Making tunnels the way you make spaceships | Hacker News Reader

Making tunnels the way you make spaceships

economist.com

30 pointsjkuria57 comments

Cost.

That's what Musk wants to drive down. He has said so explicitly and unambiguously many times. How to bring the cost of launches down? Don't throw the rockets away each time. How to bring the cost of boring down? Increase the speed of the machine (here the multiplicative factor is time, reduce time, reduce cost, ergo increase speed.) How to not throw away rockets each time, reuse them by making them land without damaging them.

I'm not sure he "makes tunnels the way you make spaceships". These are the similarities.

(1) Search for a social problem: lack of cheap access to space, traffic congestion, fossil-fuel propulsion

(2) Figure out what the factor limiting progess: rocket reuse, 3-d traffic network, battery tech for EVs

(3) Focus all your energy on that problem: figure out how to land a rocket, figure out how to bore tunnels for a fraction of the price, build a gargantuan factory optimised to drive down the cost of battery production

Look folks. There's no magic formula here. It's not rocket science. (Except in one case). I'd say what is common to all three ventures is Musk's personality. Tackling enormous problems. Getting to the heart of the problem. Self-belief. Drive. Focus. Tenacity. Bravery. Great engineering skills. Surround yourself with talented people who get shit done and have the same optimism as you. Wash, rinse repeat. What the biotech community need to be figuring out is not how to clone monkeys, but how to clone Musk.

hagiography over :)

A good chunk of the new generation of kids is watching SpaceX's success. This drives the kids' attention to Musk which ultimately kids take him as a role model. Kids will follow his steps and thread of thoughts, which is in some way "nurture cloning". We cannot conceive what Musk is actually doing on the next generation, it's going to be amazing

I watched the Falcon Heavy launch with my kid and she was gob-smacked. I don't remember anything like that growing up.

onions.

I could only draw on my experience of Challenger for reference.

Such contrast.

I could not agree more. Cost is the ultimate pain point.

The problem of cost in American tunnel projects has essentially nothing to do with technology.[0] I strongly doubt Musk will succeed in this effort, but if he does, it will be because he's managed to bypass existing labour agreements and/or violate labour and safety laws, not because he has invented a new way to dig a tunnel.

[0] https://slate.com/business/2018/02/better-technology-isnt-th...

Well shoot, somebody better tell Elon that. :)

Nobody can tell Elon anything.

I think there are a lot of things susceptible to this kind of analysis.

I also believe that the major jump is not this analysis per se, but the fact that Mr. Musk has access to capital and the ability to follow through beyond the average.

So why aren't many other billionaires launching rockets?

Bezos and Branson have tried, with some success, but Musk clearly is in a different class.

To add a little perspective: The falcon heavy has a LEO payload of 63 800kg, the Saturn 5, a launcher designed and built in the 1960ies, has a LEO payload of 140000kg, more than twice as much. There’s certainly a lot of cool tech developed at Space X and landing and reusing the boosters is a thing not achieved in the 60ies, but it’s always good to keep in mind that in terms of lift capacity, today’s heaviest launcher is less than half as powerful than a launcher half a decade out of service.

Worth pointing out however that the Saturn V cost roughly $1.2B per launch, compared to $90M for the Falcon Heavy. So it was roughly 6x more expensive per kg to LEO.

Not to mention it required a nation state to fund its roughly $40-80B development cost. Can only guess, but would expect the development costs of the Falcon 1/9/Heavy were maybe $3-5B?

That said, Space X has benefited from all of the R&D done in the past. They weren't building it in a vacuum.

You said it first and better :)

In the press conference after the Falcon Heavy start a journalist asked the question of how much the investment was. Elon answered “around half a billion [silent for 1s] propably more.”

Correct, but that was just the cost to develop the core booster of the Falcon Heavy.

That doesn't include the original costs to develop the Falcon 9 and all of its subsequent variants, nor the cost to develop the original Falcon 1.

As a counterpoint, the falcon heavy costs something like 1/10th of a Saturn V launch. So you can launch two and dock the payloads in orbit for 1/5 the cost. Source: Kerbal Space Program

KSP FTW!!!

I'm not sure what your point is?

1. No privately funded rocket venture has ever gotten nearly this far

2. Reusable boosters are a Big. Deal.

3. The BFR will have a higher LEO payload and is likely to begin testing next year

4. Payload size is a poor metric by itself, launch costs are an extremely relevant part of the equation

I could go on, but what I'm saying is that even with "a little perspective", SpaceX has achieved something hugely unprecedented, astonishing and inspiring. The prior existence of the Saturn V adds some context, but it doesn't even begin to address the question of why no other 'rich dude' has tried to put together a team to do something similar at this level, which I think is a fascinating one worthy of discussion.

They're missing the critical component: intimacy and competence with the core craft. Elon Musk can contribute and hold his own in a room with the rocket engineers doing the heavy lifting. He can understand what they're talking about, perhaps lacking only the local context, on which he can be caught up reasonably quickly.

This is not to say that Musk has done all of the engineering directly or that he understands every single component of the product, but he knows the product with the intimacy necessary to realistically evaluate and engineer in detail. He knows what's possible, what's reasonable, and what to expect if you get the best people and put them in a place where they're able to do their best work.

That, IMO, is the biggest thing that separate the wannabes from the revolutionaries. It's extremely difficult to bullshit Musk because he knows what's going on and doesn't just have to trust "an excellent communicator" to dumb it down into digestable "non-technical" nuggets for him.

There is no need for a translation layer. No nervous middle-managers trying to "soften the blow" regarding some impractical contrived fantasy that the engineers laughed out of the room. SpaceX doesn't have to waste years of resources on said fantastical contrivances because the middle managers aren't willing to shut down the idea forcefully enough, or worse, because the C-suite is unwilling to accept the engineering evaluation.

Musk knows the reality of the situation in a real, visceral way, and the value of that cannot be overestimated. That is the true superpower. As smart as Bezos and Branson may be, as far as I know, they cannot claim the ability to understand what they're trying to make, and that is the core differentiator for SpaceX.

-----

To get industrial revolutionaries like Musk, you need to combine the financial independence (true financial independence, no VC babysitters), the vision, the tenacity and courage to blaze the trail, and the technical intimacy to know when you're running afield or when someone in the chain of command is pulling your leg. Getting all of these in one package is very rare and it's what makes people like Elon Musk stand out.

There are certainly many people with some mix of these components, but not many that have them all. If we want more Elon Musks, we need to ask how we can get the more fungible parts of this recipe to more people who are only lacking those elements.

At any point, Zip2 or PayPal could've taken a misstep that would've erased Musk from the history books (perhaps one such misstep would've been Elon taking role of CEO at Zip2, something his investors did not allow). He'd be a good engineer with a passionate interest in rocketry, like many other engineers before and after him.

How many sleeping Elon Musks are out there today, waiting for their own unlikely exits to materialize?

Follow-up for people who feel like typing more (because I don't): where else has this applied? Walt Disney knew animation, even if he wasn't the best animator. Thomas Edison knew electricity, even if he wasn't the best primordial electrical engineer. Sergei and Larry knew search. It seems that the earth-shaking successes are practically always spearheaded by people who know what they do at the fine-grained, detailed level.

I was expecting Musk's company to improve the back end of the TBM. Nobody seems to pay much attention to that. There's spoil coming back, and ring segments going forward. The usual setup is to lay a temporary narrow gauge railroad behind the TBM. The back end of the TBM does much of the track laying. Much of the TBM's length is for laying track, filing spoil cars with dirt, taking ring segments off segment cars, and getting cars from the incoming track to the outgoing track.

All those rail cars could be replaced with battery-powered self driving vehicles. That would get rid of the railroad track and simplify operations at the back end of the TBM. Self-driving work cars would be more maneuverable than rail cars; they can get out of the way and pass each other.

> I was expecting Musk's company to improve the back end of the TBM.

IMHO that doesn't even start to scratch the main problem that they gloss over or sometimes even appear to purposely ignore. The main problem with tunneling in urban areas is subsidence, and all the financial compensations that must be paid to those who see their property start to show structural problems due to the resulting differential settlement. I never saw this issue being addressed by either Musk's fanboys or any representative.

Tunneling takes time not because of the amount of dirt that needs to be taken out but because even highly controlled tunneling operations can expect differential surface settlements in the order of inches. Musk didn't suddenly and single-handedly invented the concept of tunneling. People have been tunneling for quite a few decades and these problems are still hard to handle. Yet, somehow this issue is never addressed in these marketing blurbs.

It seems that the subsidence is because the tunnel does not reproduce the expansion-contraction characteristics of the spoil it replaces. A possible modernization is massively-dense conditions sampling along the proposed tunnel route, not just in the volume the tunnel replaces, but around it as well, possibly all the way up to the surface. Conditions I can think of off-hand are temperature, ambient moisture/humidity, drainage volume, and groundwater baseline versus historical. Set up the sampling as many years in advance as possible to establish good reference data. Then create hydraulically moveable tunnel rings that push against a static ring, to re-create the desired movement. Use machine learning to refine the model during the pre-construction sampling period, and to refine results during operational period. The sampling of the surrounding volume feeds into the machine learning datastream so the hydraulic movements can be anticipated and performed very slowly over time.

> Set up the sampling as many years in advance as possible to establish good reference data. Then create hydraulically moveable tunnel rings that push against a static ring, to re-create the desired movement. Use machine learning to refine the model during the pre-construction sampling period, and to refine results during operational period.

Yeah, it's definitely going to be cheaper to just pay for subsidence.

Not to mention that it's literally impossible to perform the kind of sampling you're talking about. It's basically the ultimate wet dream of the oil, gas and mining industries and they've poured trillions of dollars into developing subsurface sampling for over a century, with nowhere near the resolution you're thinking about.

I thought the depth and narrowness of the tunnels was to address this?

> I thought the depth and narrowness of the tunnels was to address this?

No, changing the stress field within the soil layers changes that. Digging a tunnel through soil causes a decompression that isn't balanced by either the tunneler or the supporting structure. That happens even if the tunnel is pre-grouted, which takes a considerable amount of time and money.

Musk's team excels at marketing, but if you ask any civil engineer who has any knowledge of tunneling (the boring company) or railway systems (hyperloop) he'll be able to point out about a dozen basic and major technical and economical problems that they simply fail to address or even consider. The reason things are the way they are isn't due to a major conspiracy, or widespread stupidity. In fact, the only stupidity I see is in how these projects promise magical solutions to problems they even failed to identify adequately.

I believe there are strong reasons why all these business pitches end with requests for investors to cough up cash to pay for the projects themselves, even though these million-dollar projects are proposed by a billionaire with a net worth of around 20 billion dollars.

Thanks for the post!

I guess there are a few things that might make these projects less of an unmitigated disaster than your analysis suggests:

1. Musk is clearly working off the common delusion that 'getting in at the ground floor' is a good idea. In reality, innovators and groundbreakers typically go bankrupt. Investors who buy the bankrupt company are the ones who typically profit off the expensive research and development, without the associated risks.

2. If you consider similar examples, for instance, the English railway system - early investors losing money is by no means a bad thing. Even if 'getting in early' is bad business strategy, it's good for society in general if moonshots get funded, even if they never turn out to be profitable.

3. These projects, while obviously unsound and based on false premisses, will typically produce useful technology, knowledge, or techniques - that can later be taken up by governments (as in the case of railways) or private investors.

So I mean, sure, these projects are stupid and will end in stupidity. But they're also a better (from a human perspective) use of investor capital than pretty much everything else it could be spent on - more Walmarts, Amazon surveillance devices, or oil exploration - even if those things are certainly safer bets in terms of profits.

To be fair, that’s also what rocket engineers said about spacex’s plans originally. The difference with musk is one of perseverance and the assumption of possibility. Instead of seeing problems, see challenges to be solved.

Purely from a business perspective isn't it smarter to use other people's money for such risky projects? Musk certainly didn't become a billionaire by gambling!

> Purely from a business perspective isn't it smarter to use other people's money for such risky projects?

IIRC Musk wants nothing to do with Hyperloop, and the boring company appears to be the marketing name they gave to a purchase of a pretty standard tunnel boring machine to dig a pretty standard, good old fashined tunnel at SpaceX's.

That's it. The rest is marketing.

Comparing with the level of investment in Tesla and SpaceX, both with regards to personal time and money poured in, it's pretty easy to notice which projects are handled seriously and what projects amount to marketing.

I sincerely doubt self driving vehicles would outperform light rail in that role. They’d still need a track, they need charging, they still require loading and unloading. Light rail provides a track, continuous electricity and has a higher loading capacity as well since it doesn’t need to lug batteries around.

Tunnel construction track usually isn't powered. Tunnel construction locomotives are usually battery powered.[1] The Channel Tunnel did use overhead wire during construction, but that was an unusually long tunnel with no intermediate access points.[2]

[1] http://www.zslocomotive.com/locomotive/ [2] https://commons.wikimedia.org/wiki/File:Channel_Tunnel_const...

I wondered if they could just use a high pressure water cutter, instead of a TBM machine.

This would results in no actual contact with the rock, so no wearing and tearing of the blades on the TBM.

Then, they can build out a robotic arm that will maneuver the water cutter.

Then, they can add Artificial Intelligence to the mix. The AI will sense where to cut the rock face at the correct angles, to allow for continuous collection. The rock pieces are carried away by a conveyor belt. And the water is recycled.

The human operator just has to sit there and babysit the machine. And if it encounters an edge case, that the robot cannot handle, then he can manually maneuver the robotic arm himself.

Something like video: https://www.youtube.com/watch?v=hmlUzUl-h5c

That's amazing tech. Would lasers not be better suited for underground cutting for tunnels?

Lasers would probably need a lot of energy to cut that much material.

Can somebody do a quick analysis how much it would probably cost using conventional/water/lasers?

Do we even have the capacity to create sustained lasers with sufficient energy output to cut through solid rock?

Actually they would vaporise the material, not cut it. So yeah, a lot of energy needed !

I don't think water jets have much effective range, maybe a few inches tops. Plus, the jet nozzles do take a lot of wear and tear and need replacement.

I'd think a typical physical rock muncher would get you more distance bored per day.

That would probably work only for hard rock TBMs. Soft rock TBM should support constant pressure to the excavation face.

Water jets cut rock with the garnet powder in the stream. [0] Which is not recycled. And they are maintenance intensive. They are great tools, but like most, they are far from pushbutton.

[0] https://www.barton.com/waterjet-abrasives/

A boring machine that spits out "cinderblocks". I like creative engineering solutions.

I live near the coast and the cost for underwater tunnels is outrageous.

I don't see why it can't eject massively wide core sections of rock with a few cinderblocks as a side portion.

If the tunnel is 10 metres diameter then instead of churning that rock to 'cinderblocks', a big long cylinder of 'rock' 9 metres in diameter or so could come out, broken up into chunks that are 50 meters or so long.

at the far end of this tunnelling effort could be a harbour under construction with the massive core sections being used to create things like the harbour walls.

Imaginably this could be possible with some 'swarm' of 'Boring' robots that could adapt the tunnel size as they go, maybe also moving the 'massive chunks' out of the tunnel too. Maybe each robot could fit in the back of a pickup truck and be fully electric.

Do you have a rough idea of how much such a cylinder weights? It has a volume of about 3200 cubic meters. Soil has a density of somewhere between 2 and 3.3 metric tons per cubic meter depending on the type of soil. So you’re talking about handling and lifting blocks of anywhere between 6 000 and 10 000 tons apiece.

And that still assumes that the cylinder is made of solid rock that doesn’t break apart.

Previous (similar) discussion:

"Boring Company to use Tesla’s technology for its tunnel project under LA"

https://news.ycombinator.com/item?id=15002501

Till now the Boring Company was "playing" with existing, used, conventional, TBM.

This article makes an error I see alot: person/group uses unusual technique X, therefore they foolishly only (or over-) use X. In this case, it's Musk thinking in terms of the underlying physics constraints. For effective altruists, it's quantifying charitable impact.

That last sentence seems like a non sequitur (or, at best, a radical tangent).

I agree! I thought the article ended there because I ran out of free monthly articles and I got cut off.

Seems like half of it is missing.

Ditto. I was intrigued and eagerly reading and then looked for the "more" button.

My last sentence? It's another example of the mistake I described.

Yes, your last sentence. What does it have to do with the article? How is an example of a mistake?

Since I was criticizing the article of making a type mistake that was not specific to Musk, I wanted to give another example of this type. Likewise, if I had said "the author is committing the typical-mind fallacy" to an audience who wasn't familiar with it, it would be helpful to provide a quick example of that fallacy that was distinct from the one made by the author.

What is bothering me: why do we only hear about Musk when it comes to his companies? Surely he cannot be doing all the thinking. I'm more interested in what the heads of engineering have to say about the technology.

Anyway, I couldn't read the article because of the paywall.

>Surely he cannot be doing all the thinking.

the value isn't in the thinking. There are a lot of different thinking around - for every taste and possibility. The value is in choosing a thinking and funding/implementing it.

It's like Edison. He got all the geniuses together and gave them a place to experiment. He probably got his hands a little dirty, but for the most part someone else is doing the hard work. Having a single name attached definitely gives the impression that Musk and Edison have/had a large ego, but there is something to be said for creating such an environment.

Well, I guess you are right. It doesn't help to have a bunch of egos in science or in your company. So what do you do? You place an even bigger ego at the top. Makes sense, but still, it doesn't make for nice/interesting reading.

This article has about as much substance as Musk's Boring Company. Little to none.

The tunnel under L.A. is already quite long and now under the 405 freeway.

So what? Do you really think it's going to relieve traffic?

Reply on news.ycombinator.com