Multi stage rockets

[MJR]

I made another cylinder similar to the first one I made. My only question is, how do you make a hollow tube? Can you do it with Anim8or? I probably will have a barrage of questions later, so get ready.

 

[n0mad23]

To assist you in your calculations for thrust/fuel etc., here's a tool that could come in handy:

http://www.astrojava.com/rvd/RocketVehicleDesign.jnlp

It's pretty complicated, but can help if you take the time to figure it out.

Good luck!

 

[Urwumpe]

I wonder how hard it would be to make a Java tool of my own, which does the multistage calculations. It is not too hard to calculate such rockets, optimizing them is the problem, as you can get really many variables and really many "local optimal" configurations.

I have some special java programs which optimized special launch vehicle configurations by using genetic algorithms, but I have not yet managed to turn it into a general form for n-staged rockets or parallel staging rockets.

 

[n0mad23]

I wonder how hard it would be to make a Java tool of my own, which does the multistage calculations. It is not too hard to calculate such rockets, optimizing them is the problem, as you can get really many variables and really many "local optimal" configurations.

I have some special java programs which optimized special launch vehicle configurations by using genetic algorithms, but I have not yet managed to turn it into a general form for n-staged rockets or parallel staging rockets.

If you do get around to such a program, I for one will be eagerly awaiting it! Especially if you can figure out some way to minimalize the amount of optimal configurations available. The sheer number of variables can be daunting.

If you do go this route, would you please consider incorporating alternative fuels as well? I'm thinking SpaceX possibilities , and would really love to see tools that would support this kind of exploration here (where an oops moment is non-lethal!).

 

[MJR]

I still haven't found out how to make a hollow tube. I don't think you can do it in Anim8or.

 

[MajorTom]

MJR,

I just uploaded my "beta" work to the Hangar. Hopefully (thanks in advance, Vash!) be available within 24 hr. for you to peruse and play with.

Feel free to make adjustments to the ini values, and add your own meshes and textures.

MT

 

[MajorTom]

I still haven't found out how to make a hollow tube. I don't think you can do it in Anim8or.

Sure you can.

If you draw a straight line at 1m distance from the z axis, and parallel to it, then select that line and then select "build-->lathe" and it will spin this line around the axis you select (z) and create a hollow tube of 2m diameter.

Technique tip: If you want the inside and outside of your tube to look different, make two tubes, one just slightly smaller than the other. The inside tube can have a "framework" (ribs and stringers, pipes and wires, etc.) texture applied, and the exterior of the tube can be clean white, with logos and such.

Another tip: watch out for "flipped normals" where the thing you just made has its "normal" vectors (perpendicular to the surface of each polygon) pointing the wrong way. Result of flipped normals is that the surface looks like a hole, rather than a solid. I think the key N will flip normals for you in Anim8or for the selected object.

EDIT: Anim8or is not helpful when it comes to "flipped" normals because the surfaces look solid in Anim8or...it's not until you get your mesh to Orbiter that you see the problem(s). But sometimes you can detect flipped normals early. In Anim8or, if you are modeling stuff that's for example, white, and you see something that should look white look yellowish instead, it may well be because it is "flipped." Keep an eye out for this.

MT

 

[MajorTom]

I wonder how hard it would be to make a Java tool of my own, which does the multistage calculations. It is not too hard to calculate such rockets, optimizing them is the problem, as you can get really many variables and really many "local optimal" configurations.

I have some special java programs which optimized special launch vehicle configurations by using genetic algorithms, but I have not yet managed to turn it into a general form for n-staged rockets or parallel staging rockets.

I would love to try this Urwumpe.

N0mad, regarding variables being daunting, you're definitely right. I had to make a number of assumptions here and there for Falcon 9 to just get the numbers to "sit still" long enough to understand them. Most of my fiddling yielded an "anemic" rocket capable of lofting only 6000kg to LEO (not the 9900 as advertised) but I may have been too conservative with the thrust and Isp values. The beta version I uploaded to the Hangar has things tweaked hard enough to get an 8500kg Dragon system to LEO with reasonable amounts of fuel remaining (about 0.5% for both stage 1 and 2 at each cutoff).

I'm going to have to take another look at all those numbers soon to make sure I'm not cheating :lol:

MT

 

[MJR]

Sure you can.

If you draw a straight line at 1m distance from the z axis, and parallel to it, then select that line and then select "build-->lathe" and it will spin this line around the axis you select (z) and create a hollow tube of 2m diameter.

Technique tip: If you want the inside and outside of your tube to look different, make two tubes, one just slightly smaller than the other. The inside tube can have a "framework" (ribs and stringers, pipes and wires, etc.) texture applied, and the exterior of the tube can be clean white, with logos and such.

Another tip: watch out for "flipped normals" where the thing you just made has its "normal" vectors (perpendicular to the surface of each polygon) pointing the wrong way. Result of flipped normals is that the surface looks like a hole, rather than a solid. I think the key N will flip normals for you in Anim8or for the selected object.

EDIT: Anim8or is not helpful when it comes to "flipped" normals because the surfaces look solid in Anim8or...it's not until you get your mesh to Orbiter that you see the problem(s). But sometimes you can detect flipped normals early. In Anim8or, if you are modeling stuff that's for example, white, and you see something that should look white look yellowish instead, it may well be because it is "flipped." Keep an eye out for this.

MT

Thank you so much. Earlier you said the hollow tube needs to be the same diameter as the bottom launcher. Do you configure that in the ini file?

 

[MJR]

Sure you can.

If you draw a straight line at 1m distance from the z axis, and parallel to it, then select that line and then select "build-->lathe" and it will spin this line around the axis you select (z) and create a hollow tube of 2m diameter.

Technique tip: If you want the inside and outside of your tube to look different, make two tubes, one just slightly smaller than the other. The inside tube can have a "framework" (ribs and stringers, pipes and wires, etc.) texture applied, and the exterior of the tube can be clean white, with logos and such.

Another tip: watch out for "flipped normals" where the thing you just made has its "normal" vectors (perpendicular to the surface of each polygon) pointing the wrong way. Result of flipped normals is that the surface looks like a hole, rather than a solid. I think the key N will flip normals for you in Anim8or for the selected object.

EDIT: Anim8or is not helpful when it comes to "flipped" normals because the surfaces look solid in Anim8or...it's not until you get your mesh to Orbiter that you see the problem(s). But sometimes you can detect flipped normals early. In Anim8or, if you are modeling stuff that's for example, white, and you see something that should look white look yellowish instead, it may well be because it is "flipped." Keep an eye out for this.

MT

Do I configure the diameter in the ini file? You said it needs to be the same size as the bottom launcher. I also don't really know where the z axis is. How do I measure a meter in Anim8or?

 

[n0mad23]

In Anim8or, scale is relative. Notice in "Options," there's a "grid" setting. At the bottom of the dialog box that opens is something that says, "enable spacing." The default setting here is "1." There's your meter.

Another trick I used a lot was to import one of Dan Steph's astronauts. This is another way of seeing what scale you're working in.

Finally, try this: Instead of making a cylinder with "Lathe," make one using the simple cylinder tool. It's not a mesh yet, and if you double click on the cylinder you've made this way, you'll find it opens a "cylinder editor" dialog box. At the bottom of this box is a field labeled "length." There you go - a direct translation into meters (if that's the relative scale you're using). For example, my current tether-sling cable is 95 km long. In this field, I entered 95000. Double checking it with Mesh Wizard, I find the same figures.

Now remember - if you use a simple cylinder like this, you have to convert it to a mesh using the "Build" - "Convert to Mesh." If you do this, and double click your cylinder you'll find it's now mesh info, and no longer gives the length of your object. In this sense, the simple cylinder is really a better measuring device as it won't give you the same flexibility as "lathe."


And - NO. Your .ini and .cfg setting have nothing to do with the scale of your mesh. The only qualification to this statement (as I understand things) is if you're using spacecraft#.dll's functions, and are scaling the mesh as a piece of animation.

 

[n0mad23]

The beta version I uploaded to the Hangar has things tweaked hard enough to get an 8500kg Dragon system to LEO with reasonable amounts of fuel remaining (about 0.5% for both stage 1 and 2 at each cutoff).

I'm going to have to take another look at all those numbers soon to make sure I'm not cheating :lol:

MT

I'm on my way to the Hanger to give this a try. I'll be looking at your "tweaks" as well, for the simple reason (MJR - are you paying attention?) that this is a great place to understand how things work.

LEO, here I come!

 

[MJR]

Yep. I am paying attention. I am trying my hardest too. So a box is 1 meter?

 

[n0mad23]

Yep. I am paying attention. I am trying my hardest too. So a box is 1 meter?

Pretty much, unless you change things. The good news is even if it isn't a meter (because of changing the settings) you can fix it with several other utilities including Mesh Wizard.

 

[MajorTom]

I'm on my way to the Hanger to give this a try. I'll be looking at your "tweaks" as well, for the simple reason (MJR - are you paying attention?) that this is a great place to understand how things work.

LEO, here I come!

Hey Guys,

Thanks for the Anim8or advice, n0mad. I've become a little rusty with Anim8or as of late. MJR, just make sure your meshes are proportionally correct, as you can always scale them later. I was finding all my meshes coming out at 10x their desired scale...so if my rocket should be 3.6 m in diameter, it was coming out at 36m diameter instead...but then later on I would use Orbiter Mesh Wizard to scale things to the correct size.

n0mad, regarding the tweaks, they're were done mainly to support my selfish need for the spacecraft to reach orbit with the somewhat tentative data I have now. However, more digging has finally revealed the answer to a major head-scratcher I met with:

The SpaceX website gives the F9 mass as 325,000 kg, yet the thrust of one Merlin, at this moment, is anywhere between 347 and 423 kN at sea level (depending on what source you read), so 9 of these would give you between 3123 and 3807 kN. Notice the weight of the F9, at 325,000 kg, is 3204 kN. (Multiply mass by 9.86 to get Newtons) So either the mass is too high, or the Merlin engines are too weak. Worst case, you get a rocket that simply sits on the pad with all 9 engines going flat out, and doesn't even begin to budge until maybe 10% of stage 1 fuel has burned. Lame!

However! Two useful clues.
1. Astronautix believes the mass of the F9 to be 290,000 kg, and Dragon only 8000 kg. (http://www.astronautix.com/lvs/falcon9.htm -- no sources given)
2. Also I saw a press release from SpaceX from Nov 2007, http://www.businesswire.com/portal/...d=news_view&newsId=20071112005019&newsLang=en saying "A planned turbopump upgrade in 2009 will improve the thrust by over 20% and the thrust to weight ratio by approximately 25%." I am not sure if this means improving the low-end, 347 kN thrust by 20% to the high-end, SpaceX website thrust values, or if it means taking the already high-end stated values and increasing them even more.

So more ini file tweaks are in order.

Yes, MJR, the ini or cfg files really don't care what your mesh looks like or how large it is. The height and diameter values in multistage have another purpose. Orbiter will treat the stage as a tube of that diameter and length, with uniform mass distribution (i.e., not all the mass is at one end), for purposes of aerodynamics, and possibly, stability. I notice long, thin rockets are nice and smooth, but short stubby ones tend to wobble when using a guidance program to control their ascent.

Cheers
MT

 

[MajorTom]

MJR, one more hint:

n0mad's Cylinder idea is a good one. What I sometimes do is create a cylinder to the approximate size I want, say 3.6m diameter by 30 m long, then there it is, sitting right there, showing me the size of the "finished product." So if I want instead to make a hollow tube, I can simply "trace" along one edge of the cylinder, then "lathe" the line into a new, hollow cylinder.

If you look at my current stage 1, it's simply a dumb cylinder with a bunch of engine shapes hanging under it. Detail can be filled in later.

MT

 

[n0mad23]

Ground Control -

I 'tweak' the same way, Major Tom. Fast and dirty to get it aloft, and then back to crunch numbers and get it right later. Sometimes this has some pretty drastic effects. For example, in my first ACME tether test, my counter-mass weight was only about 3x that of the Delta-G.

The 40 km version catching atmosphere (from the payload jettison recoil) was pretty spectacular - but not nearly as dramatic as riding the 95 km version down the well. I've got to recreate that one and post the pic on "Ooops Moments." It's like turning an entire hemisphere into a light-bulb with a insanely long filament.

Scaling issues are probably to be expected. I actually released my first (and only) add on initially at twice its real size. I kept feeling something was wrong with it, but it wasn't until I attached a Shuttle to it that I realized it was so big.

So don't worry too much about scale at this point, MJR. Instead make sure that your objects are the same size so they fit together.

MT - wasn't Armadillo's 'unique' approach during the X Prize to have the software jiggle the joystick for stability? I wonder if this is what you've observed about short, stubby rockets versus long slender ones.

 

[n0mad23]

Major Tom -

Wow! For a pre-beta, you're produced an exciting piece of work.

I found myself actually biting my nails nervously watching my fuel as the F9-Dragon sought orbit. After my third test, my nervousness slipped away.

Time after time, the F9-Dragon consistently achieves LEO - seamlessly! I can't help but think you've done some outstanding coding for this thing. I'm really impressed.

Keep after this project, will you? I'm looking forward to subsequent releases.

 

[MJR]

MJR, one more hint:

n0mad's Cylinder idea is a good one. What I sometimes do is create a cylinder to the approximate size I want, say 3.6m diameter by 30 m long, then there it is, sitting right there, showing me the size of the "finished product." So if I want instead to make a hollow tube, I can simply "trace" along one edge of the cylinder, then "lathe" the line into a new, hollow cylinder.

If you look at my current stage 1, it's simply a dumb cylinder with a bunch of engine shapes hanging under it. Detail can be filled in later.

MT

I made a "hollow tube but I am not so sure that it is hollow. Besides that, what is next? I got the bottom launcher, and the hollow tube. Do I make the third stage? Is it going to be smaller, but with the same diameter? Do I use the cylinder method?

 

[n0mad23]

When you make your tube, either as a simple cylinder or done with "lathe," you'll find that the dialog boxes have options (uncheck them) as to whether they have "caps." I believe the default is "Caps."

In terms of size, have you ever seen a multistage that's 3rd stage is bigger than the second, or first? Big to smaller to smallest. Make sure they're the same diameter, unless you're also making fairing - but even then, the top of a lower stage has to connect to the next upper one. See my link above for the java script rocket designer as well for ideas here. Also see the following post http://orbiter-forum.com/showthread.php?t=1168, and scroll down to the "rocket modeler."

For your question about axis - notice on the bottom left corner of Anim8or's work surface there are two arrows. You can't see the third because you're on the same plane. Each arrow represents another axis (X, Y if you're in Front view). If you switch views (i.e., Top) that third arrow will become visible, but one of the others will disappear.