Comparative hysteretic models for bolted moment resisting timber frame connections

Dowel-type timber connections, such as bolted beam–column joints, govern the stiffness, strength, and ductility of mass timber structures, particularly in moment-resisting timber frames (MRTFs) subjected to seismic loading. Their cyclic response is characterized by pronounced pinching, stiffness and strength degradation, which must be accurately represented in nonlinear analysis for seismic design and assessment of MRTFs. This paper reviews and compares existing hysteretic models implemented in OpenSees for simulating bolted timber connections, focusing on Pinching4, SAWS, Hysteretic, IMKPinching, and DowelType. Experimental moment–rotation data from previously published cyclic tests on unreinforced and reinforced glulam beam–column connections, as well as a two-dimensional (2D) single-bay, single-story moment-resisting timber frame tested are adopted as a common benchmark. Model parameters are calibrated using the NL2SOL nonlinear least-squares algorithm within the quoFEM framework, minimizing the error between numerically predicted and experimentally measured hysteresis loops. The models are evaluated in terms of their ability to reproduce hysteresis loop shape, pinching, strength degradation, and cumulative energy dissipation. Results reveal that the DowelType model achieves the best overall performance across all specimens, maintaining the highest stability and lowest cumulative absolute error with R 2 exceeding 0.94. While SAWS demonstrates robust all-around capability in capturing experimental hysteresis loops, a clear performance split emerges: SAWS and Pinching4 provide superior accuracy for unreinforced, slip- and pinching-dominated connections, whereas the Hysteretic and IMKPinching models excel in capturing the ductile response of reinforced specimens.