Computational Acceleration of Topology Optimization Using Parallel Computing and Machine Learning Methods – Analysis of Research Trends
•Advanced manufacturing is being integrated with lightweighting methods•Topology optimization has become a leading method in this area•Parallel computing is used for acceleration of topology optimization methods•Machine Learning methods has been actively applied for topology optimization•GPU and Mac...
Saved in:
| Published in: | Journal of industrial information integration Vol. 28; p. 100352 |
|---|---|
| Main Authors: | , , , , , , , |
| Format: | Journal Article |
| Language: | English |
| Published: |
Elsevier Inc
01-07-2022
|
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | •Advanced manufacturing is being integrated with lightweighting methods•Topology optimization has become a leading method in this area•Parallel computing is used for acceleration of topology optimization methods•Machine Learning methods has been actively applied for topology optimization•GPU and Machine Learning acceleration techniques are trending in this area
Development of advanced structures using modern manufacturing methods has become attractive since they allow to improve system efficiency and performance, fuel consumption reduction, lightweighting to decrease weight and durability of structures, and many more. Designing tools such as topology optimization (TO) has contributed to such developments and facilitated in adapting new manufacturing methods such as 3D printing and computer numerical control machining in many areas of engineering and industry.
TO requires computational resources, which can be significantly complex and time consuming when complicated designs and multiphysics problems are considered. To overcome these difficulties, computational acceleration techniques have been applied together with high performance computing. In the current work, various up-to-date research studies in computational acceleration of TO methods are analysed, classified and research trends are evaluated. Thus, the results of the work clearly shows that earlier works relied on central processing unit (CPU)-based computational acceleration techniques, while latest research studies mostly consider graphics processing unit (GPU) and machine learning (ML)-based approaches. The latter got significant attention within last few years and becoming one of the research areas in computational TO. From the reviewed works, it can be concluded that in all of the acceleration techniques, solid mechanics problems were mostly studied, while a few number of research studies are dedicated to heat transfer, fluid flow and electro thermomechanical applications. |
|---|---|
| ISSN: | 2452-414X 2452-414X |
| DOI: | 10.1016/j.jii.2022.100352 |