GUVI
Back

Scope of deep learning in mechanical engineering.

Created 3 years ago
18 Views
0 Comments
Guvigeek
@Guvigeek
Guvigeek
@GuvigeekProfile is locked. Login

So I have been hinting, right. It's very important to combine knowledge of some with deep learning. So one domain we talked about was this MBA like business domain and so on another domain was healthcare. We talked about education. Mechanical Engineering is also interesting. For mechanical Engineering one of the things that is often done is to simulate certain things. For example, what will be the amount of G force on a vehicle as it is turning at this angle or maybe what is the stress on this particular component as some water is flowing through it, but typically the lot of simulations that are done and the simulations are done with things like finite element method, where in detail, you have to model the different equations and then simulate in small units of time and space. And these simulations are very, very expensive in terms of compute power, people have been successfully using deep learning for speeding up such simulations. In fact, there is some good work happening at IIT Madras, which tries to speed up simulations of various kinds using deep learning. So, that is one good example. The other example that I'm familiar with is a work done with CAD tools, CAD is computer aided design tools, where you can use deep learning to help you in the design process. So, it can generate different combinations, and then you can start choosing between them, instead of you designing the entire thing, there can be some generative part where different options are created, and you start as a designer choosing amongst options rather than designing. So there might be many others that are not familiar with, but I think, recognizing that you can use deep learning for speeding up simulation or for improving the speed of design. All these are good templates, which can have different instances.

Comments
Please login to comment.
 
Powered by habitate.io Habitate