FE, Dynamics, Fluids and A.I.

[Trying2Learn]

TL;DR Summary

Integrating the processes of mechanics using network softare

I am aware that a lot of software exists to conduct such multi-phase macroscopic analyses.

But has anyone ever used the tools of interprocess communication to do this?

Consider, for example, these three:

• A Finite element code (For deformation) on a parallel machine
• A Dynamcis code (for dynamics) on a vector machine
• An AI code (for learning) on a PC

Consider a case where all such codes communicate through, say, sockets.

Consider that this is done in such a way that each code adheres to a prescribed data format so that they can more readily communicate?

Can anyone share any references?

The motivation for this question is as follows...

There exist many multi-phase mechanics codes (dynamics, FE, Fluids, Heat, etc.), however, the codes are massive. Perhaps, if each code was more isolated, then it would be easier:

• to learn how to do multi-phase analyses (again, at the macroscopic level).
• to share efforts.
• to integrate the codes with, say, learning tools for smart machines.

I am just thinking out of the box and would like to know if anyone has done this (I know it will not be as fast as one stand alone code).

 

[FEAnalyst]

I think that you will find this issue of Benchmark magazine (by NAFEMS) interesting: „Artificial Intelligence & Machine Learning” July 2018. It features several articles with examples of industrial cases when AI was applied to simulation processes.