Space X reusable rocket landing

[artis]

Since they started this I have been somewhat amazed by the ability to not only launch but also peacefully get back a rocket intact and landing on it's vertical axis.
I'll admit I haven't read a ton of material with regards to this so pardon if this has been asked already.

To me it seems that the main "trick" in order to make such a landing possible is to place the center of mass/gravity within the rocket at a specific place (closer to the bottom) so that when the separation happens the rocket falls back to Earth and always tends to go "bottom first" , as this happens one only then needs to make a reliable restartable engine and some other means of position control , but I assume the main point is that the rocket always stays in a stable vertical position even as it falls?

Because if it started to rotate in all directions (apart from around it's own vertical axis) I think it would be impossible to regain control of it?Is my thinking close or am I off?

 

[hutchphd]

Two points to consider:

1. It is possible to balance a very top-heavy item (bowling ball on a stick) from below
2. There are high drag steerable grid fins near the top of the first stage for a purpose

I, too, am greatly impressed by both the skill and the audacity SpaceX engineers. Gives me hope for the future.

 

[artis]

Well without knowing it's hard to say , and also I would think that in a rocket the center of mass is at neither end but rather somewhere around the physical center maybe, also the fuel capacity and it's changes probably play an effect in all of this, probably a large one as the fuel is among the heaviest part of the rocket.

Anyway it would be interesting to hear more about this.

 

[anorlunda]

also the fuel capacity and it's changes probably play an effect in all of this, probably a large one as the fuel is among the heaviest part of the rocket.

But most of the fuel is burned during the ascent when it the booster is coupled to the other two stages.

They have cold gas thrusters, and a vectoring rocket engine, and those steerable grid fins to work with.

It might be fun to think of the booster as a pendulum or an inverted pendulum, but the air drag forces may at times be much larger than gravity forces. But near touchdown the velocity and air drag, and remaining fuel all approach zero.

I'm sure that the full answer is far from simple and that the equations are very nonlinear. If someone finds a SpaceX paper discussing it, please post the link.

 

[artis]

yes @anorlunda it would be interesting indeed to know.

yup indeed they seem to use some sort of gas for horizontal stabilization during descent as the main engine is limited probably in it's capability to alter certain angles or spinning of the rocket and it's main function is probably to slow down the rocket.
I've noticed the small puffs at the sides , also in the failed landings when the rocket fell on it;s side it always seemed to explode which seems a bit odd, given so little fuel is left and there doesn't seem to be a ignition source or maybe the fuel they use is very flammable upon contact with air etc?

as those failed rockets fell on their sides they not only exploded in flames but also expelled something which resembled a fast forward evaporation of dry ice that I assume could be the gas used for rocket control?



In this video right at 0:52 we can see one of the rocket falling on it;s side and exploding , if one slows the video and plays it moment by moment like I did one can see that as the rocket hits the platform, it's top half explodes some half second at least before the bottom flame part goes up, but the top half doesn't explode with a fireball instead in a puff of smoke, is that liquid Oxygen or something in the top part?

 

[Klystron]

I heard a NASA spokesperson mention hypergolic fuels during first stage recapture and when the crew Dragon maneuvered for docking.

Hypergolic fluids are toxic liquids that react spontaneously and violently when they contact each other. These fluids are used in many different rocket and aircraft systems for propulsion and hydraulic power including: orbiting satellites, manned spacecraft , military aircraft, and deep space probes. Hypergolic fuels include hydrazine (N2H4)and its derivatives including: monomethylhydrazine (MMH), unsymmetrical di-methylhydrazine (UDMH), and Aerozine 50 (A-50), which is an equal mixture of N2114 and UDMH. The oxidizer used with these fuels is usually nitrogen tetroxide (N2O4), also known as dinitrogen tetroxide or NTO, and various blends of N2O4 with nitric oxide (NO).

The above excerpt comes from this NASA paper concerned with fires from launch accidents and propellant spills..

 

[hutchphd]

I'm sure that the full answer is far from simple and that the equations are very nonlinear. If someone finds a SpaceX paper discussing it, please post the link.

Oh yeah the nitrogen thrusters, too. How closely does SpaceX hold its information?
I did heard an interview with Elon Musk talking about how powered descent is in many ways simpler than any other recovery method...the speeds are much faster than wind, for instance.
I would also point to my toy quadcopter which weighs half a pound and provides fly-by-wire and costs $100. Sure the crashes are worse for falcon but the control requirements not that much more difficult, particularly in this era of GPS.
Ya got to love the way it "sticks" the landing...no guts no glory

 

[mfb]

In space it uses cold gas thrusters to control its orientation, in the atmosphere it uses the grid fins and the engines. The center of mass is always low for nearly empty rockets as the engines and support infrastructure are heavy. The center of drag is high because of the grid fins, that combination makes the rocket stable.

There are quite a few innovations going into this: The bottom of the rocket needs to withstand strong drag. The engines need to restart while flying against the wind. The rocket needs to fly through its own hot exhaust.

And then of course there is the task of landing at the right spot. Even a single engine at minimal throttle has more thrust than the weight of the (now empty) rocket, so the first stage cannot hover. It needs to reach zero height at the same time as zero velocity in all three axes, the right orientation and not too much rotation, all that with the available fuel reserves, with the available cold gas reserves, with the available hydraulics reserves (grid fin control), with a single attempt. And ideally all these margins are as narrow as possible to get more mass into space.
Here is some discussion of the system (PDF, the article about rocket landing starts at page 15). Written by Lars Blackmore, "Principal Rocket Landing Engineer" at SpaceX - developing Falcon 9 landing at that time, now he's working on Starship.
Here is an earlier mathematical optimization paper, focused on Mars. The first author is Blackmore, too, this was before his time at SpaceX (you can guess why SpaceX wanted to hire him).The rockets land nearly empty, but nearly empty still means 200+ kg of RP-1 (basically high quality kerosene). If the booster tips over the tanks burst, now you have 200+ kg of kerosene and a similar amount of liquid oxygen, plus some sparks from things that hit each other...
Liquid oxygen is the upper tank if I remember correctly.

 

[artis]

@mfb I actually only now realized that the "grid fins" are located on the top side of the booster as I had confused them with the landing struts that come outwards when the rocket has almost landed.

Well it makes sense to use as much passive aids as possible to stabilize the rocket's reentry. So the center of mass is at the bottom at reentry which makes sure the rocket goes bottom first and then at the op they have the grid fins , so they use the top as a sort of control arm to stabilize the rocket vertically.
Then as the rocket is almost landed , the built in GPS and gyro's etc say it;'s time to stop the descent speed and fire up the engine , so I assume that when the engine has been fired the rocket better be stabilized to it's maximum as if the engine would start with a sideways rocket it would definitely make it crash.As for the moment when to fire the engine I guess they calculated it using the rocket's mass and speed but aren't those parameters changing a bit from one landing to the next? I mean air density changes and winds change etc so how do the rocket or should I say the algorithm knows at which exact point to fire the engine each time ?

 

[mfb]

The rocket calculates when to fire based on its remaining fuel, current position, wind and all the other relevant parameters, the publication discusses them. It finds a solution, probably the one with the largest safety margins/biggest chance of success with the given parameters, and then uses that. It updates the solution over time as new data come in.

 

[artis]

Well now that they pretty much seem to have mastered the technology it would be interesting to see what will be the successful landing ratio vs the failed one.

 

[mfb]

Wikipedia has lists and statistics.
https://en.wikipedia.org/wiki/List_of_Falcon_9_and_Falcon_Heavy_launches
https://en.wikipedia.org/wiki/List_of_Falcon_9_first-stage_boosters

6/8 success rate this year so far, 15/16 last year, 12/14 in 2018 and 14/14 in 2017. Before that it was very unreliable.
One failure this year was a problem with wind data, the other one had an engine problem.

 

[FactChecker]

As for the moment when to fire the engine I guess they calculated it using the rocket's mass and speed but aren't those parameters changing a bit from one landing to the next? I mean air density changes and winds change etc so how do the rocket or should I say the algorithm knows at which exact point to fire the engine each time ?

The control laws can be quite adaptive to variations like that. Within reason, all they need to know are the desired velocity, position, and orientation. Then, given the deviation of actual from desired, they increase or decrease the commands till the two are close enough. That method is fairly robust and works in a lot of conditions.

 

[dlgoff]

6/8 success rate this year so far,

After following NASA's launches from the beginning (Mercury,...), I feel like a kid again following the SpaceX launches. Just amazing.

 

[trurle]

As for the moment when to fire the engine I guess they calculated it using the rocket's mass and speed but aren't those parameters changing a bit from one landing to the next?

At the last second of landing, the short-range radar is used to provide actual distance plus angle from landing point to booster, allowing to correct weather-related errors. GPS alone is not accurate enough, especially for high-speed landing where you have no time to average out the errors. During Flight 25 landing in 2016, booster maneuvered sharply moment before landing due to radar glitch.
https://www.space.com/38155-spacex-rocket-landing-explosions-blooper-video.html

 

[hutchphd]

Wikipedia has lists and statistics.
https://en.wikipedia.org/wiki/List_of_Falcon_9_and_Falcon_Heavy_launches
https://en.wikipedia.org/wiki/List_of_Falcon_9_first-stage_boosters

6/8 success rate this year so far, 15/16 last year, 12/14 in 2018 and 14/14 in 2017. Before that it was very unreliable.
One failure this year was a problem with wind data, the other one had an engine problem.

Here is a continuous 4k ground track of the booster from launch to landing. Very good photogrphy. New to me.

 

[artis]

that really seems spectacular,

So I suppose they developed this booster rocket and for lower weight payloads they simply use one of them but for heavier stuff they use two such boosters.

 

[hutchphd]

I believe the payload penalty for returning the booster is about 30% to Low Earth Orbit.

 

[mfb]

SpaceX has many options how to fly. Roughly ascending order of payload to low Earth orbit:
* Falcon 9, booster flies back to the launch site (typical for Dragon 1 ISS resupply missions and satellites below 6 tonnes)
* Falcon 9, booster lands downrange on the ship (most common launch, heaviest payload is Starlink at ~16 tonnes)
* Falcon 9, booster is thrown away (~23 tonnes)
* Falcon Heavy (three boosters), side boosters fly back and center booster lands downrange (all 3 FH flights so far)
* Falcon Heavy, side boosters fly back, center booster is thrown away OR Falcon Heavy, all boosters land at sea (only a theoretical option, SpaceX doesn't have enough ships)
* Falcon Heavy, side boosters land at sea, center booster is thrown away
* Falcon Heavy, fully expendable (64 tonnes)

 

[artis]

so basically they can land any booster or throw any booster away because all boosters are the same type and they can be either reused or not?
Or when they need extra heavy payloads they use a simpler non reusable booster which can then be discarded after flight? This would seem like a more logical step.

I wonder how many they have thrown away compared to how many reused , not counting the ones that failed and crashed.

 

[Tom.G]

I SERIOUSLY want one of those tracking cameras to play with!

 

[mfb]

If they throw a booster away they don't mount the grid fins and legs, but it's largely the same thing. It is simpler than having two different versions to maintain in parallel.

I wonder how many they have thrown away compared to how many reused , not counting the ones that failed and crashed.

I linked statistics in post #12. First reuse was in 2017. In 2018, 2019 and 2020 more than half of all flights were on reused boosters, despite some boosters crashing.
This year only the crewed flight was a new booster so far, all 8 other launches were reusing boosters. This will increase to 1:10 in the next two weeks, before the GPS satellite launches on a new booster end of June/early July.

 

[artis]

Well I guess it's much more than just grid fins and legs , they also must come with their control circuits and hydraulics or whatever they use to physically actuate them.As for the concept of reusable booster I wonder how "reusable" it actually is in terms of how many times one booster can launch and land over and over until the body or internal components start failing due to the large stress of various kinds it is subjected to.
Now obviously at least for human missions they would probably calculate the failure probability to know at which time or after how many cycles a booster should be retired or overhauled otherwise they could end up with another "Challenger" disaster,
what do you think?

 

[trurle]

As for the concept of reusable booster I wonder how "reusable" it actually is in terms of how many times one booster can launch and land over and over until the body or internal components start failing due to the large stress of various kinds it is subjected to.

According to very old publication in AIAA journal back to 196x, the reusable booster life is tradeoff between increased deadweight and metal structural members been subject to cracking. Simply speaking, you can make a booster which will last arbitrarily long, but it is meaningless because the payload will become too small. The metallurgy did not significantly improve in last 60 years (especially because SpaceX deliberately opted to use cheaper mass-production alloys), therefore tradeoff is still here. The optimal range is 5-10 reuses of the booster.
More interesting, with the 5 reuses the per-launch costs compared to expendable booster are 130% in 1970, and 70% according to SpaceX statement in 2020. The difference is because with advanced manufacturing tools nowadays, the manufacturing chain has become simpler, therefore less cost penalties are associated with making a smaller number of boosters.

 

[mfb]

Now obviously at least for human missions they would probably calculate the failure probability to know at which time or after how many cycles a booster should be retired or overhauled otherwise they could end up with another "Challenger" disaster,

They do that analysis for uncrewed missions, too, losing one would be bad for the business and would halt launches for a while. So far it looks like 5 flights are no problem, we'll see how many more they can make. The fleet leaders are used for Starlink launches where they are their own customer, once the boosters show that they can handle this well other boosters follow with external customers.

 

[Dullard]

One (under-appreciated?) advantage of re-usable boosters is the fact that the designer gets to examine returned hardware. Designs may be 'adjusted' to add or reduce margin, based on the results of actual use. Successful 'expendable' designs are (unavoidably) over-built, and there is no/limited feedback to modify the original design decisions.

 

[artis]

@Turtle so in 1970's it was less expensive to build single use boosters instead of trying to fabricate re-useable ones?
It seems that part of this re-use gain is in the cheaper and faster materials production achievable now with respect to the price and complexity it took decades ago?

 

[trurle]

@Trurle so in 1970's it would be less expensive to build single use boosters instead of trying to fabricate re-useable ones?

Yes, exactly.

It seems that part of this re-use gain is in the cheaper and faster materials production achievable now with respect to the price and complexity it took decades ago?

Yes if wide meaning of "materials" is used. Although i would use word "subsystems" instead of "materials". Raw materials contribution to booster cost was negligible (~1%) in 197x. What``s why 196x-197x designs sometimes utilized even rhenium combustion chambers (costing about 50% of gold on ingot level).

 

[artis]

well i guess so since now every part of a complex system can be engineered to precise limits , manufacturers usually take the route of using the cheapest material possible that can still do the job within limits ,but then again looking from this logic what's the point of a re-usable rocket booster in 2020 apart from being cool and geeky?

because in most other areas we now do the exact opposite , we for example don't build cars that last a millenium anymore instead we build cars that mostly last their warranty, the same logic goes to most other products,

 

[mfb]

Can you use your car to get home again? If yes then it is already more reusable than all current rockets.
No one would buy a car that has to be scrapped after a single trip. Even if material costs would be negligible, assembly of the car is expensive. Assembly of a rocket is expensive, too.

 

[artis]

No @mfb My thinking was different. It is a well known fact among the automotive mechanic circles that many brands made far more reliable and longer lasting vehicles back in the 80's and 90's and some into early 2000's than they do now, I guess you could say it's part of planned obsolescence.
The manufacturer now just calculates how long something has to last and then searches for the cheapest material that still fits the bill. Thanks to modern computer simulation they can do this without worries as back in the day one had to put in more "reserve" capability in order to be sure not to fail.

But I guess this logic applies more to the wider consumer industry where parts and tings are bought by the millions so the companies are interested in continually upgrading and introducing new stuff so their kind of interested that the old stuff doesn't last forever, I guess aerospace would have a bit different approach due to the high cost and not go that same route,

although who knows , I'm kinda skeptical but they are trying to 3d print boosters etc to the point where they would become extra cheap and could literally be recovered as scarp and recycled.
I guess there must be a line of cost per booster below which it would become useless to recover it, interesting to see whether we will go back to single use boosters again to close the circle once our manufacturing capability will take a another step further.

 

[mfb]

Planned obsolescence would make no sense in an industry where the manufacturer sells the rides, not the vehicles. You don't buy a rocket. You buy a ride to space. If the manufacturer of the rocket can reuse it instead of building a new rocket they save money. If they can reuse it 100 times instead of 5 times they save more money, so SpaceX tries to reuse their rockets as often as feasible. If components are known to break down earlier than other components then these components can be exchanged while the rest keeps flying.

Sure, there is no point in qualifying something for 10,000 flights because Falcon 9 will not make that many.

 

[trurle]

well i guess so since now every part of a complex system can be engineered to precise limits , manufacturers usually take the route of using the cheapest material possible that can still do the job within limits ,but then again looking from this logic what's the point of a re-usable rocket booster in 2020 apart from being cool and geeky?

because in most other areas we now do the exact opposite , we for example don't build cars that last a millenium anymore instead we build cars that mostly last their warranty, the same logic goes to most other products,

30% per-flight cost reduction for Falcon 9 reusable boosters is actually bordering no-significance zone. Currently SpaceX got a small competitive advantage thanks to reusable booster, but it is uncertain if advantage will remain. The reversion to fully expendable boosters in future, even within SpaceX, is possible.

 

[mfb]

30% price difference is a lot.

The reversion to fully expendable boosters in future, even within SpaceX, is possible.

I would like to see a reference for that claim, or at least a good argument.
Why would SpaceX choose to spend more money on flights and slow down their flight rate for no gain?

 

[artis]

please go to 0:26 and watch the booster that flipped midair , now in all the fail videos the booster always explodes , I would think that this is not some built in self destruction but merely a side effect from the booster propellants?

In this video we see that as the booster flips midair it almost immediately after explodes midair without hitting anything, is this because of the way the fuel is stored in the fuel tanks or something and it then mixes and ignites?

Because so far I have understood that there are 2 things in the booster that tend to explode , in the upper stage some sort of what appears to be a highly compressed gas and in the lower stage the rocket propellant.