Online weighted LS-SVM for hysteretic structural system identification

The identification of structural damage is an important objective of health monitoring for civil infrastructures. Frequently, damage to a structure may be reflected by a change of some system parameters, such as a degradation of the stiffness. In this paper, we propose an online sequential weighted Least Squares Support Vector Machine (LS-SVM) technique to identify the structural parameters and their changes when vibration data involve damage events. It efficiently updates a trained LS-SVM by means of incremental updating and decremental pruning algorithms whenever a sample is added to, or removed from, the training set, and robustness is improved by the use of an additional weighted LS-SVM step. This method overcomes the drawback of sparseness lost within the LS-SVM and makes LS-SVM for online system identification possible. The proposed method is capable of tracking abrupt or slow time changes of the system parameters from which the damage event and the severity of the structural damage can be detected and evaluated. Simulation results for tracking the parametric non-stationary changes of non-linear hysteretic structures are presented to demonstrate the application and effectiveness of the proposed technique in detecting the structural damage.