Structural parameters and dynamic loading identification from incomplete measurements: Approach and validation

System identification is becoming more important in structural dynamic applications including structural model update, damage detection primarily due to the rapid increase in the number of damaged or deteriorated engineering structures, and load identification for remaining service life forecasting. Time-domain identification techniques based on measured vibration data, e.g. the least-squares estimation (LSE), have been studied for a relatively long time and shown to be useful. However, the traditional least-squares techniques require that all the external excitation measurements should be available, which may not be the case for many practical applications. In this paper, by introducing a weighted positive definite matrix to the objective function and a learning coefficient using the information in the previous iterations in the identification approach to improve the convergence performance, an alternative iterative approach for both structural parameters and dynamic loading identification, referred to as weighted adaptive iterative least-squares estimation with incomplete measured excitations (WAILSE-IME), was proposed. The accuracy, convergence, and robustness of the proposed approach was demonstrated via numerical simulation on a six-story shear building model with noise-free and different levels of noise-polluted structural dynamic response measurements. The effects of the positive weight matrix, the learning coefficient, and the sampling duration on the convergence and the accuracy of the proposed approach were discussed and the results were compared with available related literature results. Results show that the proposed approach can simultaneously identify structural parameters and unknown excitations within very limited iterations with high accuracy and shows its robustness even noise-polluted dynamic response measurements are utilized. Furthermore, the approach was validated via available experimental measurements for a four-story frame structure excited by impact excitations. The results show that the proposed approach is capable of identifying both structural parameters and the unmeasured excitations with acceptable accuracy and improved convergence.

► The approach can identify structural parameters and the unmeasured excitations.
► Performance of the approach was demonstrated via simulation and experiment.
► Effects of the factors of the approach on the convergence and the accuracy were discussed.
► Performance of the approach was compared with the available related literature results.