The capacity of the structure is not just limited to the weight of the revenue vehicles, the structure needs to handle other loads it may encounter over the its useful life, such as snow (probably not much of an issue in central/southern CA), water (if the tunnel leaks) and maintenance vehicles (which will likely be significantly heavier than the passenger cars since they will have to be self-propelled).

then put down the ties, then put rail on top of those, fasten it down, etc. Ties are in the range of 6-8" thick and then there's the ballast (rocks) which aren't light either and which are generally required to keep the ties from moving.

You don't need ballast & tie track on structures.

In any event, cutting the material weights in half (even if you can, which is doubtful) does not automatically cut your costs by 90 percent. You still have the environmental process, you still have NIMBYs, you still have construction mobilization costs, etc.

I would argue that since the tube sections would all be pre-fab that means that the installation cost of the tube would be really small. Rail installation costs are no joke due to the need to precisely align horizontally and vertically the two rails.

You can make tube in 50ft sections and ship them via truck without needing any special permitting and given that the highway runs right next to the planned route, the pre-fab notion is quite a good one.

The hyperloop sections would also need to be precisely aligned.

You can make tube in 50ft sections and ship them via truck without needing any special permitting

But you may need special routing for those trucks to allow overhead clearance and turn radii, and you may need to build temporary roads for access to the installation site. There are lots of little things that tend to add up that are ignored in the whitepaper. You may be able to reduce costs by a few percent, but 90% is not a figure that I can take all that seriously.

If you wanted to do the 11ft you would need a low-boy trailer to avoid height permitting and you would need "wide load" signs for the trucks individually but no lead/chase vehicles.

The tube joining/welding/grinding jig could be 40ft (or 200 ft or more) long and fit inside the already assembled tube AND the to-be-welded tube. Precision alignment could be done via movable clamps. Once the tube is held in place via the jig it can be welded via a number of methods (friction stir, friction, MIG, TIG or SMAW) inside, outside or both.

You are not going to give yourself 6-inches of clearance for a component that weighs (using your numbers) 19.2 tons, that pipe is going on a flatbed. You also have to allow room for tie-downs, flanges, support blocks, etc. You also have to account for bridge weight limits.

I'm not saying this isn't doable (the Big Dig was able to truck in 100 ft soldier piles, although they had to do it at night), it's just that it often takes extra planning and work which increases costs.

The tube joining/welding/grinding jig could be 40ft (or 200 ft or more) long and fit inside the already assembled tube AND the to-be-welded tube.

Are you basing this on something that already exists, or is this something you are proposing?

If it's possible to make two tunnel boring machines meet at the middle of the English channel with the starting points separated by some 20 miles, then it's entirely possible to build a pipe-section-holding jig which can ensure precise alignment. I probably can't write a check for one and have it in a year (partly because I don't have $5mm, also because the plans don't exist yet) but to suggest that it would take more than three years to design, build and verify such a machine is a bit extreme. You could probably do it in a year or less provided that you're willing to pay more.

The machine as a whole doesn't exist yet but all the various component pieces of it do and they're all COTS parts. It's part tunnel boring machine (http://en.wikipedia.org/wiki/Tunnel_boring_machine) part orbital welder (http://www.mecomachineandfab.com/public/images/photos/orbita...) and the rest is "glue."

Make it 300 feet long so that it's always resting on pipe that's already affixed at two separate pylons and so that it doesn't deflect the pipe on either side of the joint it's about to make. An external crane (or if you get clever, one built onto the TBM/welder rig) lifts the pipe up, the jig grabs hold of it, positions it, and then makes the weld. Let the machine advance slightly and x-ray the entire weld for QA/QC and provided it passes, advance the machine to do the next section.

If you wanted me to make one of these for you I'd probably quote $500k-$1mm for the initial research including a small prototype (10-20" diameter) and I would guess (+- 50%) $10mm for the big one.

I built an semi-automated package dimensioner and barcode reader for about $5k in parts plus my time The NRE was probably around $40k. A fully automated one (including transport) starts at something like $80k. The company I worked for then might have lost money on the first one (only if you look at opportunity cost) but by the time they build three, they've saved at least $150k. Three of them plus the employees to operate them would have a similar throughput, higher dimensional capacity and better exception handling (what do you do with a damaged package? Or one that's leaking?) than the one fully automated system. An $80k system that can do 10x what your needs are is definitely more expensive even if the per-unit costs are the same because you're paying for idle time.

Elon Musk isn't going to call me up and ask me to build that system for him. But if he did I wouldn't have too many problems making it happen. The difficulty isn't making the machine, it's finding someone with the constitution to ride out the periods of uncertainty and doubt while all the bugs are being ironed out.

EDIT: When I've suggested that I could do the research and build a 10-20" prototype I'm referring just to the pipe positioning jig + orbital welding system. Not a small-scale hyperloop. That would be quite insane. Just a precision alignment jig and associated welding equipment to do the final fit and assembly of the pre-fabricated pieces.

For example:

If it's possible to make two tunnel boring machines meet at the middle of the English channel with the starting points separated by some 20 miles, then it's entirely possible to build a pipe-section-holding jig which can ensure precise alignment.

Of course it's possible, I've never claimed that it wasn't. The issue isn't "is it possible" it's "can it be done at less cost than current proposals." History shows numerous examples of new construction techniques that worked great on paper, but had significant implementation flaws - for example soil freezing was supposed to be a panecea on the Central Artery/Tunnel project, but there were a ton of problems in the implementation and unexpected results that needed to be dealth with. Even your Channel Tunnel example proves this point - it used a lot of innovative technology and it came in late and over 80% over budget.

You talk about a welding rig that doesn't exist as a fait accompli. You wrote "My point was to say that it's entirely possible to fit that size of a tube onto a semi trailer with no hardship at all.' - that's bull, I've seen crews have a difficult time moving smaller and lighter pieces off and on semis. "No wide corners, no permits, no height restrictions, none of that" - really? You're familiar with every possible street route along the proposed Hyperlink corridor? Because if a truck gets stuck someplace (it happens: http://www.wwlp.com/dpp/news/local/hampden/truck-stuck-at-ho... ) the project is in for a bad time. And again, we haven't even begun to address NIMBY issues.

That said (and I'm being serious here) if you have the chops to make solutions like this work, you need to be working (or consulting) for the big engineering firms like AECOM, PB, CH2M Hill, Jacobs, etc. Your ideas could revolutionize the industry.

The difficulty isn't making the machine, it's finding someone with the constitution to ride out the periods of uncertainty and doubt while all the bugs are being ironed out.

Exactly. That period is called "final design and construction." It's been the death of many construction and A&E firms.

As far as loading tubes go, you're likely going to build a tube rolling facility. These kinds of places tend to have overhead cranes. Big ones. Cranes that can grab the tube on both ends and thus have precise control over it. Most likely the loading would be done "indoors" so weather is mostly factored out. Building semi-circular restraints that grab the tube in 3-5 places along the 50ft length is done easily enough with a water-jet or you could roll a pipe into a semi-circle and weld supports to it. Once you have the custom trailer done it's relatively straightforward. You'd build the infrastructure to handle this in a highly automated fashion, similar to how container ships get loaded and unloaded. ~350 miles is 36,960 50ft sections of tube, double it since we're doing double-wide. If you know you need that much at the beginning there's a lot of work that can be done to make the whole thing reasonably efficient and cost effective.

I don't understand why you're expanding the scope to NIMBY or access roads or stuck trucks (locate the plant near the freeway) or any of this. I was talking specifically about the loading and unloading of pipe and making a jig to hold said pipe for precise welding.

Again, when I said "The difficulty isn't making the machine, it's finding someone with the constitution to ride out the periods of uncertainty and doubt while all the bugs are being ironed out." I'm referring only to building the alignment/welding machine, not the entire Hyperloop system. I realize that building the whole thing is immensely more difficult than building the alignment/welding jig. But we're inventing very little new stuff here, mostly repurposing things or making incremental changes.

All the "final design and construction" that you're talking about has the same risks and problems for the HSR as for the Hyperloop. If I have to be exposed to an 80% cost overrun I'd rather it was on a $5bn or $7.5bn project rather than an $80bn one. Maybe Hyperloop even has a 300% cost overrun and somehow the HSR project could come in 25% under budget. Hyperloop would then be (assuming the bigger version) $22.5bn versus the rail at $60bn.

The reason I don't work at a big engineering firm is that I couldn't get hired in at a senior enough level to actually get anything done. So I work for small companies. Thus far my track record is excellent but it would probably take me another 10-15 years for me to have enough successes at difficult projects for people to stop assuming it's a fluke. And by then I might not be too interested in risking my career on every project I work on.

This is why startups are even possible. The risk aversion of senior management makes it possible for people who are too young and too dumb to know that something's "impossible" to make it happen. If you've got any ideas how I could get hired at a big construction firm at a senior level I'm all ears.

There is all kinds of crazy stuff out there: https://www.youtube.com/watch?v=E0O78AyDixY

They're using a $45mm, 45 foot diameter TBM in Miami right now that's 450 feet long, has the ability to erect concrete rings inside of the machine, and can exert crazy pressure on the rock face. http://www.portofmiamitunnel.com/faqs/tunnel-boring-machine/

I'm suggesting something with less challenging specs and suggesting it could be built for less money. I don't think it's unreasonable.

In the comments to his blog post, the OP disputes that lighter weight is going to lead to significantly lower pylon cost. I didn't look through his references, but he also has an update showing how assuming that pylon costs are going to be linear in the weight is not consistent with the hyperloop document itself.

I had to reread the update a few times, but the argument is: the proposal contains two versions, for passengers and for cars. The car version will weigh more and cost more, but the cost increase is much less percentage-wise than the weight increase, so the hyperloop designers themselves don't believe that pylon will be linear in the weight.

As I said elsewhere, this is why references are not just a formality. If the designers had explained where they got their cost estimates from, we could still argue about their accuracy, but we wouldn't have to speculate where they came from.

Of course, this would have rather poor capacity, but that's a problem with the hyperloop as well. It would also be slower at maybe 200 - 250 MPH, lowering capacity further.

But if someone claimed they could simply do a 250mph light rail at drastically lower cost, people would necessarily be skeptical. Adding a bunch of really interesting high speed tech on top, shouldn't change that claim much. About the one way I can see it should change the skepticism is if the increased speed increases capacity sufficiently. Which seems not to be the case.

2) I would expect that wind pressure on a tube is higher than on a rail with an occasional train on it. If so, the supports need to be designed to handle that higher load.

3) I think higher speed trains going through a bend in the track exert higher forces per kg of train than slower speed trains. Again, if so, the design will have to accommodate for that.

4) I don't see how that 0.5^3 makes sense. If both track and train are half the weight, my conclusion would be "total structure to lift is half the weight", not "total structure to lift is a quarter the weight".

I am not an engineer, so corrections welcome.