Niğde Ömer Halisdemir Üniversitesi Mühendislik Bilimleri Dergisi

Yazarlar: Çiğdem AVCI-KARATAŞ

Konular:İnşaat Mühendisliği

DOI:10.28948/ngumuh.759297

Anahtar Kelimeler:Concrete-filled steel tube composite stub columns (CFSTCSCs),Ultimate axial load capacity,Relevance vector machine (RVM),Nonlinear regression algorithm

Özet: In this paper, the applicability of relevance vector machine (RVM) has been explored to predict the ultimate axial load capacity of concrete-filled steel tube composite stub columns (CFSTCSCs) with circular sections under axial compression loadings. As an extension of support vector machine, RVM employs Bayesian inference to achieve parsimonious solutions for regression and classification. By using MATLAB software and 150 comprehensive experimental data presented in the previous studies, a model to predict the ultimate load of circular CFSTCSCs was developed by properly training the data. Utmost care has been taken in grouping the data for training and validation. About 80% dataset for training and 20% dataset for validation have been used, respectively. The results show that the predicted ultimate axial compression load capacity of CFSTCSC members is comparable with that of the corresponding experimental data and the percentage difference is about ∓11%.

---