Application of time series analysis on performance evaluation of seismic structures

The state of arts of time series analysis and structural performance evaluation is reviewed briefly in this paper. To address the drawbacks of existing evaluation methods, a time series analysis method is introduced to evaluate the structural performance which is based on the time varying Vector Autoregressive Moving Average (VARMA) model. By modeling structures as time varying VARMA models, Kalman filter is used to estimate the corresponding VARMA coefficients. The modal and stiffness parameters of structures can be derived from the time varying coefficients which can be used to evaluate structural performance. Taking the simulation result of IASC-ASCE benchmark structure and the nonlinear dynamic analysis result of a multistory frame structure as examples, the conventional Frequency Domain Decomposition (FDD) method and the VARMA method are used to evaluate the performance of the two structures and the comparison of them shows that the VARMA method is more precise and simple which is quite suitable for the case that the natural frequencies are close and the signal to noise ratio is small. What is more important. the VARMA method is implemented in time domain thus it can describe the evolution of structural performance thoroughly and evaluate structural performance online. At the end of the paper, the development trend of the VARMA method is also prospected.