Evaluation of gusset plate moment connection at elevated temperatures using gene expression programming and finite element model
Year: 2025
Type: article
Abstract: • An innovative steel moment connection is investigated under fire conditions. • Moment-rotation-temperature relationships of the connection are explored to provide a deeper insight into the thermal and mechanical response of the connection. • The performance of the connection (as a cost-effective connection) is assessed in the multi-story buildings. • Gene Expression Programming is utilized to predict the behavior of the connection. In this study, the performance of the gusset plate moment connection under fire conditions is examined using finite element modeling and genetic algorithm. Following validation, the finite element modeling is conducted in three stages. Initially, temperature-moment-rotation curves are generated for the connection at various temperatures. Subsequently, the connection's performance in a four-story frame is assessed under two scenarios and three sub-scenarios to develop its application. In the third stage, considering various parameters, diverse connection models are expanded to generate the necessary data for genetic algorithm modeling and to predict connection performance at elevated temperatures. The finite element model results indicate that the extracted temperature-moment-rotation curves possess appropriate accuracy. Therefore, with the availability of a reliable model, it is possible to accurately predict the variations in the connection's stiffness at different temperatures. Additionally, the rotation amount, critical temperature, failure, and internal force of the connection in different stories of the four-story frame are obtained and compared across various scenarios. Finally, two genetic algorithm models are presented to predict connection performance for temperature ranges below and above 600°C. Comparison between experimental results and the genetic model shows that the relative error percentages of the proposed models are 10.42% and 12.47%, respectively, with correlation coefficients of 0.90 and 0.88.
Source: Results in Engineering
Authors: Hesam Ketabdari, Amir Saedi Daryan
Institution: Shahid Beheshti University
Language: English
Subfield: Civil and Structural Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $1,940
OpenAlex ID: W4415424483