Neural network modeling of the load-carrying capacity of eccentrically-loaded single-angle struts

✍ Scribed by Sherief S.S. Sakla

• Publisher: Elsevier Science
• Year: 2004
• Tongue: English
• Weight: 273 KB
• Volume: 60
• Category: Article
• ISSN: 0143-974X
• DOI: 10.1016/j.jcsr.2003.09.006

✦ Synopsis

In the majority of structural applications, single angles are usually loaded in such a manner that the applied load is eccentric. Eccentrically loaded single-angle struts are among the most difficult structural members to analyse and design. In many cases, the compressive resistances calculated using existing conventional design models can differ greatly from experimental results available in the literature. In this study, the potential of using artificial neural networks (ANNs) to predict the load-carrying capacity of pin-ended single-angle struts is investigated. Results of reported experimental studies on eccentrically loaded steel single-angle struts were used to train and validate the proposed ANN model. The performance of the proposed ANN model was subsequently compared to that of the AISC specification currently in use. It is shown that neural networks outperform the current AISC specification and provide an efficient alternative method in predicting the load-carrying capacity of eccentrically loaded single-angle struts. Using the developed ANN model, a parametric study was then conducted on angles connected by one leg to investigate the effect of changing the location of the applied load.