SpaceX 'Star series' simulations

[Keldor]

Perhaps counterintuitively, not as good. With 9 Raptors and a 3g limit on acceleration, I'm either throttling back or shutting down engines from the 90 second mark, when the ship is already at 34kms altitude and just over 1km/s. More than 9 engines would just add mass for no benefit. Vacuum engines would have less thrust at low altitude, which is when maximum thrust is most useable.

I was picturing them shutting off the sea level engines once they reached a certain altitude and using just vacuum engines from there on.  Launch may have to be done with the vacuum engines initially off, which is why I suggested more SL engines, though Space Shuttle style semi-vacuum engines are a possibility to reduce the need for extra SL engines at liftoff.

6 SL + 3 semi-vac fireable at sealevel, shutting off the sea level engines over the course of the launch?

[OneSpeed]

Lift and drag are both shown in the simulation. Hypersonic L/D looks like about 1.1.

Isn't that a bit too high? Even shuttle has only a Hypersonic L/D of <1.5

For shuttle L/D varies with velocity, from about 1.8 at Mach 1 down to 1.1 at Mach 30. At the post SES-10 press conference Elon Musk said Falcon 9 had a lift-over-drag (ratio) of roughly one if flown at the right angle of attack.
I've calibrated SpaceSim as best I can against the SpaceX Earth EDL physics model shown in the Dear Moon presentation, and from that, L/D is about 1.2 at Mach 19, but your mileage may vary.

[hkultala]

Perhaps counterintuitively, not as good. With 9 Raptors and a 3g limit on acceleration, I'm either throttling back or shutting down engines from the 90 second mark, when the ship is already at 34kms altitude and just over 1km/s. More than 9 engines would just add mass for no benefit. Vacuum engines would have less thrust at low altitude, which is when maximum thrust is most useable.

I was picturing them shutting off the sea level engines once they reached a certain altitude and using just vacuum engines from there on.  Launch may have to be done with the vacuum engines initially off, which is why I suggested more SL engines, though Space Shuttle style semi-vacuum engines are a possibility to reduce the need for extra SL engines at liftoff.

6 SL + 3 semi-vac fireable at sealevel, shutting off the sea level engines over the course of the launch?

With what kind of engine layout?

Only two engines should be used for the final phase of the ascent to have both reasonable g-forces and good isp. But with 3-way symmetry, this would mean asymmetric thrust, requiring the craft to fly at an angle with engines gimballed quite a lot.

And having big nozzles that can gimbal a lot... then the base of the rocket easily runs out of space.

[speedevil]

Only two engines should be used for the final phase of the ascent to have both reasonable g-forces and good isp. But with 3-way symmetry, this would mean asymmetric thrust, requiring the craft to fly at an angle with engines gimballed quite a lot.

And having big nozzles that can gimbal a lot... then the base of the rocket easily runs out of space.

It would seem plausible to just go right to one engine operation after three, not two.
Start out with the engines canted out at 4 degrees as their launch position (mounted slightly inboard of where you would otherwise put them), and be able to gimbal to ~8 in one direction, and modestly less in others.

Assuming for the moment you are talking of upgraded non-throttleable 200 ton vacuum engines, and we want to avoid >3g, to make the worst case:
This means we have 600 tons thrust with all three lit, and a minimum total mass of 200 tons.
If we were then to go to two engines, highly gimballed, this takes the new total mass to 133 tons.

This is 1500m/s of flight under this regimen.

If we initially cant the vacuum engines out at 4 degrees during boost, they are 99.7% as effective as nominal. If we assume 4450m/s until the point we turn off one engine, that is a loss of ~2m/s or so due to reduced effective ISP.

The terminal phase of flight takes ~100s now, not ~50s, but I'm struggling to find a scenario in which this incurs gravity losses.

This implies that you can live with a sharply limited gimbal angle - ~5 degrees, not the ~9+ you might want if starting out at 0 gimbal, with essentially no penalty.

[Keldor]

Only two engines should be used for the final phase of the ascent to have both reasonable g-forces and good isp. But with 3-way symmetry, this would mean asymmetric thrust, requiring the craft to fly at an angle with engines gimballed quite a lot.

And having big nozzles that can gimbal a lot... then the base of the rocket easily runs out of space.

It would seem plausible to just go right to one engine operation after three, not two.
Start out with the engines canted out at 4 degrees as their launch position (mounted slightly inboard of where you would otherwise put them), and be able to gimbal to ~8 in one direction, and modestly less in others.

Assuming for the moment you are talking of upgraded non-throttleable 200 ton vacuum engines, and we want to avoid >3g, to make the worst case:
This means we have 600 tons thrust with all three lit, and a minimum total mass of 200 tons.
If we were then to go to two engines, highly gimballed, this takes the new total mass to 133 tons.

This is 1500m/s of flight under this regimen.

If we initially cant the vacuum engines out at 4 degrees during boost, they are 99.7% as effective as nominal. If we assume 4450m/s until the point we turn off one engine, that is a loss of ~2m/s or so due to reduced effective ISP.

The terminal phase of flight takes ~100s now, not ~50s, but I'm struggling to find a scenario in which this incurs gravity losses.

This implies that you can live with a sharply limited gimbal angle - ~5 degrees, not the ~9+ you might want if starting out at 0 gimbal, with essentially no penalty.

SpaceX has mentioned wanting to have ~1000 passengers on Starship.  This puts the payload into the 100 ton range by itself.  Add in the rest of the rocket, as well as the fuel needed for the landing burn, and Starship will almost certainly be 200 tons or more at the end of the vac engine burns.

Moreover, 200 tons is Raptor's peak performance.  Minimum throttle will probably be somewhere in the range of 130 tons.  This means that peak accelleration can be kept to ~2G's before they have to turn off engines, which seems perfectly reasonable.  They shouldn't have to shut down any engines.

Even if they did, the center of mass will be pretty high up when the tanks are nearly empty.  There are 100 tons sitting on the nose, after all!  My math estimates the deflection angle needed for single engine to be somewhere around 5.  Assuming the vac raptor has the gimbal 10 meters (it's likely less!) up from the end of the bell, 5 degrees means a bit less than 1 meter of deflection.  Since it will only ever have to rotate outward significant distances to compensate, there really is plenty of room for this.

For engine configuration, I would put 6 sea level engines in two rings, like a 3 pointed star.  3 vacuum engines in the spaces between the points.

In retrospect, Starship is going to need high TWR at liftoff, much higher than a standard two stage, which stretches the first stage as far as it can while the extra weight of tanks doesn't outweigh the extra fuel they provide.  But Starship P2P is single stage, so the extra tankage weight is much more significant.  9 liftoff engines seems reasonable.  This puts it at 9 SL, 3 Vac.  Let's see if I can whip up a decent layout.

[Keldor]

Thinking about it some more, engine out on the vac engines could be tricky to deal with.  Do you relight one of the sea level engines to compensate?  What's your margin on loosing ISP that way?  Should still be better than pure SL engines, right?

Anyway, the full expansion ratio 200 3 meter bells seem too large to fit and still have deep gimbaling, but it's pretty doable with 2.5m.

With 2-way symmetry, you're able to fit 2 3 meter vac engines with room for limited gimbal while having room for a pair of landing engines for engine out landing redundancy.  You can easily fit things like 2 Vac 10 SL.

[mikelepage]

This is a speculative simulation of a single stage Starship P2P flight. With 9 SL Raptors, and a full propellant load, the initial T/W is a healthy 1.6. So, throttle back for MaxQ occurs early, at the 36 second mark. If the ship were to continue to a purely ballistic trajectory, re-entry g forces would be prohibitive (~20gs). Instead, I've used negative pitch to flatten the trajectory, reducing the re-entry flight path angle. This allows the ship to skip like a stone on a pond, extending the range out to 10,000kms. The peak g force on the first 'bounce' is just over 4. If the Starship had larger (dragon?) wings, and hence a greater lift coefficient, the peak could be reduced further, and the range extended beyond 10,000kms.

Hi OneSpeed, great simulation! Reattaching your plot for reference.

With regard to optimising the trajectory, I have been wondering if the "skipping stone" analogy is the best one.  Going so deep into the atmosphere on the very first "skip" would surely have quite a dramatic (downward) effect on range?  Intuitively I wonder if performing a small burn just as Starship is passing through 80km altitude before that first skip - to zero out the vertical motion - would result in a much less "bouncy" profile, with Starship instead skimming like an air hockey puck across the top of the atmosphere, with (potentially?) increased range. 

I would love to see a simulation similar to what you've done, but with this re-entry "skim" burn in action. Obviously the real atmosphere wouldn't be so well behaved, but you could control that with real-time attitude adjustments/RCS burns the same way the F9 boosters use their grid fins/Cold gas thrusters.  Also not sure what the size of that re-entry "skim" burn would be, but perhaps it could be minimised by using negative pitch even more aggressively to get a lower apogee?  Gemini 3 was only a 161km x 225km orbit, so going to 260+km seems a little high.

[speedevil]

In retrospect, Starship is going to need high TWR at liftoff, much higher than a standard two stage, which stretches the first stage as far as it can while the extra weight of tanks doesn't outweigh the extra fuel they provide.  But Starship P2P is single stage, so the extra tankage weight is much more significant.  9 liftoff engines seems reasonable.  This puts it at 9 SL, 3 Vac.  Let's see if I can whip up a decent layout.

If the outer engines throttle well enough, 4 outer engines lit and hoverslam sort-of-works, one 200 ton thrust raptor in the middle fixed may almost work.
~4G peak, but if the passengers will put up with this, ...

Further on the more aggressive route.
Aft cargo compartment was stated at 88m^3.
If we fill this with six stubby raptors, and come off the pad at ~2g, not ~1.2 we can ~half gravity losses, or around 300m/s saving up till 300m/s. Some of this is lost by having to then throttle back to avoid the sound barrier, but it could in principle help considerably.

[Slarty1080]

Rules of engagement
Standing orders by squadron – attack bombers only
Don’t fire unless condition xyz - < 5 hexes and directly behind or advantaged
Out of ammo > head for home
>50% damage > head for home
> 75% damage bale out
Enemy directly behind 5 or less hexes > turn!
Fired at last turn > turn!

Stage 1 move to engagement from random position
Bombers continue
Escorts turn to interceptors
Interceptors turn to bombers
Continue until first burst fired or fired upon. Then switch to dogfight stage

Dogfight stage
Bombers continue fire at approaching enemy turn for home if damage > 50%
Escorts and interceptors check front aspect turn in behind nearest enemy
If no enemy visable within 7 hexes turn rnd left or right

draw cards turn left or turn right or straight on (a few)



Speed
Guns
Bursts
Range
Engines
Fuel

Turn rate
Hexes moved

Angle behind, advantaged, disadvanted, infront

Range 1-16
Options fire long or short range
Chase

Tailing in near range    fire
Tailing in far range  depending on strategy accelerate or fire
Tailing out of range   accelerate
Advantaged in near range   fire
Advantaged in far range         depending on strategy  manouvre or fire
Advantaged out of range   manoeuvre
Disadvantaged in close range of enemy
Disadvantaged in far range of enemy
Disadvantaged to out of range enemy
Tailed by enemy in close range  avoid!
Tailed by enemy in far range 
Tailed by enemy out of range

I would assume that SpaceX have been through all of the options for engine configurations many many times and will continue to review them as circumstances change (raptor performance and sl or vac, Starship mass, mission requirements and acceptable margins).

It wouldn't surprise me if we didn't see another new engine configuration at the next announcement, seems to be one of the things they change a lot. And I would not be surprised to see multiple versions with different engine configurations flying before long.

[OneSpeed]

I wondered if you had any estimates on which (if any) of those skipping events would generate sonic booms audible on the ground? But it seems to me it goes subsonic whilst still 25km altitude, which should minimise any complaints on the ground.  (Shuttle was still Mach 1.5 at 18km).

Answering your post here because the answer might have more general applicability. From:

https://www.nasa.gov/centers/armstrong/news/FactSheets/FS-016-DFRC.html

here are some more data points for supersonic aircraft and the Space Shuttle (STS):

Aircraft	Mach	Altitude(ft)	Altitude(m)	Pressure(lb/ft²)	Pressure(Pa)
SR-71 Blackbird	3.2	80,000	24,000	0.9	43
Concorde SST	2	52,000	16,000	1.94	93
F-104 Starfighter	1.93	48,000	15,000	0.8	38
Space Shuttle	1.5	60,000	18,000	1.25	60
XB-70	1.5	37,000	11,000	2.5	120

From 'NASA/CR–2011-217077 Measured Sonic Boom Signatures Above and Below the XB-70 Airplane Flying at Mach 1.5 and 37,000 Feet' by Maglieri and Henderson, the first image attached below indicates a range of pressure signatures generated by the XB-70 Valkyrie, and measured by an F-104 Starfighter at various distances, as well as at a static ground station. This is a great example of how complex shocks formed close to supersonic aircraft 'age' as they move through the atmosphere, and by the time they reach the ground, may approach the classic N shape.

This is fine for supersonic flight, but Starship P2P re-entry is likely to be much higher and faster than for any of the above.

From 'AIAA-89-1105 Review of Sonic Boom Theory' by K. J. Plotkin, who references:
1. Seebass. AR., "Hypersonic Boom". Boeing Scientific Research Laboratories Technical Communication 030. June 1970.
and
2. Tiegerman. B., "Sonic Booms of Drag-Dominated Hypersonic Vehicles". Ph.D. Thesis, Cornell University, August 1975.

"Currently, there is one analytic model for sonic boom at hypersonic speeds. It is based on a concept by Seebass that all hypersonic bodies have effectively blunt noses (both physical blunting, such as re-entry vehicles, and aerodynamically because of the entropy layer on slender vehicles), and the resultant drag dominates the sonic boom."

For hypersonic flight (Mach 5 and above), where:

Pg = ambient pressure at ground
ag = ambient sound speed at ground
D = vehicle drag
H = atmospheric scale height
h = flight altitude

the results for the N-wave overpressure Δp, positive phase impulse I, and duration T are:

Δp = 0.59 * ( Pg^5/8 * D^3/8 * exp(-h/8*H) / H^1/4 * h^1/2 )
I = 0.16 * ( Pg^1/4 * D^3/4 * exp(h/4H) / h^1/2 * ag )
T = 1.08 * ( Pg^-3/8 * D^3/8 * exp(3h/8H) * H^1/4 / ag )

"Calculations from this theory agreed well with available data from Apollo 15 and 16
re-entry. Flight Mach numbers were from 4.6 to 15.6. and the Apollo vehicle clearly matches
the postulated model."

From my simulation, the StarshipP2P ricochets are at:

Time(s)	Velocity(m/s)	Mach	Altitude(km)	Altitude(ft)	Drag(kN)
1180	6151	19.0	53.3	175,000	2,870
1589	4810	14.5	51.1	168,000	2,318
1859	3445	10.3	47.8	157,000	1,849
2082	2018	6.3	42.35	139,000	1,359

So, plugging these numbers into the Seebass/Tiegerman model:

flight altitude h(m) =	53300	51100	47800	42350
vehicle drag D(N) =	2870500	2318000	1849000	1359000
N-wave overpressure Δp(Pa) =	42.80135556	41.67204729	41.55288667	42.6139528
positive phase impulse I(Ns) =	1417762.39	1156171.097	915659.7167	657837.1682
duration T(s) =	1.114768856	0.933718945	0.741603564	0.519523846

So, for every ricochet, the overpressure felt at sea level will be about 42 Pa, or 0.9 lb/ft², just under that for the SR-71 Blackbird figure above. The impulse will be greatest at the highest velocity, as will the duration. For people on the ground, the overpressure figure is the most important in terms of perceived volume, and would be quite a bit less than for Concorde.

[OneSpeed]

Here's an update of the L2 StarHopper 200m hop simulation. From recent public footage, it looks like the distance from the launch pad to the new landing pad is about 160m. The audio is credit bocachicagal.

[OneSpeed]

Here's a speculative simulation of Starship Mk1 performing a flight to nearly 20kms in altitude, and returning using a 'Skydiver' profile.

[Jdeshetler]

Here's a speculative simulation of Starship Mk1 performing a flight to nearly 20kms in altitude, and returning using a 'Skydiver' profile.

Amazing simulation! 

With this Skydiver profile, how long will the flight be from launching to landing?

[OneSpeed]

Here's a speculative simulation of Starship Mk1 performing a flight to nearly 20kms in altitude, and returning using a 'Skydiver' profile.

Amazing simulation! 

With this Skydiver profile, how long will the flight be from launching to landing?

Thanks Jay!

You can see the elapsed time in the second row of output. From that, it would take about 1:50 to reach 18.8 kms, and it would land at 5:20, or 3:30 later.

[livingjw]

Here's a speculative simulation of Starship Mk1 performing a flight to nearly 20kms in altitude, and returning using a 'Skydiver' profile.

Amazing simulation! 

With this Skydiver profile, how long will the flight be from launching to landing?

Thanks Jay!

You can see the elapsed time in the second row of output. From that, it would take about 1:50 to reach 18.8 kms, and it would land at 5:20, or 3:30 later.

What direction is your zero reference AoA? Zero AoA appears to be measured relative to forward out the nose, but when you land it appears to be aft out the tail? Also positive AoA would normally be measured with windward surface towards the wind. Are you using the opposite?

John

[OneSpeed]

What direction is your zero reference AoA? Zero AoA appears to be measured relative to forward out the nose, but when you land it appears to be aft out the tail? Also positive AoA would normally be measured with windward surface towards the wind. Are you using the opposite?

John

The zero reference is always the flight path vector, regardless of orientation.

[livingjw]

What direction is your zero reference AoA? Zero AoA appears to be measured relative to forward out the nose, but when you land it appears to be aft out the tail? Also positive AoA would normally be measured with windward surface towards the wind. Are you using the opposite?

John

The zero reference is always the flight path vector, regardless of orientation.

Then how do you know when you are flying backwards (tail end first)?

John

[OneSpeed]

What direction is your zero reference AoA? Zero AoA appears to be measured relative to forward out the nose, but when you land it appears to be aft out the tail? Also positive AoA would normally be measured with windward surface towards the wind. Are you using the opposite?

John

The zero reference is always the flight path vector, regardless of orientation.

Then how do you know when you are flying backwards (tail end first)?

John

I hope I'm understanding your question correctly. It's when the angle between the flight path vector and the centreline of the ship (which is the AoA) exceeds ±90°.

[livingjw]

What direction is your zero reference AoA? Zero AoA appears to be measured relative to forward out the nose, but when you land it appears to be aft out the tail? Also positive AoA would normally be measured with windward surface towards the wind. Are you using the opposite?

John

The zero reference is always the flight path vector, regardless of orientation.

Then how do you know when you are flying backwards (tail end first)?

John

I hope I'm understanding your question correctly. It's when the angle between the flight path vector and the centreline of the ship (which is the AoA) exceeds ±90°.

Maybe I am missing something, but when you are about to land, the AoA is very near zero. I expected it to be around 180 degrees.

[OneSpeed]

Maybe I am missing something, but when you are about to land, the AoA is very near zero. I expected it to be around 180 degrees.

You are absolutely correct, thanks for pointing it out. It looks like there is a bug in the AoA display  when the ship has rotated over 360°, as it did in this profile. I'll fix that ASAP.

Edit: fixed the display issue, attached an updated video.