Abstract
The paper considers the use of neural networks to predict the failure load of cold-formed steel compression members. The objective is to provide a fast method of predicting the failure load, so that the method can be used in other design algorithms, such as optimisation routines. Three types of symmetric sections are considered, and the results of neural network predictions compared with results from BS5950 Part 5. The results are in good agreement with the results from design codes. Moreover, a trained neural network gives the results significantly more quickly than a computer implementation of the code. (C) 2002 Civil-Comp Ltd and Elsevier Science Ltd. All rights reserved.
| Original language | English |
|---|---|
| Article number | PII S0965-9978(02)00051-0 |
| Pages (from-to) | 713-719 |
| Number of pages | 7 |
| Journal | Advances in Engineering Software |
| Volume | 33 |
| Issue number | 7-10 |
| DOIs | |
| Publication status | Published - 2002 |
| Event | 2nd International Conference on Engineering Computational Technology/5th International Conference on Computational Structures Technology - LEUVEN, Belgium Duration: 6 Sept 2000 → 8 Sept 2000 http://www.civil-comp.com/conf/cst2000.htm |
Keywords
- steel structures
- neural networks
- structural design
- cold-formed steel