That’s my projection as well. If the human miners need to travel around the asteroid belt to get closer to the mining locations, it would only be for trouble-shooting, and to reduce the time-delay for remote operation.
Like most heavy industry on Earth in the future, industry in the Asteroid Belt will mostly consist of ‘robot-wrangling’, organising automated systems of various kinds to perform the necessary tasks, and fixing them when they go wrong.
I think that we might be making the heat issue more complicated than it needs to be (note: Simple is not the same thing as easy). Ultimately, what it comes down to is that we’ve got a fixed amount of orbital energy, that we need to convert to some form other than orbital energy, and realistically, that “form of energy other than orbital energy” is going to be heat. The only part that we even can hope to change is what proportion of that heat energy gets deposited in the spacecraft, vs. what portion gets deposited in the atmosphere, and realistically, it’s mostly going to be the spacecraft no matter what we do, just because of the difference in thermal mass between the spacecraft and the portion of the atmosphere it’s interacting with.
Capsules aren’t intended to produce a lot of lift or aerodynamic guidance, and in fact are designed to attain only enough lift needed to moderate deceleration. With too high of a lift-to-drag (L/D) ratio, they would have an extended period of reentry with a tendency for skip-glide phenomena which would make them difficult to target. The STS Shuttle Orbiter had an enormous amount of lift which was a result of the US Air Force requirements for cross range to do a once-around-polar-orbit mission (never actually done since the Blue Shuttle program was cancelled post STS-51-L), and as a result had to perform a series of S-shaped maneuvers to waste kinetic energy. The X-37B does just fine on reentry on almost exclusively body lift alone in all regimes of flight. (I suspect it also does some S-maneuvers upon reentry but less pronounced than Shuttle, and being a long duration uncrewed spacecraft with a much smaller payload it doesn’t have the dramatic cross range and deceleration limit requirements).
Actually, it is the opposite; the TPS systems on reentry craft are designed to protect the vehicle from accepting most of the heat energy generated in the compression wave (via transfer of the vehicle momentum to the atmosphere) because that amount of thermal energy would make the vehicle inhabitable and likely cause it to fail structurally, as happens with satellites and pieces of spacecraft that are not protected (at least, we hope; nobody wants to get hit in the head with a titanium toilet seat from a deorbiting space platform and be known in the hereafter as “Toliet Seat Girl”.) The kinetic energy and momentum of the returning vehicle are transferred to the atmosphere which reacts by becoming hot gas, some plasma, and acoustic shock waves. In the case that the TPS fails, you get a result like the Columbia failure, where the thermal environment and aerodynamic forces break the vehicle and its occupants down to component parts.
The Earth’s atmosphere is actually just-so in regard for reentry, being thin enough that a solid object can fly through it at hypersonic speeds without the drag forces rapidly decelerating it to a violent stop and still develop useful lift and drag at subsonic speeds even below terminal velocity, all of which requires aerosurfaces and the associated TPS that are only a modest fraction of the total vehicle inert mass. The thin atmosphere of Mars, on the other hand, is just thick enough to cause a lot of heating, but so tenuous that you need a very large aerosurface-to-mass ratio to get significant drag, and not enough for any kind of controlled gliding or powered flight in even the low supersonic region. Flight control at hypersonic speeds in very thin media is still certainly possible but that has proven very challenging even in Earth’s atmosphere which is very well studied and pretty stable; in the Martian atmosphere that has a lot of seasonal variation in density, this may be far more challenging than is presented in conceptual designs.
Stranger
The stainless steel construction of Starship will help with this. Probably not with a Columbia-level failure (that was a huge hole), but perhaps something a bit smaller. The Shuttle used mostly aluminum construction, and while they concluded that this was a net mass savings versus materials with higher temperature resistance, one tradeoff was a kind of thermal fragility–break through the fragile TPS and all is lost due to aluminum weakening at a low temperature. Starship should be able to handle small breaks at least.
Starship won’t have anything beyond the stainless skin on the leeward side. The steady-state temperature will be pretty warm but below what the SS can handle.
At one point, SpaceX was looking at tranpirative cooling–squirting cryogenic propellant out through tiny holes in the skin, with no tiles or other passive TPS system. They seem to have given up on that, though it’s not impossible that they’ll use it for some spots on the ship.
God knows the both of you have forgotten more about physics than I’ll ever have learned, but aren’t typical stagnation temperatures at hypersonic shock boundaries well beyond the plastic deformation temperatures of even high-temperature steels? IOW, why would it even matter that the Shuttle’s framework was made of aluminum alloys vs Starship being made of stainless steel, given the temperatures either would be facing in the event of a TPS failure would be grossly in excess of the temperature needed to deform either material?
What is important is the not the temperature of the compressed gas at the shock boundary per se but the resultant heat flux that gets through to the structure, and how that causes the material to heat up and degrade mechanical properties for the duration of use. The SpaceX Starship vehicle is made of AISI 301 stainless steel (which is not a particularly high temperature material) which has a melting point at ~2550 ℉, but of course a structural material cannot be taken up to melting. Looking at MIL-HDBK-5J (not the current MMPDS-01 materials handbook but freely available online for anyone to reference, and has the same tables for this material) shows the yield (Sy) strength drops 60% of room temperature at 600 ℉, and less than 40% at 1500 ℉, where the curve ends at which it probably falls of precipitously. Ultimate strength (Sut) falls to 70% at 600 ℉ and ~25% at 1500 ℉.
Stainless steel, like most metals, is very thermally conductive so a high heating rate would transfer thermal energy into adjacent areas rapidly, which would presumably include critical structure, so unless the vehicle has a very large thermal mass to mediate the heating you can expect large parts of the vehicle structure to get hot very quickly. Of course, it is possible that the skin is spaced off of internal structure via non-conductive standoffs and insulation, and thus is not as structurally critical, but this would be a complex design to fabricate. Early in the STS (“Space Shuttle”) conceptual design, NASA and the various proposal candidates looked at titanium and steel honeycomb structures as the bottom TPS, and essentially concluded that the weight penalty and effectiveness were not acceptable, electing instead to select the high-temperature reusable surface insulation (HRSI) despite the development cost and risk.
Note that the Shuttle leading edges see a peak heat flux of around 50 BTU/sqft-sec on the RCC leading edges but only about a tenth of that on the underbelly from a Low Earth Orbit descent speed because they are far enough away from the shock front. Apollo capsules probably experienced about four or five times that much (I can’t find a canonical reference offhand), and a vehicle entering from an interplanetary trajectory is probably an order of magnitude energy above LEO (depends on the specifics of the trajectory) and corresponding heating rates. The details of reentry heating will depend upon trajectory, vehicle profile, et cetera so it is impossible to say without more design and mission information how realistic it is to just rely upon the thermal mass of a stainless steel structure vice a thermal protection system but I have a difficult time seeing how that would be workable.
Stranger
To be clear, I was speaking only of relatively small failures. Even the Shuttle lost tiles frequently, and Columbia was only lost with enormous damage to the most crucial part (the leading edge of the wing). But as Stranger has emphasized, it’s the heat flux that is the most important aspect, and small amounts of damage amount to a higher heat flux through the TPS system. Whether the structure can handle that depends on the materials and the design. Stainless certainly can’t take the full brunt of reentry, but it should still have an advantage here over aluminum (which has a significant reduction in strength even at 300 F).
The STS-27 Atlantis mission probably experienced the worst loss of TPS in any mission (aside from Columbia); this video is a nice summary. One tile in a fairly crucial location was fully lost; there just happened to be a steel plate behind that tile, which they think probably saved them.
And of course there’s the leeward side to consider. Temperatures here are much lower, and the Shuttle used lighter weight TPS here, such as Nomex blankets. But I don’t think anything was left as bare aluminum. Bare steel however should be adequate in some locations.
I’m sure that SpaceX will encounter some surprises as their design evolves, and even a few loss of vehicle incidents. So we may see the design change yet again. The most important design aspect may simply be the placement of the orbiter component: on the top of the stack instead of the side. Most of the damage to the Shuttle’s TPS came from impact of debris coming off the center tank or side boosters. Starship won’t have that problem.
Stranger - thank you for the continued wealth of information and analysis.
Other than cost, what’s the rationale behind choosing 301 over other stainless steel ? As you said, it’s high temperature strength is not as good and so is it’s corrosion resistance. Looks like there’s a fair bit of chloride and perchlorates on Mars and there are frequent dust storms. Is corrosion when the vehicle is on Mars (or subsequently) a concern ?
To be clear, we don’t know exactly what grade of stainless they’ll use in the final product. What Musk has said is this:
They’ll have their own alloy that is similar to 301 or 304 (which they’ve built prototypes out of), but we (and maybe they) don’t yet know what the final alloy will be. They’ve sometimes called it 30x as a kind of placeholder.
It’s easy to see how wildly wrong this is (no insult intended; intuition just happens to be a pretty poor guide to this kind of thing). Low Earth orbit is around 8000 m/s (more for reentry from other bodies). That means stuff in orbit has 32 kJ/g of kinetic energy. Most materials have a heat capacity somewhere around 1 J/g-K (aluminum is 0.9, for example). Divide through and you get a 32,000 K temperature increase if all that heat goes into the object. Clearly, most of that heat is going somewhere else.
The main reason I would hypothesize that AISI 301 was selected was because of ready availability. Corrosion resistance, especially after exposure to high temperatures, is certainly a concern on Earth and in a near-marine environment, but in the 300-series 317L and 347 would be more common in aerospace applications. 301 and 304 are most commonly used in food service and general material handling applications, although I have seen both of them used in cryogenic applications for commercial aerospace because they maintain fracture and fatigue resistance at low temperatures. I can’t imagine that corrosion is a primary concern on Mars given that there is essentially no water vapor in the atmosphere. The idea of creating a proprietary stainless steel alloy seems contrary to the intent of maintaining a low cost production; you might be able to concoct an alloy that is some balance of high temperature performance, cryogenic performance, and corrosion resistance, but it isn’t as if a minor tweak in some constituent is going to radically change the mechanical and thermal properties of the metal, and the cost of bespoke production of sheet steel is opprobrious unless you are willing to purchase in mill run quantities.
Stranger
Not sure what a minimum custom mill order is, but SpaceX will likely be needing several thousand tons per year; maybe multiple tens of thousands of tons. Each Starship+booster is a few hundred tons, say 500 if you include scrap. And they’re liable to be pumping out at least one per month when they really get going. That should settle down eventually given that it’s supposed to be fully reusable, but for the first few years they’ll be going through a lot of prototypes, and probably a lot of vehicles that don’t see too many uses. If Musk is serious about pumping billions into a Mars effort, there will also be a lot of Starships sent on a one-way mission (to build up propellant production and other infrastructure).
I don’t know much about steel mills but it looks like the smaller ones produce on the order of 1000 tons/day. A custom order of several thousand tons doesn’t seem unreasonable, particularly since it’ll still be in the 300-series family.
There’s also the matter of sheet thickness. They’ll want to dial that in very precisely to keep the weight down, and probably have a different thickness for each ring segment. That’s a somewhat longer-term thing, though.
But if they’re constantly tweaking their alloy, then each of those prototypes would be a different one, and so you’re not getting any economy of scale from the prototypes.
Minimills mainly produce construction grade carbon steels primarily using scrap steel and direct reduced iron. Because of the variable alloying content and large amount of carbon, they aren’t well suited to the production of low carbon stainless steel. There are specialty metal mills that can produce any alloy you can dream up but they charge a premium for such bespoke production, and they generally produce ingots or castings rather than large coils or rolled structural sections.
A mill run of steel from a large mill runs several tens of thousands of tons in heats of a few hundred tons each delivered throughout the year; the only companies that can actually command mill runs of particular metals are the major car manufacturers, a couple of the large construction equipment manufacturers like Case and Caterpillar, and steel distributors who try to forecast industry needs a couple of years in advance. The cost of a mill run of stainless steel would be several hundred million dollars, although most of that cost would be spread out over production and paid on net payment terms, so as long as you can actually use that amount I guess you wouldn’t accrue the the inventory costs though it is unclear to me how the costs of this proposed high volume of manufacture are going to covered by some kind of income from a hypothetical need for a fleet of hundreds of super heavy lift launchers.
The need to produce a specialty stainless steel alloy is still unclear. There are dozens of 300-series and 400-series steels as well as duplex steels and other industry-specific formulations made for a wide variety of applications from cryogenic service to hot corrosive environments, and it isn’t as if some slight tweaking of constituents is going to produce a magic steel with dramatic improvements in high temperature properties. The existing selection of alloys would seem to cover the array of applications. I could understand wanting to purchase an in-between gauge size for some highly weight optimized applications but you can get grades like 316L and 347 in standard sheet gauges from 0.0125 in (30 gauge) to 0.172 in (8 gauge) and plate in 0.125 in increments from 1/4” to 1” and 0.25 in increments from 1” to 2”, which would seem sufficiently optimized for this kind of application. Spending the cost and effort on bespoke steel production seems entirely contrary to the notion of keeping production costs down by using commercially available grades and sizes. The Bob Truax-proposed Sea Dragon, for instance, used marine grades of steel in standard thicknesses and actually built in shipyards to industry tolerances, which is how they got down to an estimated cost of around $100/lbm to LEO. Spending two or three times the cost for a marginally better optimized steel drives cost upward, and even though fabrication isn’t actually the main cost driver in launch costs it is a large portion of the capital resource budget, albeit amortized over the lifetime of a reusable vehicle, provided each vehicle can be reused many times without major rework or repair, which is a capability that is still not demonstrated.
Stranger
Actually, I can think of one sense in which they could be “tweaking the composition” while still remaining mostly commercial-off-the-shelf: If each individual component is one standard alloy, but which alloy is used for each component varies. So they might say that a particular piece of the structure gets exposed to slightly less heat, so for that piece, they’ll use a slightly less heat-tolerant alloy (that has some other offsetting advantage, like greater strength or lower cost).
Whether that’s actually what they meant, I don’t know.
That is what you would do, anyway. But it isn’t as if the high temperature properties of stainless steels vary by all that much or the thermal environment is so precisely characterized that you can pick or concoct so exactly optimized alloy; you basically look at the materials that can withstand that service environment and then pick one that meets other requirements like tensile and shear strength, fatigue strengths, mimimim elongation, creep and corrosive resistance, SCC class, weldability, machinability, et cetera.
Stranger
There are a lot of alloy mills making speciality stainless and alloy steels for chemical reactors, nuclear calendrias, gas turbine blades and buckets, etc etc.
Come to think of it, the metal for space ships is not that special. We’ve had reactors where the alloy steel tubes are cast centrifugally so that the impurities accumulate in the center to be machined off later.
The point is that there are mills which will take large or small orders.
In general, mills do not “take orders” from individual customers whom are purchasing less than a mill run of material because production schedules are drawn up months (for large mills even a couple of years) in advance. The vast majority of manufacturers purchase their raw through one of the many metal suppliers who stock most or all of the common grades of aluminum and steel, and coordinate the need for specialty materials among fabricators and manufacturers to do large bulk purchases and continuous supply agreements to minimize excess inventory.
Starting a new run of a specialty or bespoke alloy is not as easy as just programming a CNC machine; it requires ordering and checking all of the constituents for quality, doing a pilot run and making tweaks to ensure consistency in the final product, and then setting up a separate inventory space so it doesn’t get mixed in with other products, as well as doing all the testing and certification to assure the material meets the required specification. It is a major undertaking and not something a mill can do in any volume without a lot of planning and preparation.
There are, as noted above, specialty mills that will produce any alloy that can be made, but not in large quantities or continuous production and they charge a massive premium which is why many exotic alloys used in aerospace and nuclear power reactors have extraordinary costs per pound compared to regular grades of steel, aluminum, and titanium alloys. The economics and physical requirements of metals production just doesn’t allow for fast changes in production or serving customers who want an unusual formulation at an economical price.
Stranger
I’d call materials a slightly different category than fabrication. SpaceX clearly has put a significant focus on low-overhead fabrication, and is already pumping out prototypes at a significant rate. The contrast with the Sea Dragon is apt, since their facility in Boca Chica looks a lot more like a shipyard to me than a typical “rocket factory”. Not just because it’s largely out in the open, and with a somewhat casual approach to materials handling, but also their basic technique: fabricating lots of identical or near-identical rings and stacking them together. Big ships are built in segments, too, and I suspect the Sea Dragon would have also used that approach. Of course, most rockets are segmented to some degree, but not necessarily with the same division of labor that allows parallel optimization.
At any rate, I mentioned above that I expect any gauge optimization to be more of a long-term thing. I expect you’re right that they can get most of the way there via standard gauges. The design has a fair amount of margin in it and early flights don’t need to hit their 100-150 t targets. It’ll be more important for propellant tanker flights since that will dominate once they get going with Mars (5-10 tanker flights for each cargo flight).
Another consideration here is that the Tesla Cybertruck will need enormous amounts of stainless. The gauge and alloys will be different, and of course Tesla is a different company from SpaceX, but the volume from that may give Musk a bit of extra clout with prospective suppliers.
Should be an interesting year in any case. We’ll see tests of the Super Heavy booster, more high-altitude Starship tests, and maybe even orbit if they’re very lucky. Not out of the question that they beat SLS to orbit…
In cost estimation, raw material costs are included in the fabrication budget, and having to deal with a bunch of different thickness and alloy materials is surprisingly expensive because of all the additional inventory management and handling it entails. (On the flip side, I once worked for a manufacturer that wanted to communize all plate materials to three standard thicknesses and have only a dozen different sizes and lengths of fastener across all product lines, which produced a lot of scrap and wasted engineering effort before that effort got shut down.) Fab costs, including materials, are generally less than 20% of a total launch cost and with common metals instead of aerospace-grade composites should be even cheaper, which is why reusability in and of itself is not the cost saver that people intuit it to be,
I would not bet any amount of money on an initial launch capability (ILC) for the Space Launch System (or indeed, that it will every see ore than a handful of launches over a period f a decade), but Elon Musk is notorious for underestimating delivery on schedule, often by years. And as much as I’d like to see the capability that a superheavy space launch vehicle would offer I’m still not seeing how any of this is being paid for. Musk’s ventures of SpaceX and Tesla have enormous valuations, of course, but those literally incredible numbers are a house of cards; Tesla couldn’t make the profit to justify its current valuation in decades of even high volume production, and while the profitability of SpaceX is unknown to the public it cannot be nearly enough to cover this effort, nor is there a rapturous demand for a huge volume of 100+ ton payloads to orbit now or in the foreseeable financial future.
Stranger