Prediction of the torsional capacity of CFDST steel columns using extreme gradient boosting tree-based machine learning technique

Lenganji Simwanda, B. D. Ikotun, F. M. Ilunga, E. K. Onyari, Gatheeshgar Perampalam

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This study presents an eXtreme Gradient Boosting (XGBoost) algorithm for predicting the torsional capacity of circular Concrete-Filled Double Skin Tubular (CFDST) steel columns under pure torsion. Utilizing a dataset of 806 columns generated through non-linear finite element analysis, the of XGBoost model outperforms existing empirical models with R² values of 99.5% (training) and 97.6% (testing). SHapley Additive exPlanations (SHAP) framework aided in interpreting predictions at both global and local levels. Key influencing variables include concrete compressive strength, outer steel tube yield strength, outer steel tube thickness, and inner steel tube thickness. The study highlights the effectiveness XGBoost as a promising alternative to traditional empirical models for accurate torsional capacity predictions in CFDST steel columns.
Original languageEnglish
Title of host publicationProceedings of Nordic Steel Construction Conference 2024
PublisherThe Swedish Institute of Steel Construction
DOIs
Publication statusPublished - 2024
Event15th Nordic Steel Construction Conference - Luleå, Sweden
Duration: 26 Jun 202428 Jun 2024
https://nordicsteel2024.se/

Conference

Conference15th Nordic Steel Construction Conference
Country/TerritorySweden
CityLuleå
Period26/06/2428/06/24
Internet address

Fingerprint

Dive into the research topics of 'Prediction of the torsional capacity of CFDST steel columns using extreme gradient boosting tree-based machine learning technique'. Together they form a unique fingerprint.

Cite this