3 ms·
Does this approach get you anywhere the decades of dedicated CFD solvers haven't yet? I guess that would surprise me but I'm not up to date in CFD.
by ska 3y ago
Does this approach get you anywhere the decades of dedicated CFD solvers haven't yet? I guess that would surprise me but I'm not up to date in CFD.
- eschluntz 3y agoThe real question is whether the CFD solver is differentiable! Optimizing the bridge only works because my method for solving the forces in each beam (least squares) happens to be differentiable.
- londons_explore 3y agoThere are a bunch of papers on training a neural network to replicate the output of a cfd solver. Then the network is differentiable.
- eschluntz 3y agoooh, using a NN as a solver / simulator would make pretty much anything differentiable :)
- dist-epoch 3y agoCFD is precise, but slow. There are neural networks which learn to produce an approximation, but much faster. Useful for games, maybe exploratory analysis before deploying using slow classic CFD.
- londons_explore 3y agoI think there could be a combined approach where you use the fast neural net and the slow CFD as you state, but then every result you get from the CFD you add to the training data of the neural net and do a few more steps of training. That way, the neural net gets more accurate specifically in the area of your design. You could imagine a developers workstation where a GUI tool is letting them design stuff, and the neural net is giving them instant answers (and/or showing gradients to show how to improve stuff), and slow CFD runs are running in the background so whenever they go to lunch and another cfd run completes the results all get more accurate and detailled. Kinda similar to the way many graphics packages have 'draft' quality 3d rendering, but then when idle for a bit will improve the render quality.