Combination of mechanical and informational modeling to predict hysteresis behavior of I beam‐to‐CFT column connection |
| |
| Authors: | GR Abdollahzadeh S Yapang‐Gharavi M Hoseinali‐Beygi |
| |
| Affiliation: | Faculty of Civil Engineering, Babol Noshirvani University of Technology, Babol, Iran |
| |
| Abstract: | The behavior of beam‐to‐column connections is significantly complicated because it depends on the material, nonlinear relationship between different members, and contact of connecting elements. Their hysteretic response exhibits highly inelastic characteristics and continuous variation in stiffness, strength, and ductility. Therefore, accurate hysteretic models of bolted connections are essential to accurate seismic assessment and design. Composed (Hybrid) modeling approach is presented to represent the complex hysteretic behavior of I beam‐to‐concrete‐filled‐tube (CFT) column bolted connection. For the purpose in this study, experimental evaluation of I beam‐to‐CFT column connection with bolted end‐plate under cyclic loading is done. It must be mentioned to make the connection easily applicable practically and also prepare the column flanges continuity, a connection with the bolts penetrated into the CFT column are used since using the stiffeners in hollow sections is not easily applicable. In the composed mechanical–informational modeling framework, the conventional mechanical model is complemented by information‐based model components. Finally, moment–rotation curves of composed model have been compared with experimental results. Comparison showed that the composed model has a good agreement with experimental results. A newly designed specimen for validation purpose of the composed modeling was made and experimented. Composed model was well applied to predict the behavior of new connections. The presented results show that the composed model has a good ability to predict of hysteretic behavior. |
| |
| Keywords: | CFT column informational modeling mechanical modeling moment– rotation curve neural network |
|
|
