The Application of AFMM Aided System Identification on Ground Vibration Monitoring Data Analysis

Abstract eng:
The parametric system identification method has been used to construct soil layer models and compute average model parameters of the ground vibration monitoring data at ChingLiao. However, from the laboratory tests, the results show that the dynamic properties of soil are not always constant. An important goal of this study is to verify the soil properties changing with time. Before analyzing the data set, it is segmented into time-variant and time-invariant parts by adaptive forgetting through multiple models method (AFMM). Soil properties are estimated by both time-variant and time-invariant parametric modeling methods showing changes of system parameters. The results of this study show that the AFMM can reduce the error of estimation for time-variant model and has advantages of sorting data set in analysis. The modal frequencies decrease during main shaking area after a certain threshold strain and they are relevant to earthquake intensity. Damping ratios might be underestimated without segmentation while modal frequencies might be overestimated. The results show that there is a limit for linear time-variant models to identify earthquakes above seismic intensity 6. If non-linear time-variant models can be used to simulate the strong ground motion, the result would be more accurate.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket