Aerodynamically stable EDL

[sevenperforce]

I've been wanting to build a reusable manned upper stage that can execute a biconic lifting-body re-entry (like the Shuttle or the ITS proposal), but then come down to land on its belly via auxiliary thrusters. However, no matter how much I try, I can't get passive aerodynamic stability. I can hold pretty well through entry, but it tends to flip or spin during descent, so I can never be anywhere close to the right orientation to stick the retropropulsive landing.

Here's the design I've had the most success with so far, albeit without any actual success:

Spoiler

Belly view:

Side view:

Back view:

One action group toggles the airbrakes, deploys the elevons, and opens the rear cargo bay to add drag for EDL:

Another action group shuts down the tail engines, opens the lower bay doors, and activates the Thud aux thrusters for the landing burn:

I've got a probe core stuck in one of the cargo bays to switch control so I can tell it to stay retrograde, belly-first, but when I try, it ends up spinning wildly. What am I doing wrong?

[Das_Sheep]

First guess, ignoring anything to with the aerodynamics as it's an issue I had with horizontal configured ore lifters on Minmus - CoM vs Cot might be out of whack.

You need your centre of mass (CoM) to be aligned with your centre of thrust (CoT). There are display arrows for this in the VAB and the easiest way I found to do it is to tinker with the thrust settings of the engines to move the CoT. Also, if his is what your doing make sure all the fuel tanks that are being used by these engines are all being drained equally or you will run into more issues with thrust balancing.

If it's aerodynamic in nature I have no idea.

You know what, ignore me, never read the OP properly.

[sevenperforce]

16 minutes ago, Das_Sheep said:

First guess, ignoring anything to with the aerodynamics as it's an issue I had with horizontal configured ore lifters on Minmus - CoM vs Cot might be out of whack.

You need your centre of mass (CoM) to be aligned with your centre of thrust (CoT). There are display arrows for this in the VAB and the easiest way I found to do it is to tinker with the thrust settings of the engines to move the CoT. Also, if his is what your doing make sure all the fuel tanks that are being used by these engines are all being drained equally or you will run into more issues with thrust balancing.

If it's aerodynamic in nature I have no idea.

You know what, ignore me, never read the OP properly.

LOL. Yeah, I've got COM/COT down; that's not a problem. The problem is getting center of pressure behind center of mass while dropping belly-first.

[Das_Sheep]

Just, alllll of the reaction wheels and additional RCS?

[sevenperforce]

26 minutes ago, Das_Sheep said:

Just, alllll of the reaction wheels and additional RCS?

Precisely what I don't want to do, haha. Passive aerodynamic stability is the goal here.

[luizopiloto]

rotate those control surfaces to work as wings... do some tests to balance CoM, Drag, Body Lift and Lift... and you will be able to glide a certain distance above 12.000m to adjust your landing point, and then fall like a brick, but stable, after that...

[sevenperforce]

4 minutes ago, luizopiloto said:

rotate those control surfaces to work as wings... do some tests to balance CoM, Drag, Body Lift and Lift... and you will be able to glide a certain distance above 12.000m to adjust your landing point, and then fall like a brick, but stable, after that...

Yeah, that's what I've been attempting, pretty much. Just not having much luck with the control surfaces working the way I want.

[Spricigo]

1 hour ago, sevenperforce said:

LOL. Yeah, I've got COM/COT down; that's not a problem. The problem is getting center of pressure behind center of mass while dropping belly-first.

Which means you need to shift the CoM forward or the CoD rearward. Easier said than done, I know. 

Do  you have a picture with CoM &  CoL for us? 

[sevenperforce]

13 minutes ago, Spricigo said:

Which means you need to shift the CoM forward or the CoD rearward. Easier said than done, I know. 

Do  you have a picture with CoM &  CoL for us? 

Here you go. I replaced the rearmost elevon with a canard that I rotate when I deploy the elevons and airbrakes; that seems to help a little bit. I have control authority turned off for the canards, airbrakes, and elevons, though I'm not sure whether that does more harm than good. Only aerodynamic control authority is the elevons mounted on the side tanks. Should I give the elevons roll authority and give the canards pitch authority?

I've edited the advanced tweakables so the center tanks drain first; once they drain, the CoM stays constant since it is right at the center of the side-mounted tanks. The CoT is shown with the rear engines at zero throttle; I've differentially throttled the aux engines to make sure the CoT is right under the CoM.

[AeroGav]

So,  the OP wants an SSTO rocket that's stable going forwards on the way up, stable going forwards during re-entry, but then at the last minute can flip around and be passively stable going backwards for propulsive landing.

Passive stability  in first one direction, then another,  without staging any parts away and with a similar fuel/payload state - that's hard !

Perhaps near-neutral stability with strong control authority to pitch the craft around on landing is more realistic.

Only way I can think of doing this would be to put some control surfaces with a 30 deg max deflection angle (tail fin connector, big S elevon) right on the nose, set authority to 150%,  but disable pitch/roll/yaw.  "deploy " them on an action group, to add a large amount of drag to the front end, hopefully make it flip on command.  If a service bay is at the front and can be opened at the same time, so much the better.

 

Also, the OP is trying to do aerodynamics on a rocket, with very little wing and mostly body parts.    The problem is the stock centre of lift indicator doesn't take account of aerodynamic forces acting on parts that aren't wings or control surfaces, which, being a rocket , are most of his ship.  Therefore it is completely untrustworthy.   I strongly recommend   the mod   CorrectCoL  for this reason.    Also RCS Build Aid so he can see where his dry CoM is, though as a rocketeer i suspect there's a high chance he's already using it.

[sevenperforce]

7 minutes ago, AeroGav said:

So,  the OP wants an SSTO rocket that's stable going forwards on the way up, stable going forwards during re-entry, but then at the last minute can flip around and be passively stable going backwards for propulsive landing.

Passive stability  in first one direction, then another,  without staging any parts away and with a similar fuel/payload state - that's hard !

Perhaps near-neutral stability with strong control authority to pitch the craft around on landing is more realistic.

Only way I can think of doing this would be to put some control surfaces with a 30 deg max deflection angle (tail fin connector, big S elevon) right on the nose, set authority to 150%,  but disable pitch/roll/yaw.  "deploy " them on an action group, to add a large amount of drag to the front end, hopefully make it flip on command.  If a service bay is at the front and can be opened at the same time, so much the better.

Also, the OP is trying to do aerodynamics on a rocket, with very little wing and mostly body parts.    The problem is the stock centre of lift indicator doesn't take account of aerodynamic forces acting on parts that aren't wings or control surfaces, which, being a rocket , are most of his ship.  Therefore it is completely untrustworthy.   I strongly recommend   the mod   CorrectCoL  for this reason.    Also RCS Build Aid so he can see where his dry CoM is, though as a rocketeer i suspect there's a high chance he's already using it.

Well, slight correction -- I don't need an SSTO; this is a second stage. I've been using it bare as a single suborbital stage for testing purposes, but I would throw it on top of a booster for actual launches.

Also, I want nose-first prograde stability going up (which isn't too hard, since I have Vectors on the tail), but I want biconic passive aerodynamic stability coming down. I'm landing on auxiliary thrusters inside downward-facing cargo bays; it lands on its belly like a Star Wars ship. Nose-first on ascent, belly-first on entry, descent, and landing.

I'm not actually using RCS Build Aid (I don't have any mods, in fact; I have been playing demo KSP for about a year but I just got the full game a couple of weeks ago), but since I've set the side-mounted tanks to drain last, COM remains constant after the center tanks drain.

[Boris-Barboris]

_____
  |__|     - meh

\___/
 |__|      - better

nvm. you've already done it. I can only suggest more.

Edited May 18, 2017 by Boris-Barboris

[Spricigo]

16 minutes ago, AeroGav said:

   The problem is the stock centre of lift indicator doesn't take account of aerodynamic forces acting on parts that aren't wings or control surfaces, which, being a rocket , are most of his ship.  Therefore it is completely untrustworthy. 

 

Pretty much this.

You need to consider the surface area exposed to airstream ahead and behind CoM, if there is more area ahead it will flip. Bringing CoM closer to the geometric center may help, I would consider moving the vectors to the side stack and forward.

A side note: 7 thuds weight a lot. Exploring different engines alternatives may result in a lighter but still functional landing propulsion.

[AeroGav]

4 minutes ago, Spricigo said:

A side note: 7 thuds weight a lot. Exploring different engines alternatives may result in a lighter but still functional landing propulsion.

He'll be landing back on Kerbin almost empty, right? So it should be possible to do so with far less in the way of engine.

[StahnAileron]

@AeroGav Is probably on to something regarding body-lift aspect of KSP. KSP models body lift for parts not explicitly defines as aero-lift parts (a.k.a. "wings"). For designs that rely heavily on aerodynamic characteristics beyond static lift in straight line flight (be it a dive, climb, or level flight), the basic aero indicator is wholly non-indicative of the inherent stability of the craft is certain flight regimes.

I had this problem when I was building a shuttle design. It would flip the hell out if I exceeded a certain AoA on re-entry. Simply put, the MK3 body would start generating sufficient body lift that the CoL/CoP move forward ahead of the current CoM. It took waiting for dynamic pressure to get low enough for me to regain control, if at all.

Judging from the screenshot of the design in the editor, this is probably the problem you're facing. The CoL indicator only accounts for the control surfaces. If you're going belly first with that design, you will be generating A LOT of body lift. Probably far more than what the control surfaces can compensate for. (Besides, the CoL indicator doesn't give you a real sense of magnitude either. Just because you have the CoL in a supposedly stable point doesn't mean its exerting STRONG stabilizing forces.) You have a lot of body area in front of the CoM and not as much behind (Looks about 3:2 ratio Front:Back). This is probably the cause of the instability.

Test the design but with Aero Forces (F12) on so you can see what is generating lift and how much of it in the flight profile you want. This will inform you as to how to correct the CoL and resulting instability issue. This is how I figured out, on my own, what was screwing over my shuttle design. (Well, once I remembered body lift was a thing; I was mostly a MK2 user before that attempt at a more typical shuttle design.)

This is part of the reason I have better appreciation for the MK2 design profile: it accounts for "body lift" in the editor already by design (in most cases).

BTW, maybe it would be easier to build this in the SPH first than import to the VAB later for booster integration? I find the VAB horrible for aerodynamic-dependent designs. (It's not as intuitive to me since everything is rotated/aligned to the vertical by default.)

[Spricigo]

6 minutes ago, AeroGav said:

He'll be landing back on Kerbin almost empty, right? So it should be possible to do so with far less in the way of engine.

To be honest, I gave low consideration for the whole weight issue.

to the OP: what is the intended purpose of the vessel? How much cargo/crew you want to transport and to which orbital height? Maybe we can figure out some way to reach your goal with less.

[sevenperforce]

1 hour ago, StahnAileron said:

Judging from the screenshot of the design in the editor, this is probably the problem you're facing. The CoL indicator only accounts for the control surfaces. If you're going belly first with that design, you will be generating A LOT of body lift. Probably far more than what the control surfaces can compensate for. (Besides, the CoL indicator doesn't give you a real sense of magnitude either. Just because you have the CoL in a supposedly stable point doesn't mean its exerting STRONG stabilizing forces.) You have a lot of body area in front of the CoM and not as much behind (Looks about 3:2 ratio Front:Back). This is probably the cause of the instability.

Test the design but with Aero Forces (F12) on so you can see what is generating lift and how much of it in the flight profile you want. This will inform you as to how to correct the CoL and resulting instability issue. This is how I figured out, on my own, what was screwing over my shuttle design. (Well, once I remembered body lift was a thing; I was mostly a MK2 user before that attempt at a more typical shuttle design.)

This is part of the reason I have better appreciation for the MK2 design profile: it accounts for "body lift" in the editor already by design (in most cases).

BTW, maybe it would be easier to build this in the SPH first than import to the VAB later for booster integration? I find the VAB horrible for aerodynamic-dependent designs. (It's not as intuitive to me since everything is rotated/aligned to the vertical by default.)

I've never used Aero Forces before but I'll activate it and take a look. As for the SPH vs the VAB, I've been playing in Demo for so long that I have no trouble thinking vertically in the VAB. I built Demo shuttles in that thing, haha.

33 minutes ago, Spricigo said:

To be honest, I gave low consideration for the whole weight issue.

to the OP: what is the intended purpose of the vessel? How much cargo/crew you want to transport and to which orbital height? Maybe we can figure out some way to reach your goal with less.

I want a multipurpose interplanetary reusable second stage with integrated crew and cargo capability. The idea is to refuel this in orbit and send it pretty much anywhere in the solar system; when needed, I'll send a tanker along with it to refuel at the destination. The aux thrusters will allow me to make propulsive landings on unprepared surfaces. That's easy enough to do on the Mun, etc., but much more challenging on Earth or Duna, where I need to take aerodynamics into account.

The purpose of high thrust on the aux thrusters is to be able to lift off vertically on Duna with enough propellant to reach orbit.

[bewing]

Passive stability comes from having heavy stuff at the front end, low drag at the front end, and high drag at the back end. If you are trying to fly this thing belly first -- then it has no front end or back end, and there is no passive stability possible, really.

Even worse, the belly of your vessel has enormous drag, because it's got a huge area and it's blunt. So if you are going belly first, that means all the drag is in the front -- you would need a wing area at the top/back that is greater in size than the entire area of the belly of your ship (plus the area of your nacelles!) in order to move the CoP above the CoM. As far as mass goes, you've got it balanced. But it needs to be as far toward the belly as you can get it -- without moving the draggy bits any farther down. Which is basically impossible.

So basically, this thing is guaranteed to be massively unstable in that orientation.To force it to stay that way anyhow, you're going to need serious active stabilization, or about a hundred times the wing area you've got now.

 

[Spricigo]

@bewing the belly is blunt and draggy but so is the top and mass is fairly symmetrical up/down, not difficult to make the top just a bit more draggy and get the thing sorted along that axis. However the longitudinal axis present about 50% more area ahead of CoM and need rework.

@sevenperforce The problem with multipurpose designs its that you end up being suboptimal in everything, sometimes even bad at most things. That, unfortunately, is the case with your vessel. The engines are bad for use in space, you have to much thrust for small celestial bodies (mun, minmus), not enough deltaV to go much far from Kerbin SoI, and tthe list goes on. I strongly suggest to narrow down to a more specific purpose (e.g landing and return from Duna; or 2nd stage to LKO and reentry).  As it stands now its even difficult  to us to give advice.

edit: also, if there is constraints in regard to technology or cost.

Edited May 18, 2017 by Spricigo

[sevenperforce]

2 minutes ago, Spricigo said:

@bewing the belly is blunt and draggy but so is the top and mass is fairly symmetrical up/down, not difficult to make the top just a bit more draggy and get the thing sorted along that axis. However the longitudinal axis present about 50% more area ahead of CoM and need rework.

@sevenperforce The problem with multipurpose designs its that you end up being suboptimal in everything, sometimes even bad at most things. That, unfortunately, is the case with your vessel. The engines are bad for use in space, you have to much thrust for small celestial bodies (mun, minmus), not enough deltaV to go much far from Kerbin SoI, and tthe list goes on. I strongly suggest to narrow down to a more specific purpose (e.g landing and return from Duna; or 2nd stage to LKO and reentry).  As it stands now its even difficult  to us to give advice.

With on-orbit refueling, I have plenty of dV for pretty much anywhere. The Vector's aren't ideal for space, true, but I can replace them with Darts once I get the design perfected; I just need the Vectors to get off the ground. 

First goal is second stage to LKO and propulsive biconic EDL. The notion of using the same design for executing propulsive EDL followed by SSTO on Duna is secondary.

[sevenperforce]

7 minutes ago, bewing said:

Passive stability comes from having heavy stuff at the front end, low drag at the front end, and high drag at the back end. If you are trying to fly this thing belly first -- then it has no front end or back end, and there is no passive stability possible, really.

Even worse, the belly of your vessel has enormous drag, because it's got a huge area and it's blunt. So if you are going belly first, that means all the drag is in the front -- you would need a wing area at the top/back that is greater in size than the entire area of the belly of your ship (plus the area of your nacelles!) in order to move the CoP above the CoM. As far as mass goes, you've got it balanced. But it needs to be as far toward the belly as you can get it -- without moving the draggy bits any farther down. Which is basically impossible.

So basically, this thing is guaranteed to be massively unstable in that orientation.To force it to stay that way anyhow, you're going to need serious active stabilization, or about a hundred times the wing area you've got now.

Hmmmm, interesting.

What if I try flying it at a very high AOA rather than truly belly-first? E.g., gliding like a brick. That might be able to give me passive stability, and since the Thuds have a decent gimbal range, I can cancel my horizontal velocity pretty easily.

Doing some redesign work...

Okay, here, this seems to work a bit better. 

Spoiler

Cleaned it up. The secondary tanks are now in clipped Mk2 body segments on the sides, slightly lower than the centerline.

In place of the airbrakes and elevons for feathering, I have fixed wing strakes. The canards on the tail deploy to feather for entry. They also have pitch authority, and the elevons on the end of the strakes act as ailerons for roll authority.

I now have the Vectors mounted on the side sections rather than on the tail, which will help with first-stage mating stability and also allow me to fire them on the pad for extra TWR. Gotta love crossfeed. There's a single large cargo bay just forward of the dry COM; the rest of the main body is fuel.

Closeup showing ailerons and canards. The canards look off-kilter but they're not; the bottom segment of the side engines is rotated to place the main engine CoT a bit further back.

The Thuds have been relocated to Mk2 cargo bays. They now have slight cosine losses but it's not too bad.

Prograde drag is quite low and lift is pretty much zero. Not too concerned about aerodynamic stability on ascent; this is going to be a second stage when I actually use it.

It does tend to flip belly-first on ascent if I do the gravity turn too early...but, again, I'm not terribly concerned right now.

I've got a docking port buried in the cargo bay so I can use "Control from here" to hold belly-first entry.

Setting the canards to "deployed" to feather for re-entry. Hopefully this forces the tail up. Right now I'm in near-exoatmospheric free-fall.

Flipping around to the desired orientation for entry. One problem with using that embedded docking port to "control from here" is that there's nothing controlling which direction the nose points. Might want to try staying with the nose-first control and telling it to point prograde, then allowing the feathered canards to keep the nose up.

Aerodynamic stability seems to be holding a bit better here. This is with the canards given pitch authority. 

Nice shot of all the aerodynamic interactions. I don't know what they all mean yet, though.

Through re-entry heating, and still holding pretty nicely!

Once speed drops lower, though, it noses down. Opening the bay doors to try and pull the nose up.

.....and it just starts spinning wildly. Not sure why. Throttling up the Thuds only makes it spin faster and faster until I lawn-dart into the drink.

So, no luck there. I'll fly the same mission but try inverting the canards once I make it through re-entry.

Seems like I'm making progress...

[AeroGav]

24 minutes ago, sevenperforce said:

 

Sorry for initially misunderstanding how you wanted the thing landing,  when I heard "propulsive" I immediately thought SpaceX.    It basically flies like a normal rocket on the way up and like a normal airplane on re-entry, but at some point it makes like a VTOL airplane and comes to a stop in mid air, then settles down to the ground on propulsive thrust, is that correct?      

In which case it is just a normal airplane except for the fact that there's almost no wing and you're mostly relying on body lift - at landing weights, it should still be possible to pull out of the dive and flare to almost level flight  at 150 m/s or so.    Then fire up the vertical engines and gradually throttle them up as lift dies down.       As a stable airplane it wants to go prograde so if a really high vertical descent rate appears with no horizontal movement, ithe aerodynamics will try to push the nose down and at some point that'll overwhelm your RCS/reaction wheels/vector thrust.     However, you aren't intentionally going to get into that situation anyway on landing.

I recommend you get RCS build aid and correct CoL.  

That second design has two four ton vectors at the back and one four ton cockpit up front, so i suspect when empty the CG is aft of halfway back.  The cargo bay is forward of the halfway point.  I suspect that when you open it, you have more stuff catching the wind ahead of CG than behind it.    The design might be unstable in yaw, seeing it spin with the nose pointing due west?

Again,  the blue indicator in the VAB only shows the effects of the strakes and control surfaces you are putting on, not the body lift or the drag from open cargo bays - that's where CorectCoL really helps.

[sevenperforce]

1 minute ago, AeroGav said:

Sorry for initially misunderstanding how you wanted the thing landing,  when I heard "propulsive" I immediately thought SpaceX.    It basically flies like a normal rocket on the way up and like a normal airplane on re-entry, but at some point it makes like a VTOL airplane and comes to a stop in mid air, then settles down to the ground on propulsive thrust, is that correct?

Yeah, exactly. Although I was thinking it would stay nosed up a little on landing approach; basically, stable at a high angle of attack due to body lift.

7 minutes ago, AeroGav said:

In which case it is just a normal airplane except for the fact that there's almost no wing and you're mostly relying on body lift - at landing weights, it should still be possible to pull out of the dive and flare to almost level flight  at 150 m/s or so.    Then fire up the vertical engines and gradually throttle them up as lift dies down.

Hmm, I had been trying to punch the vertical engines, suicide-burn style. I'll try throttling them up more gradually and seeing if I can do a more controlled landing approach.

8 minutes ago, AeroGav said:

That second design has two four ton vectors at the back and one four ton cockpit up front, so i suspect when empty the CG is aft of halfway back.  The cargo bay is forward of the halfway point.  I suspect that when you open it, you have more stuff catching the wind ahead of CG than behind it.    The design might be unstable in yaw, seeing it spin with the nose pointing due west?

I don't open the main cargo bay on EDL. Though yeah, it's a touch unstable in yaw. Going to more level flight rather than trying to do a straight-down drop has helped with that.

I'm intentionally not using any RCS or reaction wheels during the descent or landing approach. If I'm throttling up the aux thrusters slowly, I might kick on the RCS at that point, just to help a bit. 

[sevenperforce]

A few changes. Going at it with the canards inverted for landing approach.

Spoiler

 

Canards feathered. I pulled up the clamp-o-tron port so I could switch controls quickly, but I'm leaving control with the nose for now, so I can tell it to hold stability assist and it will hopefully hold attitude. Pitch authority on the canards is off.

Nosing down slowly but still holding pretty well. I think these aero lines look good, right? I suppose I could tell it to hold radial-out and let the feathered canards push the tail up.

Re-entry heat fading. Are the light blue lines lift indicators?

Moving from Entry to Descent, so I feathered the canards back in the opposite direction and switched to the internal docking port for control position. Hopefully, the fact that I already have a good bit of horizontal velocity will keep me pointing in the same overall direction rather than yawing wildly. I guess that's what wings are for, right?

Not getting anywhere near the retrograde marker, but at least it's quite stable; the control authority on the elevon/ailerons is more than enough to keep it from tilting too far in either direction.

Down to 1.5 km and holding the horizon pretty nicely even though SAS is trying to keep the internal docking port to retrograde. Very stable.

Guessing I should open up the bay doors and fire up those Thuds right about now. If you squint, you can just barely see where the retrograde indicator is on the Navball.

Opening them up, holding steady....

Aaaaaand as soon as I try to throttle them up, it rolls hard to one side and crashes. Not sure why. They are all in perfect mirror alignment so there shouldn't be any problem with that.

Loving Bill's face here.

 

[Spricigo]

1 hour ago, sevenperforce said:

.....and it just starts spinning wildly. Not sure why.
 

The craft started to flip before 30km, that seems to be the point where aerodynamic forces and reaction wheels gave up. Your SAS is desperately trying to point you back to the desired orientation without success resulting in the spin.

The balance of weight and drag is still off. You can try to move fuel around to adjust the CoM, but better yet is to try to make it remain unaltered regardless of fuel level and loaded cargo.

Get rid of the vector, use 4 darts. This will put your CoM right in the middle of your vessel (provided you keep the cargo bay also there) which in turn will make balancing the drag much more easily.

Edited May 18, 2017 by Spricigo
ninja'ed by the Op, that pulled it off. Gratz!