Parameter identification of ductile timber beam-to-steel column connections: Procedure and assessment

Abstract

Ductile beam–column connections in a moment-resistant timber frame are essential part that can affect the seismic performance of the timber frame. As the first step, this study presents a ductile timber beam-to-steel column joint connected by a steel link, herein designed as a box-shaped steel bracket. A robust unscented Kalman filter (UKF) algorithm was presented to identify the model parameters of the proposed the connection. In addition, a dual-scale assessment method is proposed to evaluate the accuracy of the identified results. UKF algorithm is limited to the problem of nonpositive definite matrices in high dimensional parameter identification. To prevent the interruption of the program and improve the stability of the program, the Cholesky decomposition in the UKF was superseded by singular value decomposition (SVD), a noise estimator was embedded into the UKF to self-adapt the noise covariance matrix during the iteration. The assessment was carried out for both the steel links and timber beam-to-steel column connections. Subsequently the framework was introduced for the parameter identification of the steel links. For the hysteretic curves, performance indices, and energy dissipation, there was good agreement between the experimental and numerical results. The steel bracket improves the ductility of the timber beam-to-steel column connection. The average ductility coefficient of the steel bracket with ribs can reach 7.11. The proposed parameter identification procedure is verified to be applicable in the field of ductile timber beam-to-steel column connections. The program stability is improved and the error of most of the performance parameters is within 10%. However, this method still need to be verified for dynamic analysis in the whole building.