building a bridge game
Ondrej Certik
I just read:
http://groups.google.com/group/sage-devel/browse_thread/thread/5d8fad12107b4ba5
which was written by a participant of some NSF conference. We were
discussing with Robert, that it'd be cool to have some FEM game, and
building a bridge was one of the ideas. As pointed in the post above,
someone has already implemented it:
http://www.bridgebuilder-game.com/
you can see screenshots here from the year 2000 version:
http://www.bridgebuilder-game.com/images/screen2.jpg
http://www.bridgebuilder-game.com/images/screen3.jpg
http://www.bridgebuilder-game.com/images/screen4.jpg
http://www.bridgebuilder-game.com/images/screen5.jpg
and year 2006 version:
http://www.chroniclogic.com/index.htm?pontifex2screens.htm
but they are not using FEM. This could be a nice demo for hp-FEM. In
fact, I would really like to take a picture of some real world bridge,
then quickly (easily) model it in hermes and then see the stresses and
what happens if a train (track) goes over it.
But I don't have time to write it. :)
Ondrej
Pavel Solin
model. No relation to reality. We have a 3D mesh for a realistic
highway bridge. It is very fine, made for linear elements.
My goal is to show that hp-FEM will solve it much faster,
but we have to wait until we have curved elements in 3D.
Pavel
Pavel Solin
University of Nevada, Reno
http://hpfem.math.unr.edu/people/pavel
Visit Hermes project at http://hpfem.org
Ondrej Certik
> Unless I am missing something, this is just a naive 1D spring
> model. No relation to reality. We have a 3D mesh for a realistic
Yes, it is artificial.
> highway bridge. It is very fine, made for linear elements.
> My goal is to show that hp-FEM will solve it much faster,
> but we have to wait until we have curved elements in 3D.
Yes -- that will be just one particular bridge. I can imagine a game
where the user would actually construct the bridge.
Ondrej
Pavel Solin
You would have to go for steel bridge, since otherwise we would
have problems determining material parameters. The problem would
have to be 2D since no gamer would wait for a 3D computation to finish.
Last -- plasticity modeling is skipped completely in these games.
This is *nontrivial*. Right before the bridge collapses, you have to
consider plasticity. There is no way from the linear elastic regime
to a breakdown without that. Good project for a few months if you are
bored :)
Pavel
Ondrej
David Andrs
need that anyway :)
--
David
>
> Ondrej
>
> >
Ondrej Certik
>
>
> On Sun, May 3, 2009 at 7:35 PM, Ondrej Certik <ond...@certik.cz> wrote:
>>
>> On Sun, May 3, 2009 at 6:06 PM, Pavel Solin <so...@unr.edu> wrote:
>> > Unless I am missing something, this is just a naive 1D spring
>> > model. No relation to reality. We have a 3D mesh for a realistic
>>
>> Yes, it is artificial.
>>
>> > highway bridge. It is very fine, made for linear elements.
>> > My goal is to show that hp-FEM will solve it much faster,
>> > but we have to wait until we have curved elements in 3D.
>>
>> Yes -- that will be just one particular bridge. I can imagine a game
>> where the user would actually construct the bridge.
>
> You would have to go for steel bridge, since otherwise we would
> have problems determining material parameters. The problem would
> have to be 2D since no gamer would wait for a 3D computation to finish.
Well, the 2D model can give him some quick results, but imho our 3D
code should be able to calculate reasonably fast (e.g. couple
minutes).
> Last -- plasticity modeling is skipped completely in these games.
> This is *nontrivial*. Right before the bridge collapses, you have to
> consider plasticity. There is no way from the linear elastic regime
Isn't there some regulation, that the bridge should never get out of
the linear elastic regime? So I guess if the stresses reach some
threshold (in the linear regime), the game is over (also in real
life).
> to a breakdown without that. Good project for a few months if you are
> bored :)
Frankly speaking, I am not at all bored these days. :)
On Sun, May 3, 2009 at 7:48 PM, David Andrs <and...@gmail.com> wrote:
> Then, you should start the work on a suitable mesh generator. We will
> need that anyway :)
Yes, I think when we have the mesh generator, and robust 2D and 3D
solver, I don't mind specifying the bridge using a python script. But
it should be able to plot the bridge (and allow me to specify almost
any bridge) and solve it. So I believe we should be able to do that,
eventually.
Ondrej
Pavel Solin
Life is not so easy :)
Virtually in all re-entrant corners, stresses in
the linear elastic model go to infinity. In practice,
there always are plastic effects at very small spatial
scales there. That's why Jarda laughed when I asked
him what they do with infinite stresses: While we
have them all the time and call them "singularities",
the engineers have more advanced nonlinear models
involving plasticity effects, and they do not have
infinite stresses.
Frankly speaking, I am not at all bored these days. :)
> to a breakdown without that. Good project for a few months if you are
> bored :)
You must be kidding :))
Yes, I think when we have the mesh generator, and robust 2D and 3D
On Sun, May 3, 2009 at 7:48 PM, David Andrs <and...@gmail.com> wrote:
> Then, you should start the work on a suitable mesh generator. We will
> need that anyway :)
solver, I don't mind specifying the bridge using a python script. But
it should be able to plot the bridge (and allow me to specify almost
any bridge) and solve it. So I believe we should be able to do that,
eventually.
Ondrej
Ondrej Certik
>
>
> On Sun, May 3, 2009 at 8:02 PM, Ondrej Certik <ond...@certik.cz> wrote:
>> Isn't there some regulation, that the bridge should never get out of
>> the linear elastic regime? So I guess if the stresses reach some
>> threshold (in the linear regime), the game is over (also in real
>> life).
>
> Life is not so easy :)
> Virtually in all re-entrant corners, stresses in
> the linear elastic model go to infinity. In practice,
> there always are plastic effects at very small spatial
> scales there. That's why Jarda laughed when I asked
> him what they do with infinite stresses: While we
> have them all the time and call them "singularities",
> the engineers have more advanced nonlinear models
> involving plasticity effects, and they do not have
> infinite stresses.
Ah, interesting, I didn't know that. So we have to use some
(nonlinear) model which takes the plasticity effects into account
then.
Ondrej