Explosion during Starship tank tests


Readers!
 
Scroll down to read this post.
 
For many reasons, mostly political but partly ethical, I do not use Google, Facebook, Twitter. They practice corrupt business policies, while targeting conservative websites for censoring, facts repeatedly confirmed by news stories and by my sense that Facebook has taken action to prevent my readers from recommending Behind the Black to their friends.
 
Thus, I must have your direct support to keep this webpage alive. Not only does the money pay the bills, it gives me the freedom to speak honestly about science and culture, instead of being forced to write it as others demand.

 

Please consider donating by giving either a one-time contribution or a regular subscription, as outlined in the tip jar below.


 

Regular readers can support Behind The Black with a contribution via paypal:

Or with a subscription with regular donations from your Paypal or credit card account:


If Paypal doesn't work for you, you can support Behind The Black directly by sending your donation by check, payable to Robert Zimmerman, to
 
Behind The Black
c/o Robert Zimmerman
P.O.Box 1262
Cortaro, AZ 85652

 

You can also support me by buying one of my books, as noted in the boxes interspersed throughout the webpage. And if you buy the books through the ebookit links, I get a larger cut and I get it sooner.

During tests today of SpaceX’s Starship Mk1 test prototype there was a sudden explosion, damaging the spacecraft. Below is a clip from today’s live stream showing the explosion.

They had been doing a variety of tank and venting tests for the past day. (The link includes a video showing the first pressurization test yesterday.)

The Mk1 (Mark 1) was being built with an initial hoped-for schedule targeting the first orbital flights by next year. No one took that seriously, and today’s incident reinforces that skepticism. At the same time, SpaceX has routinely recovered very quickly from its engineering test failures, treating them as opportunities for improving their designs. It is for this reason that most knowledgeable observers of the company also expected any delays to that target schedule to never be very extended.

Moreover, earlier this week SpaceX revealed that it had already decided not to fly this version and proceed to their next version. A Mk2 version is being built in Florida, so in Boca Chica they will proceed to the Mk3, with the goal to get to build toward the final operational version, what Musk labeled the Mk5. Based on past SpaceX policy, however, expect them to begin commercial flights with the earlier versions as they upgrade to the Mk5.

UPDATE: It appears the explosion occurred during a maximum pressurization test, which means they now have data telling them the limits of their tank design.

Share

32 comments

  • Questioner

    We have seen Mk-1’s methane tank filled with liquid nitrogen burst during the pressurization and filling test. My guess: bad welds, not sufficient quality control.

  • Diane Wilson

    From discussion at NasaSpaceFlight.com, this was a maximum pressurization test with LN2. No real guess at causes at this point. Could be a bad weld, or corrosion, or something else. In any case, the top bulkhead blew off, and it appears the failure started in the top ring, where the bulkhead was joined. Lots of backwash, with gas blowing out the bottom of the stage as well. Note that neither the fairing/nose nor the Raptors were attached.

    Elon tweeted that they had already decided not to fly Mk 1, and would proceed immediately to build Mk 3. No word on whether the Mk 1 fairing would be used for Mk 3, or if that would also be scrapped. No word yet on how this affects Mk 2 in Florida.

  • MDN

    What I love about SpaceX is that they are not afraid of failure and understand that to actually make rapid progress and achieve maximum efficiencies you have to accept that occasionally tests will not go right. Since this was a max pressure test it’s not at all surprising they suffered a failure as they are trying to achieve the lightest structure possible while still meeting their spec., and unless you want to accept over-designed excess you start aggressive knowing that you can scale up later if necessary.

    You can bet that Boeing or ArianeSpace use the exact opposite approach and due to this their designs are OVER BUILT and excess weight on every flight because their organizations would crap their pants at a failure like this. Musk will simply beef up what failed and when the final design is locked it will be as robust as it needs to be, but as little over that as is reasonably feasible too (which only this methodology makes possible).

    And if the burst pressure was high enough today, they may simply decide to stand pat anyway accepting that a 147% safety margin was close enough to a 150% design goal or something like that. I remember reading a story of Lockheed making a decision like that in regard to a fuselage pressure test under Kelly Johnson’s direction. The first test blew out a bulkhead and they re-designed and beefed things up a bit. And during the second test when they got within 5% or so of the test target and the air frame started groaning and creaking its objections Kelly made the call that it was good enough to cut the test short and call it good vs. pushing his luck. As the chief engineer he knew he had already achieved the safety margin he needed and wasn’t going to blow up his schedule and budget just to hit a paper target that he knew he didn’t need to.

    Elon Musk thinks and works the same way. That’s the advantage of having corporate leadership that actually understands the technology of the company they are running.

  • Questioner

    I recall that the failed upper bulkhead of the methane tank, once already welded to the tank cylinder, was again cut off from the cylinder completely and then rewelded to it about four weeks ago. Why this happened is unclear and, as far as I know, there was no explanation in the special forums. The new weld must have been very weak, because such a tank would burst or fail differently. He would rather tear open like a sausage on the long side and not axially into two pieces, if all the welds had been the same quality. The quality of craftsmanship in the construction of Mk1 did not make the best visual impression and was often a point of critics and discussion. It may be that the whole construction concept in the open and without significant rigs, have to be revisited by SpaceX.

  • Questioner

    Mr. Z .:

    Do you have actually a direct evidence that SpaceX has actually announced the abandonment of plans for a test flight some time before the explosion of Mk1?

  • Diane Wilson

    Questioner:

    SpaceX did not announce the change in flight plans for Mk 1 before the explosion. This was “announced” in a tweet from Elon following the explosion. That was a surprise, but it’s in line with the way SpaceX operates – fail quickly and learn from failures.

  • Questioner

    Diane Wilson:

    Yes, it seems so.

    To all:

    Brian Peterson wrote at NSF 10/13/19 about his observation regarding removal of upper bulkhead:

    “Are they going to remove the top right and bulkhead? From BocaChicaGal’s latest images it looks like they’ve ground away the welds holding the top ring and build head on.”

    https://forum.nasaspaceflight.com/index.php?action=dlattach;topic=49114.0;attach=1588206;image

    Look at the appearance of the tank after brutal removal of the weld. It does not look very pretty. I am questioning that they had achieved again a good re-welding of the bulkhead to the tank cylinder afterwards. That may caused yesterday’s failure.

  • Diane Wilson

    It’s worth noting that the Mk 2 top bulkhead was attached in a completely different manner from Mk 1. For Mk 1, the bulkhead was attached to the bottom of the top ring, on the ground, and the entire assembly was then lifted onto the stack and welded into place. For Mk 2, the top ring was already attached to the stack, and the bulkhead was added (earlier this week) on top of the stack and welded to the top of the top ring. As an observer, this seems more reliable.

    Second, in all the talk of going straight to Mk 3, there has been little discussion of Mk 2. Supposedly, the “flight design” has changed a lot from Mk 1, which is the reason for going full-speed ahead on Mk 3. Will Mk 2 be another non-flight test model? Some reasonable suggestions that Musk wants to get to at least a static fire with a test article. Would that be Mk 2?

    Other things floating through NSF this morning… The break appears to occur in mid-plate, not at a weld (at least, not an externally visible weld). This is captured on video. Also, there was no venting before the explosion. There is some suggestion of an extreme overpressure, which would implicate ground support equipment, or sensors in the tank, or any number of other possibilities.

    Again, a lot of speculation and not as much information at this point, except to point out that this is how SpaceX works. They needed four attempts to get Falcon 1 launched. There were something like 10 attempts to land a Falcon 9, something that was considered fantasy at the time. They’ve tested, failed, and refined their fairing retrieval, and finally succeeded in catching and reusing fairings. The Raptor engines have had failures on the test stand. They learn from failure, and move on. “Fail forward.”

  • Ray Van Dune

    Yes it does appear that the failure occurred in mid-ring between two horizontal ring welds. I could not say if it was at a vertical weld, but I assume image processing will reveal that. I note the crane next to the ship seems well positioned to carry cameras to record any upper bulkhead failures. Could this have been a planned test-to-destruction?

  • Questioner

    Diane Wilson:

    When you have followed the whole construction and the associated effort, you feel with the workers, who put so much work into that welded construction. SpaceX should may have built only one or two tanks for such tests at first and not the whole rocket, including all other equipment. The pressure and the step was too large.

    Thank you for your hint regarding Mk2 status.

  • Diane Wilson

    Starship’s “skin” is also the side wall of the tanks. There were no separate tanks to be tested separately from the rocket.

  • Questioner

    Diane Wilson:

    I know that. :-) You misunderstood me. My failure. I meant it might have been better if SpaceX had just built only the major tank section of the vehicle ( = methane tank + LOX tank combined) for pressure and leak tests without the rest of the vehicle (nose-cone, flaps, legs and soon). It is not important, SpaceX can select between different routes. However, I believe, SpaceX should build its Starships in a building. NASA’s VAB, if available would be a great choice.

  • Diane Wilson

    Questioner:

    OK, yes, I did misunderstand. However, that’s just about what they did. The fairing and nosecone is a separate assembly, and it’s still sitting, intact, in the construction yard. The flaps were attached, but there are two considerations there. First, the flaps are heavy, and do exert some pull on the structure. Second, much of the plumbing and control circuitry runs in exterior connections, and these run “under” the flap hinges and enter the rocket interior at that point. Testing these fuel lines is part of the testing for this section of the starship, if not in pressurization tests, then certainly in the next tests to be run with LOX and LCH4.

  • Star Bird

    So just how soon will be able to send the Democrats to Mars to live maybe Marvin and K-9 will release a few of those Instant Martians(Just add Water)for a little action

  • Questioner

    Diane Wilson:

    Yes, you are right. It had a large value to build the whole and to investigate the interaction of different subsystems with the tankage/ship structure. Obviously, they already have learned now a lot when building this Mk1 thing, though it was far from perfect in terms of craftsmanship, sheet metal design, and welds. I wonder what plans SpaceX now has for Mk2, which looks much better built. As you said, the top bulkhead is attached to the main tank cylinder in different ways. See link below.

    https://twitter.com/julia_bergeron/status/1196495171468374019/photo/1?ref_src=twsrc%5Etfw%7Ctwcamp%5Etweetembed%7Ctwterm%5E1196495171468374019&ref_url=https%3A%2F%2Fforum.nasaspaceflight.com%2Findex.php%3Ftopic%3D48892.480

  • Edward

    Ray Van Dune asked: “Could this have been a planned test-to-destruction?

    Had it been, Musk would have said so in his tweets. I learned the hard way to announce in advance any tests to destruction. Damaged test articles make everyone worry — management, customer, colleagues, etc. In this case, the entire public.

    Questioner suggested: “it might have been better if SpaceX had just built only the major tank section of the vehicle ( = methane tank + LOX tank combined) for pressure and leak tests without the rest of the vehicle (nose-cone, flaps, legs and soon).

    Perhaps, but with this one test unit, they learned quite a bit that they would not have, had they only built a large test tank. It is becoming clear that SpaceX’s development style is to put a lot of effort into learning as much as they can from each test unit. They are perfectly willing to risk some hardware in order to get the lessons needed as fast as possible.

    Part of that hardware risk is not just in the possible destruction of the unit but in the obsolescence of it. MK1 was clearly going to be used for a test flight, but now they seem to be skipping even MK2 for a test flight and going straight for MK3. What did they learn in the past seven weeks that made MK2 obsolete as a flight test unit?

    Apparently quite a bit.

    I believe, SpaceX should build its Starships in a building. NASA’s VAB, if available would be a great choice.

    Early on, SpaceX made it clear that they were doing construction in the open because no building was yet available, and they wanted to begin testing and learning sooner than it would take to get a building ready. As I recall, SpaceX is working on constructing something in which they can build Starships and Super Heavies safe from the elements.

  • Dick Eagleson

    SpaceX says the Mk1 pressure test failure will not much delay Starship’s development and I’m inclined to agree. SpaceX was able to recover from the destruction of Crew Dragon 2 capsule C201 in just a few months, even though the cause was exotic and had to be extensively investigated, because it had several vehicles under construction behind C201 and could designate a new one to take C201’s place fairly expeditiously. Similarly, there are three other Starship prototypes in various stages of construction.

    The cause of Mk1’s cryogenic pressure test failure isn’t going to take nearly as long to autopsy in detail as did C201. The only mission Mk1 was slated to fly was the 20 km. “hop” intended to allow testing of the terminal “aerial ballet” landing sequence. SpaceX could still assign this mission to Mk2, nearing completion in FL- assuming it passes its own pressure tests and other pre-flight test procedures. It could also choose to wait until Mk3 and/or Mk4 – the six-engine versions that will be much closer to final operational configuration – are completed. Both are in early stages of construction in TX and FL.

    There is even the possibility of quickly running up just Mk3’s tankage and engine section, perhaps using salvaged mechanicals and rear “wings” from Mk1, and mating it to the nearly-completed upper half of Mk1 to make a “Mk2.5” that, perhaps, could fly the 20 km. landing test hop earlier than otherwise possible, though I think this is a long shot. I suspect a clear decision on this issue will await further progress on all three vehicles. Perhaps plans will firm up by, or even before, year’s end.

    As for eventually building Super Heavies/Starships indoors, I can think of only two reasons SpaceX might choose to do that:

    (1) Investigation into the Mk1 failure fingers the circumstances of construction as a material cause of the failure, or

    (2) SpaceX comes to regard some portion of Super Heavy’s or Starship’s assembly process as a trade secret to be kept from prying eyes.

    I suspect neither reason will prove to apply. SpaceX, recall, was willing to start construction tests of large prototype carbon fiber composite SHS pieces in a plastic tent on the L.A. waterfront. Carbon fiber composite is a notably fussier material with which to work than is rolled stainless and the L.A. waterfront is certainly no better an environment in which to do major spacecraft construction than is Boca Chica, TX or Cocoa, FL.

    It should also be pointed out that steel naval and commercial vessels have been built outdoors for well over a century and do not routinely come apart in service as a result. With Elon already teasing a potential SHS successor of double the current 9 meter diameter, outdoor construction seems likely to just become standard operating procedure for future large spacecraft.

  • Questioner

    Dick Eagleson:

    I believe SpaceX has found some significant flaws in Mk1’s design by means of the pressurization and filling tests. Therefore, they have already considered giving up the test flight with this vehicle. The test failure has now created only fait accompli.

    I also like the analogy to shipbuilding or even building construction in the open, which I myself introduced here some time ago. But maybe this analogy has its limits. After all, the challenges for building thin-walled pressure tanks and other similar strutural components for a launch vehicle are much higher than for a large ship. The difference in the thickness of the steel material alone is considerable. The sheet thickness of Starship Mk1 is about 3.5 mm, Mk2 even less, as SpaceX indicates. While a large ship 9 meters wide and 50 meters long is more likely to have a “plate thickness” of estimated 15-20 mm. Please correct me if you know better.

    And is only a technical feature. There are a number of others that suggest that shipbuilding is easier. BTW, there is the situation that today often only finished, even fully equipped ship sections are welded together in dry dock. It would probably be better to build Starship and SuperHeavy lying instead of standing. Simply because then much lower height of the workers must be overcome. So I have to withdraw my suggestion to use the VAB. [Remark: However, this process requires significant jigs and rigs.]

  • Questioner

    One of the most important personal qualities of Elon Musk is his consistent behavior. He never loses much time to make corrections, even if it costs a lot of money. Rings for Starship Mk3 are abandoned and now scrapped, although those rings (23 pieces), which were already ready at the SpaceX’s Florida site for installation, exhibit an improved design compared to rings for Mk1 and Mk2 (only one vertical weld and thinner sheet metal).

    They have already scrapped on 21 November 5 rings, as a horrified SpaceX fan from the NSF quite excited writes. They already started on Monday (following this report) with these actions (before the explosion of Mk1!). If that’s true, I interpret this behavior in a way that SpaceX has a much better design for the production of Mk3 and following spaceships, what makes the current technical solution obsolete.

  • V-Man

    Questioner:

    Elon is smart. You can always have more money. You can’t have more time.

  • Questioner

    V-Man:

    Correct!

    I believe that the most likely scenario which caused the bursting of the methane tank of Starship Mk1 was the combination of two factors:
    1. After deciding not to let Mk1 fly, they ventured to put more pressure on the tanks than originally planned. (However, I assume that you did not expect a tank rupture, which demolished the whole vehicle at the end).
    2. A weld gap well over 10 mm between the upper tank bulkhead (which includes a short cylindrical ring piece) and the main cylinder of the tank had to be bridged over significant portion of tank circumference. This can be proven by the corresponding photo. The resulting weld was not as strong as if it had been welded with ideal 3 mm gap. This has been obviously accepted.

    BTW, maybe the bad weld was one of several reasons why the test flight for Mk1 was abandoned.

  • Edward

    Questioner wrote: “Please correct me if you know better.

    15 to 22 mm is too thick and far too heavy. The 747 has aluminum skin that is about 2.2 mm (~0.09″). Steel, being stronger than aluminum, would be able to be thinner.

    Starship-Super Heavy together are expected to be about 11 million pounds with propellant and payload, and if we assume a steel with yield strength of 0.1 million pounds per square inch, a diameter of 30 feet (9 meters), and maximum launch acceleration of 4 G (Saturn V had this maximum), then the skin thickness at the bottom of the lower Super Heavy tank could be about 0.2 inches (~5 mm). Although I have not taken into account a safety factor, I have also included the weight of the many engines on Super Heavy (~ 0.1 million pounds) and the weight of the propellant in the lower tank (3 to 4 million pounds), both of which would not be part of the load on that tank’s skin. In addition, by the time the Super Heavy reaches maximum acceleration, much of the 3 to 4 million pounds of propellant of the upper tank will be depleted.

    The skin can get thinner the farther up the rocket we get or the lower the maximum launch acceleration.

    Going up the rocket, by the time we get to Starship the weight is about 1/4, about 3 million pounds, so the skin thickness at the bottom of the lower tank could be about 0.05 inches (~1.3 mm). Again, I did not subtract out Starship’s engines and lower tank propellant, but the 4 Gs of the Super Heavy would apply to the total amount of propellant in the upper tank, and I assume the same steel strength.

    Steels can be stronger than 100,000 pounds per square inch, especially through tempering, but tempered materials may not remain that strong after reentry. The Super Heavy may not get hot enough to affect the tempering, so perhaps a stronger steel could be used there, resulting in an even thinner skin.

  • Questioner

    Edward:

    Here is a misunderstanding, surely my bad description is responsible for it. It’s about the analogy between building a big marine ship/vessel and Musk’s Starship. My estimate for the wall thickness of 15-22 mm refers to the marine vessel. BTW, we know for sure from facts, which were discussed at NSF forum, that Starship’s Mk1 metal sheet thickness is about 3 mm. Musk said that metal sheet thickness of Mk2 is lower in order to save weight.

  • pzatchok

    That sheet metal thickness is for the outer body stainless skin.
    As far as I know it is not intended to be load bearing either so it can be as thin as the air will allow.

  • Edward

    Dick Eagleson wrote: “SpaceX, recall, was willing to start construction tests of large prototype carbon fiber composite SHS pieces in a plastic tent on the L.A. waterfront.

    This is probably what I am remembering.

    Open air steel construction is not only for ships but also for the steel beams and girders of buildings.

    It has been “traditional” for spacecraft to be built inside cleanrooms in order to assure control over contamination. Among the many contamination concerns is the cleanliness of thermal surfaces, which are pretty much every exposed surface on the body of a spacecraft. These surfaces either face something hot, such as the sun or Earth, or face the cold of deep space. Either way, thermal engineers work hard to keep balanced the amount of heat absorbed, emitted, or generated internally. Contaminated surfaces can change the absorption or emission properties, messing up careful calculations on how to keep these in balance. Building Starship or Super Heavy outdoors should horrify thermal and contamination-control engineers in the space business.

    So what to do with a spacecraft that will spend virtually its entire between-missions life out in the elements? Design it so that open air contamination does not much affect the thermal balance, of course. Designed with this in mind, how necessary is it to build the spacecraft inside an environmentally controlled cleanroom?

  • wayne

    pzatchok–
    excellent point

  • wayne

    tangentially—

    this looks familiar….

    Amazing Stories
    April 1947
    “All aboard for the Moon”
    https://archive.org/details/Amazing_Stories_v21n04_1947-04.Ziff-Daviscape1736

  • Questioner

    The processing of a 1/8 inch thick metal sheet (eg bending), which exhibit normal dimensions, by a craftsman requires some definitive effort. What’s interesting is what you can see in the following video: The apparent thickness and sturdiness of the sheet metal is lost and turns into the laxity of a piece of paper due to large dimensions of Starship. The upper bulkhead (shown in the linked video), which was freed from Mk1 by that burst test has a diameter of nearly 30 feet.

    https://forum.nasaspaceflight.com/index.php?topic=48895.760#lastPost

    Pzatchok:

    There is special no outer and no special inner body stainless skin. “Starship” has only a single wall (“skin”) made from single walled sheet metal, aside from later payload/fairing section and bolted thermal protection tiles at windward side.

  • Questioner

    Do they try a repair (Mk1)?

    They are working on a new bulkhead. I assumed that the damage was so great that it was not worth repairing?

    https://forum.nasaspaceflight.com/index.php?topic=48895.780

  • Edward

    Questioner wrote: “The apparent thickness and sturdiness of the sheet metal is lost and turns into the laxity of a piece of paper due to large dimensions of Starship.

    The geometry (shape) of the paper-like sheet metal has a lot to do with how much it will flap around. The edges of the bulkhead moved quite a bit when loose, but when assembled to the rest of the rocket then the same edges have significantly less flexibility.

    Pressurizing the propellant tanks provides a large amount of stiffness, too. It is similar to crushing a soda pop can when it is empty vs. before it is opened.

    There is only a single wall on a rocket, where the propellant tank walls are also the “outer” structural wall. This saves a tremendous amount of weight. A major disadvantage to a single wall is that cryogenic propellants boil off quicker. However, before launch it is relatively easy to continue topping off the tanks, so boil off is less of a problem than the weight of a second wall would be.

    They are working on a new bulkhead. I assumed that the damage was so great that it was not worth repairing?

    Questioner’s “#lastPost” link gives a comment: “From clearing up Mk1’s flying bulkhead to machining the monolithic steel rings for what is likely to become Mk3.” This indicates to me that part or all of MK3 is being constructed at Boca Chica. The new bulkhead may be for MK3.

  • Questioner

    Edward:

    It looks like Mk3 is built in other (reverse) way than Mk1. This time they get started with the bulkhead and followed by the most upper ring underneath and the whole thing is pushed up, keeping the current weld in work down at ground level, which simplify automated welding. Just like as in this video.

    https://www.youtube.com/watch?time_continue=1&v=w-ARF26P2rI&feature=emb_logo

Leave a Reply

Your email address will not be published. Required fields are marked *