I still wasn't quite sure what SPH is, so I googled it. From what I've read, it seems like it'd be extremely helpful.
For example, Saturn's rings are lagging way too much in my computer, to the point it almost crashes the game. So, with SPH, the rings wouldn't be calculated as individual particles (their orbits and behavior wouldn't be individually calculated) and the rings would be like just one thing... right? That sounds amazing. I'm totally a SPH enthusiast now
Sorry to be a downer, but that is about as wrong as it can be really :-/
I tried a few different ways of describing this in a brief message, in a way which was both correct and informative and readable, and that was rather hard, so the following is my best shot at it, which is not perfectly accurate, but should bring the point across.
SPH is a method to model the behavior of a material by describing the material as a collection of its parts, where each part is a "particle". It is commonly fluid materials, but could just as well be solid materials.
Assume you have a big blob of water. There is no common physical equation for how such an object behaves when you push it a little here or heat it a bit there. You only have continuum equations which are differential equations which basically state things such as
if you for a point in water with a certain density have a pressure gradient (difference in pressure from side to side) like this, the water will accelerate like that, and
if you for that water point has a velocity divergence (how much the water expands or pushes together) like this, then the pressure will change over time like that and the internal energy will change like that, and
if the difference of the velocity field between the point and its surroundings is this and the viscosity is that then the acceleration will be this and the internal energy will change by that much.
When the internal energy changes by this and the heat capacity is that then the temperature changes like this and the pressure like that...
and so on and so on
The problem is that those values are not really known in a computable form without the blob being discretized. This means that it much be split into simpler building blocks which can be handled, so you can calculate the state of things in each block as well as the derivatives describing how much that state is changing between neighboring blocks.
One way is finite element or finite volume where you make a mesh and calculate this in different ways. Another is finite difference where you overlay a grid on the blob and describe the state at grid points. Another is SPH, smoothed particle hydrodynamics, where you "cut out" little blobs of this larger blob called particles.
So instead of describing a huge complex blob you consider its parts (and this is the part which is not technically right, but brings across the idea) and when one little part needs to find out how much it should accelerate due to pressure difference it will look to its neighbor blobs and see what their position and their pressure is.
You may have a blob which sees to its right a high pressure blob and to its left a low pressure blob. It will then accelerate left away from the higher pressure.
Not sure if this simplistic explanation is clear at all, but feel free to ask again.
Anyway, the main point here is that the rings of Saturn will with SPH still be described using particles, just different ones, and the calculations involved will be much more complex than what we have for pure N-body. The second point is that those rings are really not one material where the parts affect each other by forces other than gravity, so SPH would be rather pointless.
SPH is actually used in modelling accretion disks, so the idea is not totally off, but the point there is that planet formation and momentum transferring in an accretion disk requires viscosity to be modeled, because it is a driving factor there.
The point of sph in US² is that the current fragmenting rigid body model for planet collisions is really not describing a lot of what goes on. It does not capture how the bodies touch and compression waves move through the bodies heating things up (the current shock waves is not that). Not does it describe how the bodies flatten and fragment in highly non spherical forms etc etc. It also doesn't properly describe the stretching and fragmentation of a body in a steep gravity well and it doesn't describe tidal deformation and how that leads to heating and less relative rotation etc etc.
That was a longer bla bla message than intended. While probably rather unclear, I hope it gives a little usable information.
Topic: Supernovae (Read 6759 times)
