Conference hydra::amiga_v1

5180.0. "Lightening Simulation?" by STAR::GUINEAU (but what was the question?) Wed Nov 06 1991 14:07

Below is a Usenet thread I started on Lightnening Simulation. I think with the
right algorithm, the Amiga could do a neat job...

Any ideas on how to approace this???

john



Article 1901 of alt.fractals:
Path: nntpd.lkg.dec.com!datum.nyo.dec.com!pa.dec.com!decwrl!uunet!convex!mips.mitek.com!utacfd.uta.edu!letni!rwsys!trsvax!kent
From: [email protected] (Kent Bowling)
Newsgroups: alt.fractals
Subject: Re: Lightening Simulation?
Message-ID: <[email protected]>
Date: 4 Nov 91 21:31:53 GMT
References: <[email protected]>
Organization: Tandy Electronics Research and Development, Fort Worth, Texas, USA
Lines: 13

It's funny that you mention this, because during a thunderstorm the other day,
I started thinking of ways to simulate lightening.  I am going to try out an
expirement an see how it turns out.  I was thinking of basically setting up
a 3 dimensional plasma cloud (much like the 2D version in fractint except in a
3D box).  Since electricity is "lazy" and follows the route of least
resistance, one could start at a point and constantly check all of the points
around it to see the one with the smallest value (that is not above the current
point).  As far as the lightening branching out, you could have it set up so 
that when it finds to steps of the same value, you could give it a certain
probability of it splitting (say 50%).  This should create some kind of results
that might look like lightening.  The hardest part (which isn't TOO hard), would
be displaying the 3D lines.
						Kent



Article 1905 of alt.fractals:
Path: nntpd.lkg.dec.com!datum.nyo.dec.com!pa.dec.com!decwrl!uunet!van-bc!rsoft!mindlink!a42
From: [email protected] (Marc North)
Newsgroups: alt.fractals
Subject: Re: Lightening Simulation?
Message-ID: <[email protected]>
Date: 5 Nov 91 07:30:52 GMT
Organization: MIND LINK! - British Columbia, Canada
Lines: 16

> [email protected] writes:
> 
> point).  As far as the lightening branching out, you could have it set up so
> that when it finds to steps of the same value, you could give it a certain
> probability of it splitting (say 50%).  This should create some kind of
> results that might look like lightening.


I'm very curious about this. Perhaps you could elaborate for us how you will
control the length of the forks? I would suggest that a realistic simulation
for lightening must have an accurate representation of both the major and
obvious forks, as well as the minor and more delicate forks.


--
[email protected] -- ULYSSES Systems Corp. (USC)



Article 1906 of alt.fractals:
Path: nntpd.lkg.dec.com!star.enet.dec.com!guineau
From: [email protected]
Newsgroups: alt.fractals
Subject: Re: Lightening Simulation?
Message-ID: <[email protected]>
Date: 5 Nov 91 14:04:59 GMT
References: <[email protected]>
Sender: [email protected]
Reply-To: [email protected] ()
Organization: Digital Equipment Corporation
Lines: 38


In article <[email protected]>, [email protected] (Marc North) writes:
|>> [email protected] writes:
|>> 
|>> point).  As far as the lightening branching out, you could have it set up so
|>> that when it finds to steps of the same value, you could give it a certain
|>> probability of it splitting (say 50%).  This should create some kind of
|>> results that might look like lightening.
|>
|>
|>I'm very curious about this. Perhaps you could elaborate for us how you will
|>control the length of the forks? I would suggest that a realistic simulation
|>for lightening must have an accurate representation of both the major and
|>obvious forks, as well as the minor and more delicate forks.
|>

I like this approach as well. I think you would also require some method
of attenuating each fork as it looses energy to the surroundings.

The energy at a given point in a fork could be represented by the "diameter"
of the fork. This way as a fork looses energy, it gets thinner, just like
real lightening.

Also, would the energy lost in a strike be returned in part to the plasma cloud?

> point).  As far as the lightening branching out, you could have it set up so 
> that when it finds to steps of the same value, you could give it a certain
> probability of it splitting (say 50%).  This should create some kind of results

As for dividing, I'd give it a small range of "equi-potential" steps to split 
on, not just ones of equal potential. This would allow more control over how
much branching a strike does.


--                			
W. John Guineau                         grep meaning life | more
VMS Development
Digital Equipment Corporation		[email protected]



Article 1907 of alt.fractals:
Path: nntpd.lkg.dec.com!tkou02.enet.dec.com!jrdzzz.jrd.dec.com!pa.dec.com!decwrl!usenet.coe.montana.edu!caen!zaphod.mps.ohio-state.edu!qt.cs.utexas.edu!yale.edu!cs.yale.edu!musgrave-forest
From: [email protected] (F. Ken Musgrave)
Newsgroups: alt.fractals
Subject: Re: Lightening Simulation?
Message-ID: <[email protected]>
Date: 5 Nov 91 13:21:35 GMT
References: <[email protected]>
Sender: [email protected] (Usenet News)
Organization: Yale University Computer Science Dept., New Haven, CT 06520-2158
Lines: 23
Nntp-Posting-Host: systemsy-gw.cs.yale.edu


  As someone with some experience in developing fractal models of natural 
phenomena, I can tell you this: lightening is hard to do (well).

  It is an instance of dielectric breakdown in 3 dimensions, and thus quite
context-sensitive.  This makes for expensive algorithms, if you want to 
model it properly.

  A simple model might be to solve a Laplacian on a 3-D random field.
Elegant perhaps, but computationally infeasible!

  But then, if you can come up with a cheap hack that looks good, that'd
be in some ways better (faster, hopefully, at least!)

  I'd be interested to know if you make significant progress on this.

								Ken

-- 
"But what do we do about moose and squirrel?"	 -Boris Badanov

F. Kenton ("Ken") Musgrave	[email protected]	(203) 432-4016
Yale U Depts of Math and CS	Box 2155 Yale Station	New Haven, CT 06520



Article 1912 of alt.fractals:
Path: nntpd.lkg.dec.com!datum.nyo.dec.com!pa.dec.com!decwrl!uunet!stanford.edu!msi.umn.edu!cs.umn.edu!cs.umn.edu!sandvig
From: [email protected] (Cary Sandvig)
Newsgroups: alt.fractals
Subject: Re: Lightening Simulation?
Message-ID: <[email protected]>
Date: 5 Nov 91 23:09:59 GMT
References: <[email protected]> <[email protected]>
Sender: [email protected] (News administrator)
Organization: The Geometry Center
Lines: 30
In-Reply-To: [email protected]'s message of 4 Nov 91 21:31:53 GMT
Nntp-Posting-Host: poincare.geom.umn.edu

In article <[email protected]> [email protected] (Kent Bowling) writes:
|It's funny that you mention this, because during a thunderstorm the other day,
|I started thinking of ways to simulate lightening.  I am going to try out an
|expirement an see how it turns out.  I was thinking of basically setting up
|a 3 dimensional plasma cloud (much like the 2D version in fractint except in a
|3D box).  Since electricity is "lazy" and follows the route of least
|resistance, one could start at a point and constantly check all of the points
|around it to see the one with the smallest value(that is not above the current
|point).  As far as the lightening branching out, you could have it set up so 
|that when it finds to steps of the same value, you could give it a certain
|probability of it splitting (say 50%). This should create some kind of results
|that might look like lightening. The hardest part (which isn't TOO hard),would
|be displaying the 3D lines.

Ok, my $0.02 worth.  I think this would be a good start but needs further
attenuation.  For example: as the arc is formed, the "air" is is forming in
as well as the "air" immeadiately adjacent to it would be become more conduct-
ive, the greater the flux through it.  This, however, would happen more slowly
than the propogation of the arc; but, would allow for other branching effects
as well as automatically fix the radius problem.  A greater flux can jump
across a greater resistance, so a greater flux will 'involve' a greater amount
of the "air" in that area (so it would be thicker).  Unfortunately this means
that you also have to check along the entire surface of the arc for new
branches every itteration.  This would get expensive, but I susspect it would
provide good looking results.  I also think a fairly sparse "plasma cloud"
will work best in this case.
--

				-Frang
		       (the beast a man can be)