Seismic performance of timber mortise-tenon joints reinforced with a self-centering variable-friction damper

To improve the seismic performance of traditional timber mortise-and-tenon (M–T) joints, this study proposes a cylindrical self-centering variable-friction damper consisting of a steel casing, rubber friction elements, and a disc spring stack. The damper dissipates energy through displacement-dependent friction between the rubber elements and steel casing, while the disc spring stack provides restoring force. Low-cycle reversed loading tests, FE simulations, and retrofitting tests on straight-tenon and stepped-through-tenon joints were conducted to evaluate the damper performance and strengthening effectiveness. The effects of rubber hardness and length on the hysteretic behavior were investigated, and a simplified moment–rotation model was developed for strengthened joints. Results showed that the 80 Shore A rubber with a length of 50 mm exhibited the highest energy dissipation capacity, reaching 466.22 kN·mm per cycle at a displacement amplitude of 10 mm. The FE model reasonably reproduced the main hysteretic characteristics, with peak force errors below 19% and energy dissipation errors within 10%. Comparisons of the tested strengthened and unstrengthened specimen pairs indicated increases of 58.87% and 8.24% in initial stiffness, 43.57% and 192.18% in peak bending moment, and 27.13% and 42.37% in equivalent damping ratio, respectively, while residual pull-out displacements decreased by 23.5% and 28.0%. These results represent the comparative responses of the tested specimens due to the limited sample size. The proposed moment–rotation model reasonably describes the skeleton response of strengthened joints within the investigated rotation range of −0.15 rad to 0.15 rad, providing a simplified approach for describing the nonlinear rotational behavior of reinforced traditional timber joints.