Hey guys,
I'm doing a bit of research with my university (UNSW in Sydney, Australia) concerning game engine physics and structural analysis. I'm studying civil engineering, so I'm no programmer or mathematician but I have been trying to get a fair understanding of all this stuff over the past couple of months.
What I'm basically investigating is the ability of engines to perform real-time structural analysis. So far I've looked at the FEA work from Pixelux, Volitions Geomod 2.0 stuff and a few others. The current path I'm looking into involves taking this piece of software (Frame3D from Dr. Software, a standalone real-time structural analysis engine) and potentially be able to port its analysis and visualization methods into Bullet, with the Bullet engine handling all the collisions constraints (because buildings are of course made up of a lot more than just structural elements) and non-structural joins. And then of course further add-ons could then be included to form the more aesthetic elements around the hidden building structure.
In your opinion does this look feasible? Would Bullet itself be able to perform these calculations within its engine or would such an add-on be required?
If you have any other useful information on the topic in general, it'd be great to hear what you know.
Cheers guys,
Jaryd
Structual Analysis with Bullet
Re: Structual Analysis with Bullet
It sounds like a reasonable idea. The solver of a physics engine iterates until it arrives at an impulse at each contact/constraint/joint that satisfies the system.
Simple example... using a simple physics engine, if you look at the contacts at the bottom of a single stack of blocks then you can see the force at the bottom does correspond to the weight of all the things on top.
Some things you should watch out for... There are tricks a physics engine can do to perform better. This includes detecting certain structures and *not* propagating impulses in order to achieve faster convergence. When the motivation is a stable stack of bricks (and not an accurate simulation), its possible that the engine might cut corners. Also note that engines will regularly turn things off and on, so this might affect the numbers the simulation generates. Of the various optimizations, I believe the warm-starting technique should not cause this problem for you. Anyways, it might be wise to get a sense of how the underlying simulation works so you dont get too frustrated during your research.
I think many people on this forum would be interested in learning how suitable game physics engines are for static analysis. After you do your research, please follow-up with a post on what you find out. As a civil engineer, you are the most qualified to determine how suitable these game engines are for real-world problems.
good luck
stan melax
Simple example... using a simple physics engine, if you look at the contacts at the bottom of a single stack of blocks then you can see the force at the bottom does correspond to the weight of all the things on top.
Some things you should watch out for... There are tricks a physics engine can do to perform better. This includes detecting certain structures and *not* propagating impulses in order to achieve faster convergence. When the motivation is a stable stack of bricks (and not an accurate simulation), its possible that the engine might cut corners. Also note that engines will regularly turn things off and on, so this might affect the numbers the simulation generates. Of the various optimizations, I believe the warm-starting technique should not cause this problem for you. Anyways, it might be wise to get a sense of how the underlying simulation works so you dont get too frustrated during your research.
I think many people on this forum would be interested in learning how suitable game physics engines are for static analysis. After you do your research, please follow-up with a post on what you find out. As a civil engineer, you are the most qualified to determine how suitable these game engines are for real-world problems.
good luck
stan melax