Theoretical Lagrangian Points?

[Iforgotthis]

Now I know KSP does not work in the same way as real life; there are not Lagrangian points. But, let's talk theoretically.

Let's say that KSP functions in the same way as the real solar system, and Lagrangian points are permitted to exist. What altitudes would they sit at in the system? Kerbin is smaller than Earth, would this affect the position of the points? As well as this, Kerbin has two moons, would that affect the Lagrangian points in the Kerbin system?

It'd be interesting to know what you think!

[Kasuha]

You can find all the math needed for that on Wikipedia.

[Claw]

The best you can do to simulate Lagrange points in KSP is with L4 and L5. If you put a craft just in front (or just behind) the smaller body's SOI and put it in the same orbit at the same speed, it will sort of act like an L4/L5 point. Unfortunately it isn't "stable" and you'll still have to station keep. From there, you can simply bump the craft's orbit out slightly (if in L4) or in slightly (if in L5) and you can enter the smaller body's SOI. This will allow gravity slingshots but that's about it.

I believe someone is working on an n-body plugin.

[Sirrobert]

The best you can do to simulate Lagrange points in KSP is with L4 and L5. If you put a craft just in front (or just behind) the smaller body's SOI and put it in the same orbit at the same speed, it will sort of act like an L4/L5 point. Unfortunately it isn't "stable" and you'll still have to station keep. From there, you can simply bump the craft's orbit out slightly (if in L4) or in slightly (if in L5) and you can enter the smaller body's SOI. This will allow gravity slingshots but that's about it.

I believe someone is working on an n-body plugin.

Would that even be possible?

[stoani96]

as far as i know, unity has a built-in function....

at least i've found an asset for n-body particle simulation (hardware driven) some time ago (some sort of GPU gems), so it should not be too difficult to build it in, but i think, that squad is just too "lazy" to do this; not only the work, but also the changes in gameplay and the testing, that this affords.

sorry for my english

[Sirrobert]

as far as i know, unity has a built-in function....

at least i've found an asset for n-body particle simulation (hardware driven) some time ago (some sort of GPU gems), so it should not be too difficult to build it in, but i think, that squad is just too "lazy" to do this; not only the work, but also the changes in gameplay and the testing, that this affords.

sorry for my english

Right, they are to lazy....

Offcourse let's not count the fact that timewarp would be impossible with n-body physics

[Claw]

Would that even be possible?

Would what be possible? n-body? I'm not versed on Unity, but I don't see why not. Internal to KSP it was a matter of choosing how to implement physics, and Squad chose SOIs. I don't know how easy that is to override (in a similar fashion to FAR overriding stock aerodynmaics). Although I suspect a lot of fidelity will be lost in the floating point errors.

KSP is pretty hard on the CPU already, so I don't know how much worse doing n-body would be...

as far as i know, unity has a built-in function....

at least i've found an asset for n-body particle simulation (hardware driven) some time ago (some sort of GPU gems), so it should not be too difficult to build it in, but i think, that squad is just too "lazy" to do this; not only the work, but also the changes in gameplay and the testing, that this affords.

sorry for my english

I would say it's not a matter of being "lazy." Spheres of Influence is considered a valid form of simplification for orbital mechanics. It works pretty well in most cases. Lagrange points is one big failure. Also, craft won't experience orbital tides. But then again, it's just a game and it probably isn't the extra physics load to elongate a craft's orbit ever so slightly. (Especially when my cooling fans run at full speed constantly already...)

[CosmosX]

First, i've gotta say that there could be some points already said by other people.

you said physics would work same as real world? then where could be no L points. since there are two moons, the L point will be unstable as time passes. It's really easy to caculate L point between two points, but between three points would be almost impossible.

Now, let's make this problem little bit easier. if minmus diapeares, the L point will have same position with Earth-Moon L point. since all gravitational forces are same in KSP with real world(Kerbin's g = Earth's g, Mun's g = Moon's g), physical caculation will be exactly same. for sure, this work has some problems since radiuses are different. BUT, differences are so small that we can ignore that.

I might have some problems with my thoughts, so tell me if you see any problems.

[Iforgotthis]

You can find all the math needed for that on Wikipedia.

I'm sorry, that's going to take some explaining. I like maths, but that's a little confusing.

[cantab]

The Kerbin-Mun L points would still be there. Minmus might make them unstable, but I expect its mass will be low enough to not matter, at least not for L1 and 2.

[Kasuha]

I'm sorry, that's going to take some explaining. I like maths, but that's a little confusing.

L3, L4, and L5 are obvious. You can even place some satellites to those points and they will stay there as long as you make sure their orbital period is the same as Mun's.

L1 and L2 are solutions of an equation and explanation of all values is in there. For example L1:

The location of L1 is the solution to the following equation balancing gravitation and centrifugal force:

where r is the distance of the L1 point from the smaller object, R is the distance between the two main objects, and M1 and M2 are the masses of the large and small object, respectively. (The quantity in parenthesis on the right is the distance of L1 from the center of mass.)

So you just need to look up KSP Wiki for values for M1, M2, and R and solve the equation for r - you can try to use Wolfram Alpha for that.

[Claw]

Also, if we're just talking theoretical and discount KSP's inability (as stated in the OP), all of the Lagrange points would be effected by the existance of Minmus in the Kerbin-Mun system. The Lagrange points in the wiki are built off a two body system with a dominant main body and the secondary body (as well as the satellite that sits in L-points, but it is negligible.)

If you throw Minmus into there, all the points would be effected and some of them might be even less stable than they already are. L4 and L5 would probably shrink to relatively small areas, if they would even exist because Minmus would tend to pull objects out and drag them forward in orbit. L5 objects would be thrown into Mun and L4 objects would be deflected into a higher orbit.

[Vanamonde]

squad is just too "lazy" to do this

The company simulating real-time spaceflight with calculus and rocket science and whatnot is too lazy to do some math? That's an interesting chain of logic you've got going there.

Meanwhile, discussing Lagrange points is a science matter, so, thread moved. Asking for them and N-body in the game is on the What Not to Suggest list, so please restrict the discussion to the science side.

[Sirrobert]

Would what be possible? n-body? I'm not versed on Unity, but I don't see why not. Internal to KSP it was a matter of choosing how to implement physics, and Squad chose SOIs. I don't know how easy that is to override (in a similar fashion to FAR overriding stock aerodynmaics). Although I suspect a lot of fidelity will be lost in the floating point errors.

KSP is pretty hard on the CPU already, so I don't know how much worse doing n-body would be...

I meant an N-body plugin that someone else wants to make

[Ralathon]

Would that even be possible?

Yea, it is certainly possible, although the solvers are a bit of a challenge under high timewarp. Right now KSP uses patched conics to simulate orbits and those easily permit time warp. Imagine a circular orbit that has a period of 2 hours. Now you can plug in any time and you can calculate exactly where your spaceship will be. After 5 hours it will have completed exactly 2 and a half orbits. You don't need to calculate what the spacecraft was doing in those 5 hours, all you need to figure out its position is the time.

Multibody physics doesn't give such a neat result. At any point in time there are multiple bodies pulling on your ship causing your orbit to change. So now you need to model every step the spacecraft makes along its orbit to figure out where it is after those 5 hours. This isn't that hard to calculate in real time or low time warp, it is fairly cheap on the CPU. The problem starts with high timewarp. At the highest timewarp you need to calculate a ridiculous amount of steps to get an accurate guess of where your spaceship is per physics step. Now add in the extra challenge of doing this for all the debris and other spaceships you have running in the background, and your performance will tank.

There is another reason why this isn't implemented in the stock game: It adds very little in terms of gameplay.

The only real differences if KSP used n body physics would be:

- Lagrange points.

- orbital drift.

- some fancy new types of orbits, like halo orbits.

Lagrange points or halo orbits sound really cool, but they wouldn't really add that much to the game. You could hang a station or probe there to do... what? On the other hand n-body physics would also make your orbits unstable. So every time you timewarp to Duna your entire satellite constellation around Kerbin is out of whack thanks to the Mun's gravity. This means you'll spend a lot of time stationkeeping, not the most interesting task in the world.

[Iforgotthis]

The company simulating real-time spaceflight with calculus and rocket science and whatnot is too lazy to do some math? That's an interesting chain of logic you've got going there.

Meanwhile, discussing Lagrange points is a science matter, so, thread moved. Asking for them and N-body in the game is on the What Not to Suggest list, so please restrict the discussion to the science side.

Oops... Apologies, I should have looked at the different forums first!

[Sirrobert]

Yea, it is certainly possible, although the solvers are a bit of a challenge under high timewarp. Right now KSP uses patched conics to simulate orbits and those easily permit time warp. Imagine a circular orbit that has a period of 2 hours. Now you can plug in any time and you can calculate exactly where your spaceship will be. After 5 hours it will have completed exactly 2 and a half orbits. You don't need to calculate what the spacecraft was doing in those 5 hours, all you need to figure out its position is the time.

Multibody physics doesn't give such a neat result. At any point in time there are multiple bodies pulling on your ship causing your orbit to change. So now you need to model every step the spacecraft makes along its orbit to figure out where it is after those 5 hours. This isn't that hard to calculate in real time or low time warp, it is fairly cheap on the CPU. The problem starts with high timewarp. At the highest timewarp you need to calculate a ridiculous amount of steps to get an accurate guess of where your spaceship is per physics step. Now add in the extra challenge of doing this for all the debris and other spaceships you have running in the background, and your performance will tank.

There is another reason why this isn't implemented in the stock game: It adds very little in terms of gameplay.

The only real differences if KSP used n body physics would be:

- Lagrange points.

- orbital drift.

- some fancy new types of orbits, like halo orbits.

Lagrange points or halo orbits sound really cool, but they wouldn't really add that much to the game. You could hang a station or probe there to do... what? On the other hand n-body physics would also make your orbits unstable. So every time you timewarp to Duna your entire satellite constellation around Kerbin is out of whack thanks to the Mun's gravity. This means you'll spend a lot of time stationkeeping, not the most interesting task in the world.

Thanks, I always just asumed patched conics were hardcoded in the game and weren't so easy to change

[going-to-the-MOON]

At the highest timewarp you need to calculate a ridiculous amount of steps to get an accurate guess of where your spaceship is per physics step. Now add in the extra challenge of doing this for all the debris and other spaceships you have running in the background, and your performance will tank.

If the performance drops from that, then why not use patched conics for other vessels, and n-body for only the current vessel? Would that not affect performance as much? (Would also take away the task of station keeping.)

[R0cketC0der]

If the performance drops from that, then why not use patched conics for other vessels, and n-body for only the current vessel? Would that not affect performance as much? (Would also take away the task of station keeping.)

Sure, but then what's the point of even adding it? Halo orbits wouldn't be possible, orbital drift wouldn't happen and if you leave a vessel on a transfer trajectory your orbit prediction would be totally wrong.

[going-to-the-MOON]

True.

In any case, what is the likeliness of n-body physics actually being officially added into the game?

[sjwt]

True.

In any case, what is the likeliness of n-body physics actually being officially added into the game?

Its on the what not to suggest list.

Any 'N-body' solutions out there are either lacking in one of or more then one of the following features

1) 3d

2) 4d

3) More then 3/X bodies

4) Calculating mass of all bodies

5) Real time

6) Adaptable to more then its specific situation.

7) Any other way they can cheat to make a headline.

[Nibb31]

Zero. .