Nuclear Fission Chain Reaction

[tekmaster]

I am in need of a tutorial or someone who could guide me through the process of creating a nuclear fission chain reaction in Houdini. I'm a complete n00b to Houdini, so any help would be muy bueno.

[tekmaster]

Im kinda in a hurry to get this done, so the faster someone could reply, le better.

[meshsmooth]

Well your first problem is what part of the nuclear reaction you want to create. The atoms splitting and smashing into the other atoms creating a chain reaction? Or are you just talking about a mushroom cloud?

if it is the mushroom cloud here is a link to a max tute that may help

http://www.the123d.com/tutorial/3dsmax3/explode01.shtml

This tute is for Houdini but may end up getting you into more trouble because then you will want to color a fog shader and I don

[tekmaster]

Well, I really just wanted to show neutrons smashing into larger isotopes and spliting them, also displaying the enegry released when the bombardment has finished. I want to show it at a large scale showing 100s of these chain reactions.

Maybe someone very skilled with Houdini can have a one on one with me and he can teach me on completing an effect like this.

[meshsmooth]

as far as finding tutorials on Houdini there are few sources but as for methods P-flow tutorials can give you ideas on the concepts that could be used

[anakin78z]

This hipfile may give you some ideas. It shows particles colliding with a survace and splitting every time they hit. You may be able to rework it to do what you want.

-z

recursive_splits.hip.gz

[Jason]

I'm afraid I can't be of much actual help (time constraints), but look into the Collision POP and the Split POP if you wan't a wild, particle based thing. If it has to be very controlled , you may have manually animate one of the these and use CopySOP (maybe stamping to vary time) to duplicate a systems of these.

I get the impression you will end up doing some quick particle based collisions/splits to get a feel for it and then animating a hero sequence to match to look and feel. Then back to adding in a tweaked particle approach again and finding a nice way to blend the two together - either a wipe or a fast zoom out or something.

damn, jens beat me to it;)

[talos72]

What you can do is set up a stream of particles moving in Y. using bounding box $BBY, setup groups for different parts of the mushroom cloud: ie. group any particle whose $BBY is greater than a normalized value or height. You can setup multiple groups corresponding to different heights of your upward particle stream.

Next you can setup various attractor systems using metaballs in SOPs to control swirling motion of your plume. Apply different attractor forces to different bounding box groups (having to do with BBY values). Using an Interact POP, for example, with high charge values to really spread the particles once they pass a certain height. Next, you can also use say Winp POP to push the particles in specific directions to get swirling.

It seems to boil down to combining various forces (Wind, attractors, interact) to get one final effect.

Just a thought

The rest boils down to endless tinkering of parameters to get the specific look.

[zeinin]

Allright, i just did this last night, i will post screenshots/hip tomorrow.

I assume you were talking about a nuclear fission chain reaction using the Bohr model of the atom. In that case, you would want to display a nucleus of 238-235 bosons (IE protons/neutrons) and about 92 elections in various orbits.

The way i did this was pretty simple: you take a sphere that has been subdivided into 238 points, spawn a particle on each. then, you declare one the "king", having the rest of the particles follow. An interact pop is chained with a few property pops to give the atom the proper distribution. essentiall, the "king" particle repels all the others, but all of the other particles want to follow it. the interact prevents the bulk of interparticle colisions. the next step is splitting off 92 electrons and putting them into a few groups to simulate energy orbits. they orbit around the king particle at various diameters through the orbit pop.

To split the atom, all you need to do is kill the king particle. the orbit pops will loose thier power, the electrons will scatter and the bosons will tear themselves apart. there are a few ways to do this: a simple keyed kill pop, a stream pop to do collosions without actually colliding, or an actual collision pop.

there are a few shortcomings: i was only able to simulate one atom in the popnet. even thougth the atom is created from a single source particle (it is all splits and the like), if you put more than one atom in the particle network, the whole thing breaks apart. the atoms do not know that they "own" certain groups of particles. i am sure i can figure something out with the lineage pop later. other problem is general touchyness: the whole system is dependant on certan particles having certain masses, and bringing the network to the semi-stable point was a lot of tweaking.

To simulate a chain reaction, you copy primitive spheres over the popnet (you might want to grab color attributes or someting for shading), and render out the result in a geometry rop. all you have to do then is copy the geometry sequence over a point-sorted template with an animation offset driven by $PT. the result is a psuedo chain reaction. I want to make a real chain reaction later, IE one that actually has neutrons killing the king particle on other atoms, causing the atom to split and kill more king particles, but it was just innefficient compared to this solution.

anyway, i will post it later

-Clive