angle?

      Musk lives in Bel Air (~20 minutes from the proposed LA station)
      His Tesla Motors factory is in... wait for it... Fremont. ;)

      So yeah, I'm sure his own personal travel habits weighed in, whether consciously or subconsciously.

      Delete
    2. I don't think the guy is in it with big oil, but that doesn't mean he may not being doing this to kill HSR. Maybe he wants to sell more cars. Maybe he owns property by the HSR line and is a run-of-the-mill NIMBY. Or maybe he's a libertarian and hates all things government.

      If he was in it for the technology, why not propose the project for, say, Texas?

      Also, why make a pre-announcement, schedule a reveal date, and then host a press conference for an idea, rather than say, submit it to a Science magazine for consideration?

      As I said in my post, I stayed far away from all the engineering stuff because that's his strength, not mine, so I went with the assumption that his numbers were fine.

      But when it comes to infrastructure? Guy is out of his element. He builds cars, not highways. And yes, he might want a quick trip from Bel-air to Fremont....but who else does?

      Is he attacked HSR because it wouldn't operate as his private shuttle?

      Delete
  20. Rei, even assuming narrow tunnel diameter, the tunnel costs are outlandishly low. Musk himself doesn't buy his own argument about narrow tunnel diameter, actually: a 50% increase in diameter from the passenger-only tube to the passengers-and-cars version leads to only a 17% increase in tunnel cost in his proposal. And that's for a twin tubes each about two fifths the diameter of an HSR tube.

    The cost of a high-precision, tight-tolerance metal tube is much more than the cost of materials. ROW irregularity needs to be zero at the speeds in question; even on full-speed HSR there's a leeway of a few millimeters in gauge and track irregularity. The cost of the concrete piles supporting such a tube is about the same as the cost of the concrete piles supporting a vanilla trains, which is ten to fifteen times as high as the cost mentioned in the proposal without a reference. The cost of HSR civil infrastructure is a lot more than the cost of the raw steel and concrete, too.

    The proposed mass of the larger-diameter tube is 4 metric tons for both tubes per linear meter plus about 1-2 (can't tell the extra length of car bays) for the capsule, versus about 2 for an HSR train and 8 for a two-track concrete slab capable of supporting trains. It's a factor of 2 difference, and the dependence of cost on viaduct weight is much, much less than linear. In the proposal, 60% mass growth from the passenger tube to the passengers-and-cars tube leads to 24% growth in tube cost. Somehow the passengers-and-cars version's cost figure is still about one tenth the actual cost of viaducts on California HSR. The cost numbers don't pan out.

    There are three possibilities. One is that he made a computational order of magnitude error, in which case he should immediately acknowledge it, retract his claim of lower cost, and think long and hard on why he persisted in making this error for a year now. Another is that he's clueless about construction costs and made up numbers. A third is that he's defrauding the public in order to reduce public support for a train system that's competing with the cars he's trying to make money on.

    ReplyDelete
  21. Did you realize that Fremont is where the Tesla plant is and Sylmar is the closest part of the LA territory to SpaceX headquarters? Coincidence or personal bias?

    ReplyDelete
  22. I definitely appreciate the insightful critique! However, he was pretty clear that he has no plans to pursue the design because he is too busy, and he was putting it out open source so that others could solve the problems associated with it. Maybe he made some bogus claims, but it is pretty cool that he is attempting to make huge advancements in the technology/transportation world.

    ReplyDelete
  23. I don't know the specific details about why there's a low difference in tunnel cost between the two tubes - all I can say is that there's a huge difference in excavated volume between that and conventional rail tunnels. Two fifths the diameter means 16% of the volume. Times 2, 32%.

    As for steel: Inner diameter, 2,23m. Outer diameter, 2,25m. Inner cross section, 3,9057m². Outer cross section, 3,9761m². Tube steel cross section: 0,0704m². Tube steel volume: 39.640m². Total tube mass: 310.000 tonnes. Price at $400USD/tonne: $124m. High speed rail parts, due to manufacturing and higher tolerance, generally costs double that, about $800USD/tonne (for example, see the details of the Tata contract for the Mecca-Medina line), which would be $233m in the Hyperloop case. Musk budgets $650m. So his budget is for three times the steel manufacturing expense as conventional high speed rail. I think he's being quite realistic. And it should be noted that he's not exactly new to mass production of large high-tolerance parts, since SpaceX rockets work on a high-precision common core system with automated manufacture.

    You make a claim about how Musk's numbers are wrong and how he didn't include a reference... but then neither do you. I at least have reason to *suspect*, given all of the amount of work gone into this (including 3d modeling and CFD sims, and the fact that this is a guy who lives and breathes economics with his companies) that he consulted people. I have no reason to give you that benefit of the doubt - nothing personal, just stating things as it is. And your reasoning doesn't make sense. Why would pillars carrying a tiny fraction as much weight per mile cost as much to build per mile as something carrying orders of magnitude more? That's illogical.

    The small tube (I've been doing the math on that one, so I'm going to stick to it) cannot be anywhere close to 4 tonnes per meter. It has a cross section of 0,0704m², aka, 0,0704m³/m. Steel weighs 7800kg/m², so about 550 kilograms per meter. The pillars are spaced every 30 meters, so each pillar supports 16.500 kg worth of track (note: I previously thought they were every 100 meters - corrected here). A fully loaded capsule is 26.000 kilograms. Hence each tower supports 16500 kilograms continuous, 42500 in very brief pulses every 30 seconds.

    Do you have solid stats on HSR trains and tracks? Because I've had trouble finding them. Regardless, the mass would be borne in long, sustained pulses.

    Viaduct cost has to be proportional to weight. Maybe not linearly, but it should be noted, the cost of building a viaduct that supports zero mass is zero dollars.

    Passenger capsule mass is not tube mass/tube cost. Tube mass is proportional to its cross-sectional area. Tube cost is proportional to its materials cost and manufacturing cost.

    Defrauding? Okay, let's stake a few steps back from the Rhetorical Zone here. Making a technological proposal is now fraud? Ugh, I hate when debates about technology turn into personal attacks. You realize that this doesn't win any points to your point of view, right? It simply makes people who are on the fence less likely to listen to you.

    ReplyDelete
    Replies
    1. "The small tube (I've been doing the math on that one, so I'm going to stick to it) cannot be anywhere close to 4 tonnes per meter. It has a cross section of 0,0704m²."

      You're off by a factor of more than 2. The Hyperloop document claims that the thickness for the larger-diameter version is 23-25 mm. 0.025 * 3.3 * pi * 7,800 * 2 = 4,000. For the small version, do 0.023 * 2.23 * pi * 7,800 * 2 = 2,500. The cross-sectional area of each of the smaller tubes is 0.16 m^2.

      "Viaduct cost has to be proportional to weight."

      Musk doesn't believe it is: 60% growth in weight for him means 24% growth in cost. I have not seen a single source suggesting it is proportional, let alone linear, and one solid indication suggesting it isn't, namely the fact that slab track is more expensive to install than heavier ballasted track. Details of how difficult the fabrication and installation are are important.

      "So his budget is for three times the steel manufacturing expense as conventional high speed rail. I think he's being quite realistic."

      The total cross-sectional area of HSR viaduct concrete, exclusive of the pylons but inclusive of the box girder, is 10 m^2, so $1,000 per meter. Somehow, HSR viaducts are 50-80 times more expensive than that.

      "Defrauding?"

      Yes, defrauding. The energy cost the document claims is true for trains is 5 times too high for HSR, the technology Musk chooses to bash.

      Delete
  24. It was proposed at $6 billion, ok lets say the smart people at Tesla and Space X are wrong. Double it. Triple, no, quadruple it! Now we are at $24 billion. And lets say to take a few things you pointed out into account, lets just double that! $50 billion dollars, more than 8 times original projected cost... That would be ridiculous if it cost that much, and that is still at least 10 BILLION dollars cheaper then the original PROPOSED cost of the high speed rail. Proposed at $60 billion, the high speed rail will more likely be 70-90, possibly even 100 billion.

    Starting at a cheap $6 billion, the hyperloop would have have more extra budget to work out any problems. Like it or not, the hyperloop or something very much like it is the future of transportation.

    ReplyDelete
    Replies
    1. Cheaper than HSR isn't so great if it takes longer, doesn't carry as many people, and skips critical cities. We already have that - they're called airplanes.

      Delete
  25. What about the fact that this thing is supposed to go 600-800mph? Or just under the speed of sound? Is that something most normal people can tolerate for 30-40 minutes at a time? I can't imagine anyone younger than 14, older than 60, or with any aversion to high speeds being able to complete the trip comfortably...

    ReplyDelete
    Replies
    1. Everybody does it daily on an airplane. Even the newborns. And even on short jumps of just 30 to 40 minutes. That is not one of the many problems this design has. Of course it means that the visual cues of speed must be on a reasonable distance from the pod. I believe the installation on pylons takes care of the ground at least.

      Having said that, the problem is not the speed per se, it's the lateral acceleration requirements the speed creates. Airplanes take care of that by having not many limits on how gentle turns they can make (when cruising). Sky has no curve radii.

      Delete
  26. Even if Elon Musk's Hyperloop were technically feasible and cost the amount it did, the system has an incredibly low capacity - and this is assuming that systems exist to keep the cars 30 seconds apart (in the event of the failure of one pod, a second pod is likely going to crash into the first due to the limits of how fast one can decelerate without causing brain damage in passengers.)

    ReplyDelete
    Replies
    1. At 30 seconds apart, if car A leading car B were to suddenly come to a complete stop, car B would have 30 seconds to slow from 780 mph to 0. That translates to 26 mph/sec, or 1.2 g. Assuming there's a 5 second delay in the system response and you want at least 1000 feet for a safety buffer, that works out to 32.5 mph/sec, or just under 1.5 g, or half the acceleration experienced during a Space Shuttle launch. That's not going to be comfortable for some people, but it won't cause any injuries.

      Delete
    2. Yeah, but the sudden change to deceleration will. At least it will require to people be strapped for the whole journey. And even then, the deceleration cannot instantly jump to 1.5 g. Just go and experience a train making emergency braking, which is much much smoother change in forces.

      Delete
  27. Watching Musk (and others) use this ridiculous proposal to attack HSR is galling. It doesn't pass the most basic engineering and physics smell test, but let's use it to derail a technology that's highly evolved and proven across the rest of the modern world!

    I'm glad that Musk put considerable resources into good ideas like Tesla, but I've now lost all respect for the guy. His proposal looks like the senior thesis of someone who knows enough physics to be dangerous, but not enough to actually make things work.

    Chris F.
    Mountain View, CA

    ReplyDelete
  28. I've read through this article and comments -- what seems to be missing is an answer to this question:

    Would it work at all -- even along some other route? Even 15 years from now?

    Even assuming the cost estimates are off and there is a huge problem with effective braking I'm still left wondering this: In 20 years could it revolutionize transportation as we know it?

    There seems to be some serious small-time thinking here. The Wright brothers flew in 1903 and 40 years later a World War is being fought using the absurd idea of planes that launch from the decks of ships.

    One could easily name a hundred (thousand?) reasons why flight shouldn't have happened and yet, here we are. Even with all the problems, we flew.

    Put another way -- are you saying hyperloop could never work anywhere or just that his proposal fails to present the technology in a viable method?

    ReplyDelete
    Replies
    1. The proposal would make more sense for the Texas triangle. Almost perfectly flat, and from what I understand, the "free" ROW in the highway median actually does connect the major cities.

      I'm not attacking the technical issue (although others have found fault in that), I'm looking at the infrastructure side of it.

      Heres the difference between the Wright brothers and Musk. As far as I know, they didn't use their wealth and fame to try and get rail projects stalled because their flying machine would be ready in 5 years for much cheaper than rail technology.

      The problem isnt that hes acting as an inventor, is that hes acting with political motives.

      Delete