AN OPEN ACCESS JOURNAL
 JJCE Submission Home


Effectiveness Of Modern Data-based Prediction Models For The Axial Load Capacity Of Cfdst Columns

Submitted2022-03-29
Last Update2022-10-01
TitleEffectiveness Of Modern Data-based Prediction Models For The Axial Load Capacity Of Cfdst Columns
Author(s)Author #1
Author title:Research Scholar
Name: Nikhil Bembade
Org: Walchand College of Engineering, Sangli, Maharashtra, India. 416415
Country: India
Email: nikhil.bembade@walchandsangli.ac.in

Author #2
Author title:Professor Doctor
Name: Shrirang Tande
Org: Walchand College of Engineering, Sangli, Maharashtra, India. 416415
Country: India
Email: hod.apm@walchandsangli.ac.in

Other Author(s)
Contact AuthorAuthor #1
Alt Email: nikhilbembade@gmail.com
Telephone: +919822245672
KeywordsCFDST, Poisson Regression Model (PRM), Artificial Neural Network (ANN), predictive models.
AbstractApplying different composite materials is advantageous and common in the construction field nowadays. By Considering various advantages like enhanced strength, stiffness, fire resistance for various loading conditions, and less construction time, Concrete-Filled Double-Skin Tubes (CFDST) members are used. CFDST members can be utilized as columns, beams, beam-columns, stub-columns. Various approaches and formulae based on different codes are available for estimating the axial load capacity of CFDST columns. This study aims to develop data-based models for predicting the axial load capacity of CFDST columns and check their effectiveness and applicability. One hundred twenty-five experimental data samples related to the axial load capacity of CFDST are collected from various research papers. The Artificial Neural Network (ANN) model and Poisson Regression Model (PRM) are developed to predict the axial load capacity of CFDST columns using this data. In addition to this, axial load capacities for one hundred twenty-five experimental samples are calculated using two standard code based equations. The applicability and effectiveness of these data-based models are checked by comparing all predicted and calculated axial load capacities with experimental capacities. Statistical comparison considering parameters like the coefficient of determination (R2), Mean Absolute Deviation (MAD), Mean Square Error (MSE), Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE) revealed that all of these models could effectively predict axial load capacity of CFDST columns. Data-based models PRM and ANN are pro over two standard code-based equations regarding accuracy and robustness with the R2 value of 0.994 and 0.999, respectively.
Paperview paper 6660.pdf (1272KB)

http://www.just.edu.jo