Model updating hybrid testing method based on dual adaptive unscented Kalman filter algorithm

Abstract

Model updating hybrid testing method provides crucial technical support for assessing the seismic performance of engineering structures. The model-based unscented Kalman filter (UKF) algorithm and its improved variants have become the mainstream identification choice for hybrid testing due to their high practicality and precision. However, when the statistical characteristics of system noise involve uncertainties, existing UKF-based identification algorithms may suffer from filter divergence, reduced accuracy, and decreased efficiency in MUHTM. To address these issues, this paper proposes a novel model updating hybrid testing method based on dual adaptive UKF algorithm (MUHTM-DAUKF). Firstly, the DAUKF algorithm is proposed, which integrates a Sage-Husa adaptive noise estimator module to dynamically adjust statistical characteristics of the noise and an adaptive variance module to diminish the risk of filter divergence. Furthermore, the MUHTM-DAUKF is proposed, which utilizes the DAUKF algorithm to identify and update the constitutive model parameters based on measured data from experimental substructures. This enhances the accuracy of numerical substructures and improves the overall reliability of MUHTM. Lastly, the effectiveness and accuracy of the proposed methods are validated by numerical simulations and experimental tests. It is shown from the numerical simulation results that the DAUKF algorithm is feasible for parameter identification, whilst the MUHTM-DAUKF exhibits superior accuracy and computational efficiency compared to the MUHTM based on adaptive UKF algorithm (MUHTM-AUKF) and the MUHTM based on dual adaptive filter approach (MUHTM-DAFA). The experimental results further validate the effectiveness and reliability of the MUHTM-DAUKF and the superiority of the MUHTM-DAUKF over the MUHTM-AUKF and the MUHTM-DAFA. These findings indicate that the proposed MUHTM-DAUKF has strong potential for seismic performance assessment of complex engineering structures.