SpaceX Falcon 9 - AMOS-6 - (Pad Failure) - DISCUSSION THREAD (2)

[Jim]

This is the kind of thinking that propels buggy whip companies all over the world into collapse. Seriously? The way we do it is the best way possible? Are you even listening to yourself? There is always a better way.

Not with chemical rockets.

[M.E.T.]

This is the kind of thinking that propels buggy whip companies all over the world into collapse. Seriously? The way we do it is the best way possible? Are you even listening to yourself? There is always a better way.

Theoretically there is always a better way. Practically there are constraints (we don't have time for the best way, we need to make money so take what we have, ...) and what is much more important: There are so many variables and parameters in such a complex system like a rocket that no one can exactly predict what will turn out as the best way.

Putting a copv into lox? Saving weight, increasing performance, increasing risk? No one could have predicted that this event would happen. But everyone can predict that as newer the design, as higher the innovation, as higher the rate of change, as higher the risk will be that something goes wrong. In the aftermath we are always more clever: no, copv in lox didn't pay out. Whatever the performance win was, it's eaten up by the cost off one rocket and one satellite and all the other damage done.

I know, it's boring to say there are lessons learned in the last decades about complex systems, the rate of change, the probability of loss of mission, etc. and how often a rocket failed because of a small failure or change or .... But still the same lessons seems to be (at least to me) valid. SpaceX is driving fast. Looking at their launch manifest for next year on wikipedia with 31 launches (no simple prediction, really real rockets with real customers) I would say this can't work.

To clarify, what is it that you are saying can't work? The fueling method they were using (albeit revised to address the specific root cause of the last failure)?

Or the increased launch cadence in general that they are pursuing? Before the failure they were achieving launches about every 3 weeks, at their shortest turnaround time.

If they can resume that rate, they are already on about 18 launches per year. If they can do a launch every 3 weeks in the first half of the year and ramp it up to a launch every 2 weeks from mid-year onwards, they could get to around 21 launches in 2017.

And if reusability kicks in properly by mid next year, that could potentially even be exceeded.

So 31 launches in 2017 - probably not. But I would think between 20 and 24 are probably achievable. With the rate in 2018 probably reaching 30 launches or more, (barring any further accidents of course).

[Jim]

Or the increased launch cadence in general that they are pursuing? Before the failure they were achieving launches about every 3 weeks, at their shortest turnaround time.

They were more like surges than sustained cadence.

[MikeAtkinson]

Well then can the 2nd stage be stretched?

The rocket equation shows that there is a point at which a heavier second stage starts to reduce payload, especially to higher energy orbits.  Falcon 9 v1.2 may already be using close to the optimum sized second stage. - Ed Kyle

Welcome to the forum. Ed is correct. Let me try to put this in simpler terms for someone who may not have studied physics or rocket science. By making the 2nd stage larger, you make it heavier. Now the first stage engines are struggling to push that heavier second stage up, burning and wasting a lot of stage 1's fuel and oxidizer. The first stage burns up all that propellant at a lower velocity and altitude than it normally would have. Now the second stage kicks in. Its engine is optimized to operate in a vacuum by having a huge expansion nozzle, but due to the lower altitude, it is now firing with some atmospheric pressure pushing back against its escaping exhaust gasses. This reduces its efficiency. Also, the weight of the second stage burdens its one engine even more than it did the first stage engines. You wind up getting less mass into space than you would have had you just left the design the way it was. Upper stages need to be designed to match what the first stage can do. Too big or too small and you don't get optimum performance. There is a sweet spot in upper stage design that matches what the first stage can provide.

fphowell, welcome to the forum.

Your question has subtleties, over and above the above answers. Ed is probably correct in the the current stage is near optimum. But optimum for what? The optimum for a large payload to LEO and a smaller payload to GTO are different, so are the optimums for expendable, barge and RTLS reusables, the optimum for FH to various destinations in its various possible reusable modes are different again. We don't know which SpaceX has designed for, or perhaps some compromise between the various missions. The GTO performance seems the most important and perhaps SpaceX have optimised for that.

It should be noted that optimum can mean different things, it could be optimised for cost, performance, reliability or something else.

We know that SpaceX are moving to max thrust Merlin 1D. Normally if thrust increases by x% then stages should also be increased by about x% to give maximum performance, all other things being equal. But  we have been told that the 1st stage cannot be extended. So an increase in the 2nd stage of more than x% is likely optimum.

We also know that SpaceX are enhancing the 1st stage of reusability (block 5), this is probably going to reduce its performance slightly, but then other enhancements (e.g. lighter legs?) will improve its performance. Until we see what SpaceX have done it will be difficult to determine which of these two factors is dominant.

It may be that SpaceX will move to two slightly different upper stages, one optimised for GTO satellites and another more robust one for larger payloads to LEO. This has disadvantages in multiple configurations being manufactured, but might be worth it if payload to GTO can be increased significantly. Any differences would be subtle and possibly not easily visible.

With 1st stage reuse, 2nd stage and launch ops become the dominant costs, it could be that SpaceX decide to reduce those even if it leads to a non-optimum 2nd stage in performance terms.

[M.E.T.]

So what was limiting their cadence during 2016 (apart from the 2 month grounding)?

Production rate of cores? Availability of launch facilities? Something else?

And are they not addressing these bottlenecks? Their entire business model is based on volume, after all. So unless the cadence constraints are addressed, their business model is flawed. Hence, they will be giving it a lot of attention, as Shotwell has repeatedly stated.

Before the failure they were heading to around 15 launches this year. As it stands they will likely end 2016 on 10 successful launches and one pad failure.

[Comga]

Well then can the 2nd stage be stretched?

The rocket equation shows that there is a point at which a heavier second stage starts to reduce payload, especially to higher energy orbits.  Falcon 9 v1.2 may already be using close to the optimum sized second stage.

Abandoning first stage recovery would allow better design margins on the stages, especially on the second stage.  It might be possible then to abandon the entire extra cooling idea that imposes the hazardous late-loading restrictions.

 - Ed Kyle

Much as this exchange and the replies are informative, they don't have much to do with the Amos-6, or any specific mission.  It is more appropriate for some thread in the General section or the L2 Modeling section. Comments like Ed's are also found and analyzed in general discussions in the Reuseable Rocket section.

Whether Jim is right or wrong that there is no better way than what has been done, subcooled LOX and cooled RP1 is what SpaceX is doing and what seems to have caused the Amos-6 failure.   Can we keep this thread on that specific failure discussion unless and until SpaceX concludes that they will turn back from subcooling?


Ditto for discussion of the number of launches in 2017 or what limits their launch pace, unless the failure investigation says they have to take extra days or weeks per launch to be safe.  We will have a poll on 2017 launches next month, so hold on for that.

[RDoc]

Liquid helium? I thought it was cold gaseous He at high pressure.

That was assumed to be the case. But it has been suspected for some time now that SpaceX has been messing around with partially liquified helium in loading their COPV's to prevent having a large, and potentially problematic, thermal gradient between the LOX touching the outside of the COPV's and the gaseous Helium inside the COPV's.

I'd thought that they were just chilling the He so that after expansion it wound up at the LOX temp. If they were loading liquid He, that does sound a bit fraught. If nothing else, I'd think it would be extremely complex to characterize all the transients inside the COPV during loading. It seems like maybe it was a bit too complex.

[su27k]

Limits that don't need to be explored:
There is no need to to reduce prop loading times by a few minutes to the absolute minimum
Or push He into the vehicle as fast as possible.
Shave a day off pad flow and risk customer's spacecraft.

I grew up with sci-if that had Pan-Am shuttles making routine flights to space. Which is where SpaceX wants to get to.  Go ask United Airlines if they can throw away a day, or an hour, turning around an airliner.
Yes the envelope does need to be pushed, no today's tech is not good enough.  They failed in a way that no one has ever failed before, and I'm fine with that.  So long as it doesn't happen again.

I wonder if there's an alternative explanation instead of blindly pushing envelopes. It looks like some people, especially in Commercial Crew circle, are worried about the current propellant loading scheme because it doesn't give time for propellant to reach a steady state, for example ISS advisory committee was saying they're worried about "stratification of oxidizer temperature".

It seems to me that one way to move propellant closer to steady state is to load them even faster, this gives less time for initial loaded propellant to warm up and thus ensure propellant state is more homogeneous. I wonder if by pushing propellant load speed SpaceX is actually trying to address a safety concern, instead of creating one just for kicks.

[sdsds]

one way to move propellant closer to steady state is to load them even faster

That is a truly fascinating idea!

Control systems thinking suggests that when we move things quickly we better be confident in our model of the system's dynamics. It can work: SpaceX itself has demonstrated this with high-g descents to zero-v "steady state" landings. Specifically regarding propellant loading, though, there isn't much that can be done about the thermal inertia of the tankage and other structural elements of the vehicle. (The risk of thermal shock damage would cause me to stick with the slow and easy approach. But then again I'm a total coward compared with the fearless gunslingers at SpaceX! ;-) )

[Space Ghost 1962]

Echoing Comga's point above - stay on topic to AMOS-6 anomaly and take theory/SX "shoulda' coulda' woulda' " to appropriate thread.

We've got a very specific issue here about prop loading in the context of a submerged COPV. Mostly a materials/thermodynamics/cryochemestry discussion at this point.

BTW, LHe has various interesting properties including fluid flow and wetting that could play havoc in a subcritical LOX environment as it "boils" (phase changes) to GHe. Lots to talk about there alone.

[Dante80]

It seems to me that one way to move propellant closer to steady state is to load them even faster, this gives less time for initial loaded propellant to warm up and thus ensure propellant state is more homogeneous. I wonder if by pushing propellant load speed SpaceX is actually trying to address a safety concern, instead of creating one just for kicks.

I echoed this on the CC schedule thread, here.

Regarding the use of densified propellants in general, you also have to put this anomaly in context. When the OG-2 and SES-9 campaigns were under way, SpaceX had to work a lot of kinks and bugs both on the rocket and on the pad/GSE due to the apparent correlation between performance and LOX temperature.

Back then, some people (including George Sowers I think) were saying that this was not the way to go, that it was too much of a hassle etc etc.

Then SpaceX - as was inevitable - ironed out the kinks, and those people ate crow. Because the next few campaigns were the fastest, most clock-work and error free that the company ever had. Far better than the v1.1 campaigns, I might add.

And then a second stage exploded on the pad during fueling. And..here we are. Everything points that SpaceX encountered a not studied before failure, tied both to procedure and to the novelty that densified propellants brought to the rocket. They may well iron this out quickly by adapting their procedures, but the fact that they are still treading new grounds with this does not means that they have caught anything and everything.

At least, yet.

[AncientU]

It seems to me that one way to move propellant closer to steady state is to load them even faster, this gives less time for initial loaded propellant to warm up and thus ensure propellant state is more homogeneous. I wonder if by pushing propellant load speed SpaceX is actually trying to address a safety concern, instead of creating one just for kicks.

I echoed this on the CC schedule thread, here.

Regarding the use of densified propellants in general, you also have to put this anomaly in context. When the OG-2 and SES-9 campaigns were under way, SpaceX had to work a lot of kinks and bugs both on the rocket and on the pad/GSE due to the apparent correlation between performance and LOX temperature.

Back then, some people (including George Sowers I think) were saying that this was not the way to go, that it was too much of a hassle etc etc.

Then SpaceX - as was inevitable - ironed out the kinks, and those people ate crow. Because the next few campaigns were the fastest, most clock-work and error free that the company ever had. Far better than the v1.1 campaigns, I might add.

And then a second stage exploded on the pad during fueling. And..here we are. Everything points that SpaceX encountered a not studied before failure, tied both to procedure and to the novelty that densified propellants brought to the rocket. They may well iron this out quickly by adapting their procedures, but the fact that they are still treading new grounds with this does not means that they have caught anything and everything.

At least, yet.

The contrast that you highlight -- pushing the limits vs. too much hassle -- is not just a race to improve turn-around time or improve on schedules.  The need for maximum efficiency is fundamental to reusable rockets.  Each push into the unknown is a complication, a hassle, but each is needed to get the Payload Mass Fraction up to where reuse is achievable... with a substantial, profitable payload mass (such as a GEO satellite).  Setting PMF records as F9 FT is currently doing isn't a frivolity. Record PMF is a requirement.

If you don't push limits, you get... well, just look around the space launch industry.

[Jim]

It seems to me that one way to move propellant closer to steady state is to load them even faster, this gives less time for initial loaded propellant to warm up and thus ensure propellant state is more homogeneous. I wonder if by pushing propellant load speed SpaceX is actually trying to address a safety concern, instead of creating one just for kicks.

No, the term steady state includes the vehicle (tanks, ducts, bottles, components, etc) reaching plateau temps.

[Jim]

The contrast that you highlight -- pushing the limits vs. too much hassle -- is not just a race to improve turn-around time or improve on schedules.  The need for maximum efficiency is fundamental to reusable rockets.  Each push into the unknown is a complication, a hassle, but each is needed to get the Payload Mass Fraction up to where reuse is achievable... with a substantial, profitable payload mass (such as a GEO satellite).  Setting PMF records as F9 FT is currently doing isn't a frivolity. Record PMF is a requirement.

If you don't push limits, you get... well, just look around the space launch industry.

That is talking out both sides of the mouth.   If that were true, they would be using higher ISP propellants and engines.  They don't because those are hassles.  The whole design of the F9 for operations and not launch vehicle performance.  That is why densification is out of family for them.


also, it isn't record PMF.

[spacenut]

It looks like densification works on the first stage, but not on the second.   Can they do both standard liquification and densification?  Or only one?  I guess it depends on the infrastructure.  There might have to be some compromise on this for SpaceX and the result will be lower capability, then move on to Falcon Heavy. 

[mfck]

It looks like densification works on the first stage, but not on the second.   Can they do both standard liquification and densification?  Or only one?  I guess it depends on the infrastructure.  There might have to be some compromise on this for SpaceX and the result will be lower capability, then move on to Falcon Heavy.

No, it does not look like that. It's like saying that driving the cars manually does not work.

[edkyle99]

Here is where operations overrides launch vehicle performance
...

Not to mention first stage recovery systems dry mass and payload performance given up for first stage recovery.

 - Ed Kyle

[TomH]

It looks like densification works on the first stage, but not on the second.   Can they do both standard liquification and densification?  Or only one?  I guess it depends on the infrastructure.  There might have to be some compromise on this for SpaceX and the result will be lower capability, then move on to Falcon Heavy.

No, it does not look like that. It's like saying that driving the cars manually does not work.

If spacenut is making a sweeping generalization (i.e. that densification cannot work on S2), I disagree with him. If he is saying that densification seems not to be working on S2 under the present design and loading procedures, then he is unquestionably correct. IMO, the problem can be rectified. Saying it can't be fixed and giving up on densification altogether for S2 in not necessary.

Spacenut, you should clarify which of these you mean.

[spacenut]

I am under the impression that densification doesn't work well on S2 because of the location of the helium tanks.  I am also under then impression that it actually works on S1 since there have been no failures on S1 other than one engine was shut down early a couple of years ago.   Therefore should it be abandoned for S2 unless it is redesigned or flaws worked out?

[mulp]

I am under the impression that densification doesn't work well on S2 because of the location of the helium tanks.  I am also under then impression that it actually works on S1 since there have been no failures on S1 other than one engine was shut down early a couple of years ago.   Therefore should it be abandoned for S2 unless it is redesigned or flaws worked out?

By that logic, the S2 design was perfect in July 2016.