Starship heat shield

[Hog]

The nose cone has glued tiles, just as the Shuttle had. Those don't fall off, unless they were not glued properly. (Anyone watching Starbase live today would have seen how hard it is to remove glued tiles. Chisel and big hammer needed.)

The ones most in danger imho are in the middle of the ship, held by studs. If one is not flush enough to the ones around it that corner could break off. If a tiles is not installed the correct way it could fall off too.

What no one can know for sure, until at least the first orbital flight, is whether or not losing one tile causes the loss of others around it. I don't know what kind of forces the plasma can exert on a tile, besides the heat. Is it capable of getting under the next tile and flicking it out?

Emphasis mine.
During STS times this was known as the "zipper effect".

The "zipper effect" was a concern during shuttle development. But flight experience proved that the "zipper effect" was highly unlikely to be triggered by a lost tile. Increased turbulence and thermal loading to the exposed side of the next tile was observed, but was not strong enough to force the debonding of that next tile. STS-27 being the best observations of this lack of zipper effect.

Not sure if considered a part of Shuttle's development period or not, but the "zipper effect", or the lack thereof, was accentuated by media reports of "supposed tile losses" during OV-102 Columbia's transit atop the Shuttle Carrier Aircraft from California to Florida.  What appeared to be larger sections of tiles that were lost during this SCA transit, were actually tile spacers that were installed onto the orbiter vehicle specifically for the delivery flight.  While SOME tiles were indeed lost, the vast volumes of lost tiles due to this "zipper effect" were incorrect and hugely overstated.  To the un-knowing, the off-nominal appearing "false tiles" appeared like missing or damaged tiles which combined with earlier worries from NASA engineers about the possibility of the "zipper effect" seemed to propagate the apparent myth of the "zipper effect".  Another case or knowing "just enough to be dangerous".  There's a LOT of this in the world of spaceflight enthusiasm.

There was a huge push to get the Thermal Protection System(TPS) installed on Columbia as the launch date was looming, so these "false tile sections" were fashioned onto Columbia specifically for her cross country SCA flight to Florida. The remaining tiles would be installed at KSC at the rate of 1.7 tiles per person per week.

attachments
1)Columbia tile situation upon arrival at KSC
2) tile history slide
3) Columbia delivery flight

[OTV Booster]

Re: loosing tiles. I can't put my finger on it but there has been a NASA paper linked on NSF showing a relationship between radius of curvature and shock stand-off.


Assume the very tippy tip tile on the nose were lost. The radius of curvature would normally be the nose radius. Remove the tile and the radius of curvature that is of interest would be the top edge of the next tile down on the centerline. This edge is sharp and the radius of curvature of the shock front would be correspondingly small. My guess is that there would be some risk of shock impingement where the tile is missing. If the shock does not impinge and the shock stand-off distance small, heat input might be mitigated by any cryo cooling the header tank would give. Then the problem becomes landing propellant loss.


The entire centerline would be a concern with the the (I'm at a loss for the correct word) point towards the bottom of the ogive that is the first point of atmospheric 'contact' being of special concern. Moving away from the centerline, the shock stand-off distance increases and the risk associated with tile loss goes down.


This looks like a map for a backup TPS system, if future experience warrants. All the fiddly shapes, including the fins, makes my brain hurt but there may be a similar map here too. Key point: a backup TPS need not be full sized or of constant thickness.

SS does not enter "nose" first. It enters belly first. All references to "nose radius" in the re-entry literature refer to the curvature of the entering body in its direction of travel. If the tippy-tip TPS falls off, I don't know how to calculate the "radius" of the result. The Starship nose is sticking out toward the side of the plasma flow.

No, please. I understand SS will not reenter nose first. Reread the my post with the assumption of a realistic AoA.

AIUI, SS is intended to reenter with an AoA of ~70deg. This moves the point of first contact from the centerline of the barrel up onto the ogive. At this point there are two radii to consider. Up/down and left/right, and it gets complicated.

At the edge where the tippy tip tile is missing a new up/down radius is introduced along the centerline. Flow will 'flow' over the lip and most likely down to the newly exposed surface. The shock will most probably also make an adjustment around the lip but my dedicated CFD computer is down right now (/s) and I can't tell if there will be any shock impingement. Greatest danger might be the side face of the tile opposite the flow.

Down on the barrel, let's assume an AoA of 90deg. It makes the visualization easier. The shock front is closest along the centerline and compressive force and it's heating are greatest. Surface flow is from the centerline to the sides. Remove one centerline tile.

The compressive forces stay the same but are now on the skin, not a tile. AIUI, the shock front would dimple in at a rate determined by the radius of curvature of the tile edge. Here it gets interesting. With the shock front dimpling in and the missing tile forming a pocket, there may be a Venturi effect with the semi trapped flow excessively heating the adjacent tile edges. This effect decreases as the missing tile distance from the centerline increases.

This is admittedly all a mind experiment. The only number I have is 42, and Elon's already used it on SS.

[OTV Booster]

The nose cone has glued tiles, just as the Shuttle had. Those don't fall off, unless they were not glued properly. (Anyone watching Starbase live today would have seen how hard it is to remove glued tiles. Chisel and big hammer needed.)

The ones most in danger imho are in the middle of the ship, held by studs. If one is not flush enough to the ones around it that corner could break off. If a tiles is not installed the correct way it could fall off too.

What no one can know for sure, until at least the first orbital flight, is whether or not losing one tile causes the loss of others around it. I don't know what kind of forces the plasma can exert on a tile, besides the heat. Is it capable of getting under the next tile and flicking it out?

I've a hunch they'll start glueing the centerline tiles too. That'll suck for repair but the need for repair will go down. ISTM that the possibility of a missing tile starting a cascade is highest in the centerline. See previous post.

[OTV Booster]

Re: loosing tiles. I can't put my finger on it but there has been a NASA paper linked on NSF showing a relationship between radius of curvature and shock stand-off.


Assume the very tippy tip tile on the nose were lost. The radius of curvature would normally be the nose radius. Remove the tile and the radius of curvature that is of interest would be the top edge of the next tile down on the centerline. This edge is sharp and the radius of curvature of the shock front would be correspondingly small. My guess is that there would be some risk of shock impingement where the tile is missing. If the shock does not impinge and the shock stand-off distance small, heat input might be mitigated by any cryo cooling the header tank would give. Then the problem becomes landing propellant loss.


The entire centerline would be a concern with the the (I'm at a loss for the correct word) point towards the bottom of the ogive that is the first point of atmospheric 'contact' being of special concern. Moving away from the centerline, the shock stand-off distance increases and the risk associated with tile loss goes down.


This looks like a map for a backup TPS system, if future experience warrants. All the fiddly shapes, including the fins, makes my brain hurt but there may be a similar map here too. Key point: a backup TPS need not be full sized or of constant thickness.

SS does not enter "nose" first. It enters belly first. All references to "nose radius" in the re-entry literature refer to the curvature of the entering body in its direction of travel. If the tippy-tip TPS falls off, I don't know how to calculate the "radius" of the result. The Starship nose is sticking out toward the side of the plasma flow.

Starship entry AoA is ~70°. The nose still still be first, and since the nose itself has an angle of ~60°, the stagnation point will just be a little to the nadir of the very tippy top.

Stagnation point! That's the term I was looking for. Thanks.

[adrianwyard]

The nose cone has glued tiles, just as the Shuttle had. Those don't fall off, unless they were not glued properly. (Anyone watching Starbase live today would have seen how hard it is to remove glued tiles. Chisel and big hammer needed.)

The ones most in danger imho are in the middle of the ship, held by studs. If one is not flush enough to the ones around it that corner could break off. If a tiles is not installed the correct way it could fall off too.

What no one can know for sure, until at least the first orbital flight, is whether or not losing one tile causes the loss of others around it. I don't know what kind of forces the plasma can exert on a tile, besides the heat. Is it capable of getting under the next tile and flicking it out?

I've a hunch they'll start glueing the centerline tiles too. That'll suck for repair but the need for repair will go down. ISTM that the possibility of a missing tile starting a cascade is highest in the centerline. See previous post.

If glue turns out to be necessary to fix most tiles I wonder if it would make sense to switch to larger replaceable units, e.g. an entire barrel-section's worth of glued tiles on a lightweight structure. This is far from a 'rapidly reusable' ideal TPS but I could imagine a large supply of barrel sections being on hand, so if a Starship lands with a couple of broken tiles on the belly, you swap out its entire section in less time than a custom repair. It's far from ideal, and doesn't address the complex geometry sections but it would address ~70% of the TPS area?

[RamsesBic]

The nose cone has glued tiles, just as the Shuttle had. Those don't fall off, unless they were not glued properly. (Anyone watching Starbase live today would have seen how hard it is to remove glued tiles. Chisel and big hammer needed.)

The ones most in danger imho are in the middle of the ship, held by studs. If one is not flush enough to the ones around it that corner could break off. If a tiles is not installed the correct way it could fall off too.

What no one can know for sure, until at least the first orbital flight, is whether or not losing one tile causes the loss of others around it. I don't know what kind of forces the plasma can exert on a tile, besides the heat. Is it capable of getting under the next tile and flicking it out?

I've a hunch they'll start glueing the centerline tiles too. That'll suck for repair but the need for repair will go down. ISTM that the possibility of a missing tile starting a cascade is highest in the centerline. See previous post.

If glue turns out to be necessary to fix most tiles I wonder if it would make sense to switch to larger replaceable units, e.g. an entire barrel-section's worth of glued tiles on a lightweight structure. This is far from a 'rapidly reusable' ideal TPS but I could imagine a large supply of barrel sections being on hand, so if a Starship lands with a couple of broken tiles on the belly, you swap out its entire section in less time than a custom repair. It's far from ideal, and doesn't address the complex geometry sections but it would address ~70% of the TPS area?

Thermal expansion and contraction would crack such sections.

[adrianwyard]

If the idea had merit I'm confident SpaceX engineers could work up expansion joints. The bigger question is how much weight would such a panelized approach add?

[BT52]

If the idea had merit I'm confident SpaceX engineers could work up expansion joints. The bigger question is how much weight would such a panelized approach add?

Yeah i got similar idea. But with bottom mesh as support and tiles are glued on it. Sorta strapping mechanism.

No matter how you put it. Bayonet clip with steel  frame inside tile should be still more weight efficient.

Ofc most efficient is just glue all of it directly on shell... But Shuttle showed that this is tough cookie to maintain and quick turn around is not really possible. That's why we both juggling with other means to quickly remove panels and cure them in more clean, assessable and controlled environment.

[john smith 19]

Thermal expansion and contraction would crack such sections.

If the skin were aluminum and they didn't use a stress isolation pad yes.

But SS is steel. I can't recall if it's TCE is 1/3 or 1/10 that of aluminum. It's definitely a lot lower.

Maybe low enough not to be a problem.

Truth is without a test flight to orbit and at least an attempt to get back to ground we lack any solid data. 

[DanClemmensen]

The nose cone has glued tiles, just as the Shuttle had. Those don't fall off, unless they were not glued properly. (Anyone watching Starbase live today would have seen how hard it is to remove glued tiles. Chisel and big hammer needed.)

The ones most in danger imho are in the middle of the ship, held by studs. If one is not flush enough to the ones around it that corner could break off. If a tiles is not installed the correct way it could fall off too.

What no one can know for sure, until at least the first orbital flight, is whether or not losing one tile causes the loss of others around it. I don't know what kind of forces the plasma can exert on a tile, besides the heat. Is it capable of getting under the next tile and flicking it out?

I've a hunch they'll start glueing the centerline tiles too. That'll suck for repair but the need for repair will go down. ISTM that the possibility of a missing tile starting a cascade is highest in the centerline. See previous post.

If glue turns out to be necessary to fix most tiles I wonder if it would make sense to switch to larger replaceable units, e.g. an entire barrel-section's worth of glued tiles on a lightweight structure. This is far from a 'rapidly reusable' ideal TPS but I could imagine a large supply of barrel sections being on hand, so if a Starship lands with a couple of broken tiles on the belly, you swap out its entire section in less time than a custom repair. It's far from ideal, and doesn't address the complex geometry sections but it would address ~70% of the TPS area?

Thermal expansion and contraction would crack such sections.

How does thermal expansion compare with expansion due to pressurization? I would think that pressurization would cause more on a tank that large.

[RamsesBic]

A ship with about atmospheric pressure is built with rather large gaps between the tiles. At operational pressure (6 bar) it will expand - that is fine for the tiles, the gaps might increase a bit.
But when it is re-entring the atmosphere (still at 6 bar internal pressure) and the temperature reaches its maximum those gaps will narrow to be almost zero due to expansion of the tiles. This is the part that would cause a very large tile to crack when it has nowhere to expand.
Of course the very large tile would also have trouble when the ship goes from 1 bar to 6 bar.

Another question is how would you install such a big tile? How many studs would you need? Will you use a frame? Then how much weight will it add? It is already hard to install a small tile with 3 holes onto 3 studs.

How much would get exposed if such a very large tile is lost?

I really can't see any benefits from using very large tiles (or shields if you prefer).

I am still not giving up on the present tiles. We just have to wait and see.

[sebk]

Re: loosing tiles. I can't put my finger on it but there has been a NASA paper linked on NSF showing a relationship between radius of curvature and shock stand-off.


Assume the very tippy tip tile on the nose were lost. The radius of curvature would normally be the nose radius. Remove the tile and the radius of curvature that is of interest would be the top edge of the next tile down on the centerline. This edge is sharp and the radius of curvature of the shock front would be correspondingly small. My guess is that there would be some risk of shock impingement where the tile is missing. If the shock does not impinge and the shock stand-off distance small, heat input might be mitigated by any cryo cooling the header tank would give. Then the problem becomes landing propellant loss.


The entire centerline would be a concern with the the (I'm at a loss for the correct word) point towards the bottom of the ogive that is the first point of atmospheric 'contact' being of special concern. Moving away from the centerline, the shock stand-off distance increases and the risk associated with tile loss goes down.


This looks like a map for a backup TPS system, if future experience warrants. All the fiddly shapes, including the fins, makes my brain hurt but there may be a similar map here too. Key point: a backup TPS need not be full sized or of constant thickness.

SS does not enter "nose" first. It enters belly first. All references to "nose radius" in the re-entry literature refer to the curvature of the entering body in its direction of travel. If the tippy-tip TPS falls off, I don't know how to calculate the "radius" of the result. The Starship nose is sticking out toward the side of the plasma flow.

No, please. I understand SS will not reenter nose first. Reread the my post with the assumption of a realistic AoA.

AIUI, SS is intended to reenter with an AoA of ~70deg. This moves the point of first contact from the centerline of the barrel up onto the ogive. At this point there are two radii to consider. Up/down and left/right, and it gets complicated.

At the edge where the tippy tip tile is missing a new up/down radius is introduced along the centerline. Flow will 'flow' over the lip and most likely down to the newly exposed surface. The shock will most probably also make an adjustment around the lip but my dedicated CFD computer is down right now (/s) and I can't tell if there will be any shock impingement. Greatest danger might be the side face of the tile opposite the flow.

Down on the barrel, let's assume an AoA of 90deg. It makes the visualization easier. The shock front is closest along the centerline and compressive force and it's heating are greatest. Surface flow is from the centerline to the sides. Remove one centerline tile.

The compressive forces stay the same but are now on the skin, not a tile. AIUI, the shock front would dimple in at a rate determined by the radius of curvature of the tile edge. Here it gets interesting. With the shock front dimpling in and the missing tile forming a pocket, there may be a Venturi effect with the semi trapped flow excessively heating the adjacent tile edges. This effect decreases as the missing tile distance from the centerline increases.

This is admittedly all a mind experiment. The only number I have is 42, and Elon's already used it on SS.

Actually, on the center-spot (or center line at 90° AoA) behind the stagnation point there would be no local shock even if a tile fell off. The bow shock would be around half a meter away (bow shock standoff is about 1/9 of the radius of the curvature) and behind the shock you have stagnation point, i.e. the point where there's no major air movement. There will be turbulence and stuff (unless things are so smooth the flow is laminar) but they will be subsonic.

But as you'd start moving to the side, you'd start getting air movement again. If you'd move totally to the side you'd actually get supersonic flow (in the order of Mach 2.5 if I remember correctly; I could be badly wrong, though). So somewhere on the side of the barrel, likely at some point between 30° and 60° to the side you actually get supersonic flow behind the bow shock too. So that's the place where there's potential for shock impingement onto the skin if a tile is missing. OTOH heating would be less there, because bow shock standoff is significantly larger back there, and the flow has lower pressure (so it's colder and contains less heat to transfer).

You could (extremely roughly!) think about the skin of the vehicle and bow shock forming a kind of a nozzle through which the stagnation point air expands to escape behind the sides of the reentering vehicle.

[edzieba]

The bow shock would be around half a meter away (bow shock standoff is about 1/9 of the radius of the curvature) and behind the shock you have stagnation point, i.e. the point where there's no major air movement.

The problem is that once a tile is removed (or the TPS otherwise mechanically disrupted) you no longer have a radius equal to the radius of the vehicle. You have a radius equal to the lip of the edge of the adjacent tiles, which is a few mm at most. The shock in that region is then happy to reattach to the surface of the vehicle and allow hot gas to impinge on the vehicle.

[eriblo]

The bow shock would be around half a meter away (bow shock standoff is about 1/9 of the radius of the curvature) and behind the shock you have stagnation point, i.e. the point where there's no major air movement.

The problem is that once a tile is removed (or the TPS otherwise mechanically disrupted) you no longer have a radius equal to the radius of the vehicle. You have a radius equal to the lip of the edge of the adjacent tiles, which is a few mm at most. The shock in that region is then happy to reattach to the surface of the vehicle and allow hot gas to impinge on the vehicle.

No, as sebk stated the relevant radius of curvature has to do with how the gas flows around the vehicle and this will not change drastically for disruptions that are significantly smaller than the shock standoff distance.

The dangers are shocks in the local (up to supersonic) flow and increased heat transfer downstream due to turbulence. The space shuttle flew experiments to study this where one tile had an up to 0.5 inch tall ridge with radii of a few mm (see for example the attached pdf) which did not suffer any damage.

[john smith 19]

The dangers are shocks in the local (up to supersonic) flow and increased heat transfer downstream due to turbulence. The space shuttle flew experiments to study this where one tile had an up to 0.5 inch tall ridge with radii of a few mm (see for example the attached pdf) which did not suffer any damage.

Interesting paper. 

But note those tile locations were selected to avoid damage to begin with. That's not to say things could have been much more serious in other locations.

What is good new s was that all predicted temperatures were substantially higher than the actual results.

So the design tool results were conservative and while you can't be quite sure how much lower the actual temperatures will be, you can be sure they will be lower.

In fact the difference looks big enough that there may be a systemantic element that needs to be found to improve the match.

The joker in the pack is to what exten those tools are "first principles" or wheather they use fudge factors drawn from wind tunnel models of Shuttle. Obviously if the latter then transferring the results to SS will be very doubtful.

[InterestedEngineer]

The bow shock would be around half a meter away (bow shock standoff is about 1/9 of the radius of the curvature) and behind the shock you have stagnation point, i.e. the point where there's no major air movement.

The problem is that once a tile is removed (or the TPS otherwise mechanically disrupted) you no longer have a radius equal to the radius of the vehicle. You have a radius equal to the lip of the edge of the adjacent tiles, which is a few mm at most. The shock in that region is then happy to reattach to the surface of the vehicle and allow hot gas to impinge on the vehicle.

No, as sebk stated the relevant radius of curvature has to do with how the gas flows around the vehicle and this will not change drastically for disruptions that are significantly smaller than the shock standoff distance.

The dangers are shocks in the local (up to supersonic) flow and increased heat transfer downstream due to turbulence. The space shuttle flew experiments to study this where one tile had an up to 0.5 inch tall ridge with radii of a few mm (see for example the attached pdf) which did not suffer any damage.

A void and a protuberance are not the same thing when it comes to hypersonic flow. 

So while the paper is interesting, it doesn't apply to missing tiles.

[OTV Booster]

Re: loosing tiles. I can't put my finger on it but there has been a NASA paper linked on NSF showing a relationship between radius of curvature and shock stand-off.


Assume the very tippy tip tile on the nose were lost. The radius of curvature would normally be the nose radius. Remove the tile and the radius of curvature that is of interest would be the top edge of the next tile down on the centerline. This edge is sharp and the radius of curvature of the shock front would be correspondingly small. My guess is that there would be some risk of shock impingement where the tile is missing. If the shock does not impinge and the shock stand-off distance small, heat input might be mitigated by any cryo cooling the header tank would give. Then the problem becomes landing propellant loss.


The entire centerline would be a concern with the the (I'm at a loss for the correct word) point towards the bottom of the ogive that is the first point of atmospheric 'contact' being of special concern. Moving away from the centerline, the shock stand-off distance increases and the risk associated with tile loss goes down.


This looks like a map for a backup TPS system, if future experience warrants. All the fiddly shapes, including the fins, makes my brain hurt but there may be a similar map here too. Key point: a backup TPS need not be full sized or of constant thickness.

SS does not enter "nose" first. It enters belly first. All references to "nose radius" in the re-entry literature refer to the curvature of the entering body in its direction of travel. If the tippy-tip TPS falls off, I don't know how to calculate the "radius" of the result. The Starship nose is sticking out toward the side of the plasma flow.

No, please. I understand SS will not reenter nose first. Reread the my post with the assumption of a realistic AoA.

AIUI, SS is intended to reenter with an AoA of ~70deg. This moves the point of first contact from the centerline of the barrel up onto the ogive. At this point there are two radii to consider. Up/down and left/right, and it gets complicated.

At the edge where the tippy tip tile is missing a new up/down radius is introduced along the centerline. Flow will 'flow' over the lip and most likely down to the newly exposed surface. The shock will most probably also make an adjustment around the lip but my dedicated CFD computer is down right now (/s) and I can't tell if there will be any shock impingement. Greatest danger might be the side face of the tile opposite the flow.

Down on the barrel, let's assume an AoA of 90deg. It makes the visualization easier. The shock front is closest along the centerline and compressive force and it's heating are greatest. Surface flow is from the centerline to the sides. Remove one centerline tile.

The compressive forces stay the same but are now on the skin, not a tile. AIUI, the shock front would dimple in at a rate determined by the radius of curvature of the tile edge. Here it gets interesting. With the shock front dimpling in and the missing tile forming a pocket, there may be a Venturi effect with the semi trapped flow excessively heating the adjacent tile edges. This effect decreases as the missing tile distance from the centerline increases.

This is admittedly all a mind experiment. The only number I have is 42, and Elon's already used it on SS.

Actually, on the center-spot (or center line at 90° AoA) behind the stagnation point there would be no local shock even if a tile fell off. The bow shock would be around half a meter away (bow shock standoff is about 1/9 of the radius of the curvature) and behind the shock you have stagnation point, i.e. the point where there's no major air movement. There will be turbulence and stuff (unless things are so smooth the flow is laminar) but they will be subsonic.

But as you'd start moving to the side, you'd start getting air movement again. If you'd move totally to the side you'd actually get supersonic flow (in the order of Mach 2.5 if I remember correctly; I could be badly wrong, though). So somewhere on the side of the barrel, likely at some point between 30° and 60° to the side you actually get supersonic flow behind the bow shock too. So that's the place where there's potential for shock impingement onto the skin if a tile is missing. OTOH heating would be less there, because bow shock standoff is significantly larger back there, and the flow has lower pressure (so it's colder and contains less heat to transfer).

You could (extremely roughly!) think about the skin of the vehicle and bow shock forming a kind of a nozzle through which the stagnation point air expands to escape behind the sides of the reentering vehicle.

The CFD system is still down <g> so I'm trying to use my head and model this one air particle at a time. Again assume a 90deg. AoA for simplicity.


The heatshield is in vacuum so there are no aerodynamics. Then it encounters one particle of air directly on the centerline. The particle hits and that is one tiny bit of compressive heating to both the heatshield and the particle. The energetic particle is now conceptually sharing a frame of reference with the ship and being hot and energetic, rebounds.


A key concept here is 'mean free space' between air particles. As the air gets denser the mean free space goes down and a rebounding particle has an increasing chance of impacting an incoming* particle of air. When a rebounding particle hits an incoming particle a nascent shock wave forms.


That first partial that impacts an incoming particle would bounce back towards the ship and contribute again to compressive heating. The particle it impacted would also rebound and have some chance of hitting yet another incoming particle and the process would cascade. This is all at the stagnation point.


To one side or the other the process would follow this pattern but with a radial component. This is what forms the flow under the shock. Particles are always bleeding in from the shock and adding to the flow. Particles are always piling up and entering the ships frame of reference. This pileup is the shockwave.


Not sure where this is going. Gotta think on it some more. Somebody point me in the right direction if I'm off base.


* Really just unlucky enough to be in the way.

[Nomadd]

 I heard that 24 and 25 will have tiles but 26, 27 and 28 will be operational Starlink missions with no tiles. The only thing that makes sense to me is that there's a major shakeup with the heat shield that won't be ready for a while but they still want to get some Starlinks launched.
 Then again, not all that much makes sense to me.

[john smith 19]

The CFD system is still down <g> so I'm trying to use my head and model this one air particle at a time. Again assume a 90deg. AoA for simplicity.


The heatshield is in vacuum so there are no aerodynamics. Then it encounters one particle of air directly on the centerline. The particle hits and that is one tiny bit of compressive heating to both the heatshield and the particle. The energetic particle is now conceptually sharing a frame of reference with the ship and being hot and energetic, rebounds.


A key concept here is 'mean free space' between air particles. As the air gets denser the mean free space goes down and a rebounding particle has an increasing chance of impacting an incoming* particle of air. When a rebounding particle hits an incoming particle a nascent shock wave forms.


That first partial that impacts an incoming particle would bounce back towards the ship and contribute again to compressive heating. The particle it impacted would also rebound and have some chance of hitting yet another incoming particle and the process would cascade. This is all at the stagnation point.


To one side or the other the process would follow this pattern but with a radial component. This is what forms the flow under the shock. Particles are always bleeding in from the shock and adding to the flow. Particles are always piling up and entering the ships frame of reference. This pileup is the shockwave.


Not sure where this is going. Gotta think on it some more. Somebody point me in the right direction if I'm off base.


* Really just unlucky enough to be in the way.

What you're kind of describing is quite similar to how the effect of a neutron moderator was worked out from first principles in a book called "Elementary Pile Theory"

It's about momentum transfer between two bodies. In your scenario you would start by picturing it from the PoV, or frame of reference of SS, and then move to considering how both bodies react to their impact and recoil in the frame of reference of them  moving against the Earth.

You're starting at Newtonian flow (straight line impacts, like rays of light) then as the density rises into (IIRC) slip flow. so you need to consider what the Knusden number is. This is the area where "Temperature" is measuring the average speed of molecules. This is also where you have 7-11 species airflow models as the "airflow" can have multiple temperaturs. NASA's Chul Park would be a relevant author here.

[sebk]

Re: loosing tiles. I can't put my finger on it but there has been a NASA paper linked on NSF showing a relationship between radius of curvature and shock stand-off.


Assume the very tippy tip tile on the nose were lost. The radius of curvature would normally be the nose radius. Remove the tile and the radius of curvature that is of interest would be the top edge of the next tile down on the centerline. This edge is sharp and the radius of curvature of the shock front would be correspondingly small. My guess is that there would be some risk of shock impingement where the tile is missing. If the shock does not impinge and the shock stand-off distance small, heat input might be mitigated by any cryo cooling the header tank would give. Then the problem becomes landing propellant loss.


The entire centerline would be a concern with the the (I'm at a loss for the correct word) point towards the bottom of the ogive that is the first point of atmospheric 'contact' being of special concern. Moving away from the centerline, the shock stand-off distance increases and the risk associated with tile loss goes down.


This looks like a map for a backup TPS system, if future experience warrants. All the fiddly shapes, including the fins, makes my brain hurt but there may be a similar map here too. Key point: a backup TPS need not be full sized or of constant thickness.

SS does not enter "nose" first. It enters belly first. All references to "nose radius" in the re-entry literature refer to the curvature of the entering body in its direction of travel. If the tippy-tip TPS falls off, I don't know how to calculate the "radius" of the result. The Starship nose is sticking out toward the side of the plasma flow.

No, please. I understand SS will not reenter nose first. Reread the my post with the assumption of a realistic AoA.

AIUI, SS is intended to reenter with an AoA of ~70deg. This moves the point of first contact from the centerline of the barrel up onto the ogive. At this point there are two radii to consider. Up/down and left/right, and it gets complicated.

At the edge where the tippy tip tile is missing a new up/down radius is introduced along the centerline. Flow will 'flow' over the lip and most likely down to the newly exposed surface. The shock will most probably also make an adjustment around the lip but my dedicated CFD computer is down right now (/s) and I can't tell if there will be any shock impingement. Greatest danger might be the side face of the tile opposite the flow.

Down on the barrel, let's assume an AoA of 90deg. It makes the visualization easier. The shock front is closest along the centerline and compressive force and it's heating are greatest. Surface flow is from the centerline to the sides. Remove one centerline tile.

The compressive forces stay the same but are now on the skin, not a tile. AIUI, the shock front would dimple in at a rate determined by the radius of curvature of the tile edge. Here it gets interesting. With the shock front dimpling in and the missing tile forming a pocket, there may be a Venturi effect with the semi trapped flow excessively heating the adjacent tile edges. This effect decreases as the missing tile distance from the centerline increases.

This is admittedly all a mind experiment. The only number I have is 42, and Elon's already used it on SS.

Actually, on the center-spot (or center line at 90° AoA) behind the stagnation point there would be no local shock even if a tile fell off. The bow shock would be around half a meter away (bow shock standoff is about 1/9 of the radius of the curvature) and behind the shock you have stagnation point, i.e. the point where there's no major air movement. There will be turbulence and stuff (unless things are so smooth the flow is laminar) but they will be subsonic.

But as you'd start moving to the side, you'd start getting air movement again. If you'd move totally to the side you'd actually get supersonic flow (in the order of Mach 2.5 if I remember correctly; I could be badly wrong, though). So somewhere on the side of the barrel, likely at some point between 30° and 60° to the side you actually get supersonic flow behind the bow shock too. So that's the place where there's potential for shock impingement onto the skin if a tile is missing. OTOH heating would be less there, because bow shock standoff is significantly larger back there, and the flow has lower pressure (so it's colder and contains less heat to transfer).

You could (extremely roughly!) think about the skin of the vehicle and bow shock forming a kind of a nozzle through which the stagnation point air expands to escape behind the sides of the reentering vehicle.

The CFD system is still down <g> so I'm trying to use my head and model this one air particle at a time. Again assume a 90deg. AoA for simplicity.


The heatshield is in vacuum so there are no aerodynamics. Then it encounters one particle of air directly on the centerline. The particle hits and that is one tiny bit of compressive heating to both the heatshield and the particle. The energetic particle is now conceptually sharing a frame of reference with the ship and being hot and energetic, rebounds.


A key concept here is 'mean free space' between air particles. As the air gets denser the mean free space goes down and a rebounding particle has an increasing chance of impacting an incoming* particle of air. When a rebounding particle hits an incoming particle a nascent shock wave forms.


That first partial that impacts an incoming particle would bounce back towards the ship and contribute again to compressive heating. The particle it impacted would also rebound and have some chance of hitting yet another incoming particle and the process would cascade. This is all at the stagnation point.


To one side or the other the process would follow this pattern but with a radial component. This is what forms the flow under the shock. Particles are always bleeding in from the shock and adding to the flow. Particles are always piling up and entering the ships frame of reference. This pileup is the shockwave.


Not sure where this is going. Gotta think on it some more. Somebody point me in the right direction if I'm off base.


* Really just unlucky enough to be in the way.

Your understanding is pretty good.

Just few additions:

To actually have a shock formed, its standoff distance must be incomparably larger than the mean free path of air molecules. Otherwise it will be "too blurred", i.e. the variance of actual free paths of actual particles will be too large. Once would hit another just after a millimeter while another would fly freely for half a meter.

So the bigger the reentering object and lower its curvature, the earlier will the shock form. That's why you could have the bow shock of our solar system forming in a gas so rarefied that human scale objects could fly through thousands km per second without forming anything.

Another part is that those molecules will emit thermal photons, i.e. they will lose energy via radiation. And to make things even more interesting depending on their states they may be also absorb or not photons which fly around them.

The interactions are complex, you get multiple blurry layers differing by temperature and actual chemical species present (the temperatures are high enough to cause ionization not to even mention chemical reactions which would be rare in calmer conditions). For example, simplifying air to just nitrox, you'd have O2, N2, O, N, O+, N+, O3-, O3,  various species of NOx, (NO2, N2O4, NO+, N2O, etc...), even some O+2 and N+2, etc...

The net effect is that the bow shock itself is about 8000K hot, but between it and the vehicle nose things get a lot cooler, so the skin could be for example at 1260K.

NB, there's an interesting phenomenon that if the air is rarefied enough that the bow shock doesn't form, you get extremely high temperatures very close to the skin. The effect is that the skin will absorb much higher fraction of the heat when there's no bow shock, than when there's the cushion formed by the bow shock and the layers behind. So for example in Shuttle you'd already get high heating and plasma visible behind the windows while the braking force was like 0.1g or so.