500 Meter Bridge During Earthquake

[Mike Pauza]

I've been experimenting with MD using structural engineering type meshes. Below is a pic of a 500 meter bridge simulation during an earthquake. Because of the size of the mesh, and the prerequisate stiffness, I had to use a step size and resolution of about 10000 to acheive stability/numerical convergence, but I did get a "BIG & STIFF" simulation working in MD which surprised me!!! The simulation generated at about 1/3 fps on a fast PC. Email me if you want the 2MB DivX video ( sharonpauza@aol.com or mike@3dpiranha.com ).

-Mike Pauza

[Mike Pauza]

Does this suck or something? I thought it was really cool, but nobody's commenting.

[Randog]

Looks cool to me man. I'd love to see the video. Got a URL for a download?

[legomanww]

That looks pretty cool, I went WHOA! when I saw it. I'd like to see the video.

[Titus]

Actually I find this really cool. I've spent the last couple of years with Maya and there are some features I really miss like the soft bodies and particles.

[richpr]

It's great, just different ;)

[robpauza]

Hey Mike,

Just judging from the image and seeing some of your previous simulation videos, I bet it looks really sweet animated.

I'd be glad to host the video for you. Just drop me an e-mail.

-Rob

[starbase1]

I suspect you have to see it moving to get the full effect - I'd certainly like to see a vid clip!

Nick

[Lynx3d]

Darn...now i wanna play bridge builder/pontifex again, and i thought i got over the addiction :D

[Mike Pauza]

Hey thanks everybody. My brother is going to have this video and some other MD examples on-line tonight if all goes well.

-Mike

[legomanww]

Cool. Can't wait to see it.

[Mike Pauza]

Here are the links:




http://mywebpages.comcast.net/robpauza/Cup_Soft.avi (0.46 MB)

http://mywebpages.comcast.net/robpauza/Cup_Hard.avi (0.43 MB)

http://mywebpages.comcast.net/robpauza/BridgeEarthquake2.avi (4.5 MB)

http://mywebpages.comcast.net/robpauza/CubeArray_Twist.avi (1.8 MB)

http://mywebpages.comcast.net/robpauza/CubeArray_Stretch.avi (1.3 MB)

http://mywebpages.comcast.net/robpauza/Truss_Shake.avi (0.26 MB)

http://mywebpages.comcast.net/robpauza/Truss_Snake.avi (0.89 MB)




I plan to keep creating these little sims so please comment. :)

-Mike Pauza

[legomanww]

The bridge during the earthquake is neat. What kind of scale earthquake are we talking here? Seems like a big earthquake.

The trusses are neat, although either they are a lot bigger scale than they seem, or they are too soft. But still looks cool.

I like the cube array twist.

I don't quite understand the cups, why do they slide at the end?

Looks good though.

[Lightwolf]

Wow Mike!
:cool: f***ing wicked sh*t !!! ;)
Very cool stuff. What were your calculation times on what kind of a machine? It looks like MD can be really capable if used properly (If the ui was just a tad better, and the whole loading / saving of attributes less flakey, I'd even use it on tight deadlines...).
Cheers,
Mike

[Mike Pauza]

legomanww:

Hey, I'm just messing around with the earthquake motion and the bridge stiffness...just wanted to create some interesting motion and to see if MD could handle it. :) When I mean accurate, I mean it deforms like it's supposed to.

The cups slide at the end because MD isn't very good at friction (or collision in general) and I didn't take the time to tweak it.




Lightwolf:

Woah, thanks!

On a 2.5GHZ P4 the cup sims generated at about 5 fps, the cube arrays at about 5 fps, the trusses at about 200 fps, and the bridge at about 1/3 fps...but it had 24,000 springs in it. No collision detection except for for the cups.

I'm finding that MD is very capable at some things, and totally inadequate for others...some of the internal code seems to be good and some seems to be crap. Like most everybody else I hate the interface, but I'm getting it to kind of work by using "structural meshes" that are built for simulation in mind.



-Mike

[Lightwolf]

Hi Mike,
I think collision detection is one of the weakest part in MD. Even though I got some decent results with plenty of tweaking, a save and load of a scene combined with a simulation re-run could lead to totally different results :( (I guess this has to do with the way MD saves...)
I whish I had more time and patience for it (...well, as well as a faster machine :D ).
You really ought to write a tut on your techniques...
Cheers,
Mike - still impressed :cool:

[Lynx3d]

whoa, that earthquake anim is impressive IMHO!
Hm not too long ago such simulations where the best computers can do, and now you just do it with a "consumer level" 3D software... :)

[Mike Pauza]

Thanks for the encouragement guys. Hopefully I can put a tutorial together once I flesh everthing out.

Tonight I'm going to try and build a self flying glider.

-Mike Pauza

[Mike Pauza]

I got an MD Glider working. Yeah baby yeah!

It's not very pretty, and it took me way too long, but I'm excited. Hopefully I'll have video links up tonight or tomorrow. I can't really "fly it" interactively because there's no way for me to control flaps or give it any power (am I wrong about that?). Anyway, this sort of sim shows the power of MD's parallel air resistence and sets the stage for more complex dynamic machines...like ILM's pod racing simulations for example.

I just started a new thread "Md Flying Machine" if your interested in this.

-Mike Pauza

[stib]

cracking stuff, mike. You've inspired me to start playing with MD now.

I'm interested in the truss-snake sim. How did you set it up?
You know you could do a great tutorial based on your MD work..

[edit]dang this was supposed to go on the thread with the bridge. oh well...

[merlinux]

i like you're simulations but u certainly now other ways to do that like "inertia" for the snake sample...

[Mike Pauza]

stib :

I'll get a big tutorial together sooner or later...but basically what I did was build a truss/tower sort of structure (go find one to look at if you don't know how they are put together) out of line polygons...and created a surface out of it, made a "base" surface out of some end polygons...brought the model into layout, keyframed the object, and in MD I made the base surface to be "fixed" so that it would move with the keyframing...the rest of the truss then has to deform accordingly (once you get the parameters how you want them, and the calculation resolution and step number high enough to solve it correct that is).




merlinux:

Yeah I know there are other (faster) ways of doing a snake-like animation for simple cases, but I was just fooling around to see what MD is capable of.




-Mike Pauza

[Consty]

*wave* hello, new here. Just checking through the forum looking at the pretty pictures :).

I have a possibly silly newbie question: What's MD, and is it part of lightwave? I still haven't found where all the bits are in the program, there's a lot to play with there!
Consty

[Mike Pauza]

MD is Motion Designer, which is part of LightWave.

[lasco]

good work.

question : is this "just" an artistic work or did
you expect to do a "real" earthquake simulation,
I mean in a scientific sense ?
If it's the case then I'd just say it's maybe not enough understandable
(i'm talking about the movie of the bridge) : another point of view (not inreplacement but added to this one) for instance a side view would I think show
better what happens, as the perspective deforms too much the object itself,
I mean as the scale of the object is deformed by the perspective the scale
of the deformations DUE tu the earthquake is not really understandable…
(hey, you write it's a 500 meter bridge but there it could be 100 meter or 2000 meters, who would guess ?)

And also about "scientific" things (sorry my english's really bad) could it
be possible to change for example the color of the mesh (WF) at "special" points,
where there are the more tensions for example…

See what I mean : guess it might be even better it was more "technical"…

[Mike Pauza]

lasco:

It's not really meant to be an accurate earthquake simulation...just a little something I wanted to try. :)

-Mike

[lasco]

all right,
nothing to add then it's perfect !

(though I'd like to see a side view of this animation ;)

[Mike Pauza]

Here's a 3/4 view:

http://www.3dpiranha.com/BridgeEarthquake.avi

[Bionic Antboy]

That's some pretty cool stuff! I've been wanting to dig into MD some more, though work doesn't really call for it too often.

For the bridge quake, have you tried "twisting" the ends independantly on the Y scale?

I remember some movie (don't recall which) that had a nice feel to it, and I believe they must have set up the ends to shift up and down independantly. Had a good look to it.

[Mike Pauza]

Thanks Bionic!

MD currently has no hierarchy control, so you can't really twist the two ends separatly...I did key a jerking rotation from one end though to create that wave that propagates through it.

-Mike

[lasco]

very interesting 3/4 view !

[Bionic Antboy]

Originally posted by Mike Pauza
Thanks Bionic!

MD currently has no hierarchy control, so you can't really twist the two ends separatly...I did key a jerking rotation from one end though to create that wave that propagates through it.

-Mike

What about a bone rig that just twists on the one axis, to give that extra dimension of motion? I know that's not really an MD solution, but it might give it the "extra touch".

It's been so long since I played with MD, I can't remember if such bone-controlled twisting motions would translate well into the MD calculations of the upper parts of the bridge structure or not.

Of course, having a hierarchical approach that allows separate control points would be a better solution :)

[Mike Pauza]

Can you attach bones to MD? That would be cool in this case...I'll give it a try.

Thanks.

[Bionic Antboy]

Sorry if I mis-posted...

I don't think there's any way to attach bones to MD directly, but you can manipulate the object using bones, and that distortion will be taken into account by MD after the fact. Not the most "realistic" solution technically, but...

if you do a pretty straight forward bone rig, with just a string of bones rotating on their bank, then apply your MD, you should get some usable bounciness, kind of like what one sees in footage from real storms/quakes/whatever. :)

Since the bridge is a little more complex than a ribbon, it may take more work, but it's just an idea.