Patched COnics

[YourEverydayWaffle]

Hey all.

I've been wondering about this. What are patched conics, and how do they apply in both KSP and real life? I haven’t really been able to find a good explanation lately that isn't incredibly complex. I’m sure somebody here knows. Thanks in advance!

[wasmic]

Well, patched conics don't apply in real life. It's an approximation of real-life orbital mechanics, which can get very complex for a computer to handle.

The system that KSP uses with SOIs are what's called Patched Conics Approximations. This means that while you are in Kerbins SOI, Kerbin and only Kerbin will tug in your spacecraft. IRL, if you were in orbit of the Earth, both the Earth, the Moon, the Sun, Jupiter and every other massive object in the solar system would tug at your spacecraft at the same time. They don't in KSP.

This means that in KSP, there are no such things as lagrange points and orbit precession. Also, orbits with an eccentricity of barely above 1 will be unrealistic. Mostly, though, KSPs approximation is very close to the RL n-body gravitation. n-body gravitation CAN be simulated in a game, but it was dropped out of KSP for performance reasons.

[YourEverydayWaffle]

Well, patched conics don't apply in real life. It's an approximation of real-life orbital mechanics, which can get very complex for a computer to handle.

The system that KSP uses with SOIs are what's called Patched Conics Approximations. This means that while you are in Kerbins SOI, Kerbin and only Kerbin will tug in your spacecraft. IRL, if you were in orbit of the Earth, both the Earth, the Moon, the Sun, Jupiter and every other massive object in the solar system would tug at your spacecraft at the same time. They don't in KSP.

This means that in KSP, there are no such things as lagrange points and orbit precession. Also, orbits with an eccentricity of barely above 1 will be unrealistic. Mostly, though, KSPs approximation is very close to the RL n-body gravitation. n-body gravitation CAN be simulated in a game, but it was dropped out of KSP for performance reasons.

Ah, I see. So it's the simplification of physics involving SOIs. Thanks.

[Idobox]

It mostly allows analytic solutions. It means that rather than compute the new position every time step (like 24 times per second); with errors accumulating, you can tell where the ship will be at a time t by using a relatively simple formula.

Without patched conics, it would be impossible to have time acceleration (just physics acceleration), and the computer would have to determine the position of every ship, debris, etc at each time step.

[pizzaoverhead]

For the visual learners: The orbits can be thought of as sections of a cone, or "conic sections", where a plane intersects a cone. They're useful because they model simple two-body orbits well. You can see how the angle of the plane changes a circular orbit into an elliptic orbit, and that a parabola or hyperbola is an escape orbit.

Patched conics join orbits from different spheres of influence together. There are a few different ways of joining conics from different bodies together, depending on whose point of view you're seeing it from. You can choose which one to use by changing the patched conics setting in game. You can see what these are like in this thread.

[TechnicalK3rbal]

For the visual learners: The orbits can be thought of as sections of a cone, or "conic sections", where a plane intersects a cone. They're useful because they model simple two-body orbits well. You can see how the angle of the plane changes a circular orbit into an elliptic orbit, and that a parabola or hyperbola is an escape orbit.

Patched conics join orbits from different spheres of influence together. There are a few different ways of joining conics from different bodies together, depending on whose point of view you're seeing it from. You can choose which one to use by changing the patched conics setting in game. You can see what these are like in this thread.

http://upload.wikimedia.org/wikipedia/commons/4/48/Conic_sections_2.png

Darn, you beat me to it.

[YourEverydayWaffle]

That's cool. I was actually learning about Pascal's studies of conics today.

[KASASpace]

All things parabolic, hyperbolic, elliptical, and circular are conic sections.

So, using a pair of cones and intersecting a plane with them you get one of those conic sections depending on what angle the plane is.

Now, since orbits are elliptical (one of Kepler's Laws.....),

and Ellipses are conics sections....

You should see the connection.

You basically use these cones to draw the orbit.

Okay, patched conics is multiple conic sections, with "patch points".

These patch points indicate SOI change, and a new conic is drawn when you get there.

[Eric S]

Ah, I see. So it's the simplification of physics involving SOIs. Thanks.

Right. Patched conics may be an approximation, but they're a good one for most cases. The Apollo missions were actually planned out using patched conics.

[FirstSecondThird]

I learned in my astrophysics classes back in college that patched conics is not an accurate representation for physics at all.

[magnemoe]

I learned in my astrophysics classes back in college that patched conics is not an accurate representation for physics at all.

It depend on the forces acting on you. Going too the Mun is pretty realistc, you can do stuff like free return trajectories.

Patched conics has two benefits, its easy to calculate, on high time warp the game only have to check if an object pass within an SOI, if not it just has to update the position after an formula.

With n-body physic you would need constant calculations on all objects, how longer between the calculation who higher error rate, in short forget high time warps.

Second issue is that it would make the game less predictable, you go to Minmus and pass Mun larenger point and is thrown off target even if Mun is 60 degree ahead. Going to Dress and Jool make you miss a lot.

[Wavechaser]

Well, patched conics don't apply in real life. It's an approximation of real-life orbital mechanics, which can get very complex for a computer to handle.

The system that KSP uses with SOIs are what's called Patched Conics Approximations. This means that while you are in Kerbins SOI, Kerbin and only Kerbin will tug in your spacecraft. IRL, if you were in orbit of the Earth, both the Earth, the Moon, the Sun, Jupiter and every other massive object in the solar system would tug at your spacecraft at the same time. They don't in KSP.

This means that in KSP, there are no such things as lagrange points and orbit precession. Also, orbits with an eccentricity of barely above 1 will be unrealistic. Mostly, though, KSPs approximation is very close to the RL n-body gravitation. n-body gravitation CAN be simulated in a game, but it was dropped out of KSP for performance reasons.

Even the L points don't really exist, the points right in front/back of Kerbin, and share the same orbit with it, would be extremely useful for communication relays.

[Wavechaser]

I learned in my astrophysics classes back in college that patched conics is not an accurate representation for physics at all.

True, but if it simulates every body in the Kerbol system, the game would most likely crash as soon as it loads into your saves.

Plus, patched conics is good enough for KSP.

[pizzaoverhead]

if it simulates every body in the Kerbol system, the game would most likely crash as soon as it loads into your saves.

False. It's already been done.

[bmp0341]

"The current version is not stable,

it can corrupt or otherwise destroy saves,

it has been known to crash, and predictions are horrendously slow."

Uhh... Do you know how to read?

[Padishar]

Uhh... Do you know how to read?

Do you? That bit of text is from the section about the previous version.

I'm not saying that it is perfectly stable now (it isn't) but it is certainly better than "the game would most likely crash as soon as it loads into your saves".

[YNM]

Certainly it'd NOT run well, with Tantares and FASA (yeah !), on a 2GB RAM 1GB video card (GTX 210 if anyone's interested) and dual core ~3 Ghz i3 computer.

Patched conics itself... Already said about it once. Should search up my past posts...

EDIT: Here is it

- (1st) Patched because the trajectory is determined by the energy and momentum at that particular instance, assuming a simple limited two-body orbit.

- Conics because the trajectory results in conic sections, which is either a circle, ellipse, parabole or hyperbole.

- (2nd) Patched because the orbit is limited to SOIs. So you can't get a transfer to Mun using trajectory akin to GRAIL's trajectory, or you can escape mun in an elliptical orbit (eccentricity less than 1), for example. Also explains why a few near-Kerbin asteroid can get itself to orbit Kerbin, while lacking perturbations.

There's no link between patched conics method and GR at any rate, because trajectory in GR are always spiral.

Edited July 27, 2015 by YNM

[Mudkip909]

" I haven’t really been " Whoa, easy there with the special characters!

[Vanamonde]

1 hour ago, Mudkip909 said:

" I haven’t really been " Whoa, easy there with the special characters!

This is a very old thread. Some of the text was messed up during forum software changes.