Rapid on-site peak ground acceleration estimation based on SVM and P-wave features in Taiwan

Rapid on-site peak ground acceleration estimation based on SVM and P-wave features in Taiwan

This study extracts pressure wave(P-wave) features from the first few seconds of recorded vertical ground acceleration at a single station. These features include the predominant period, peak acceleration amplitude, peak velocity amplitude, peak displacement amplitude, cumulative absolute velocity, and the integral of the squared velocity. The support vector regression method is employed to establish a regression model, which can predict the peak ground acceleration according to these features. Some representative earthquake records of the Taiwan Strong Motion Instrumentation Program from 1992 to 2006 are used to train and validate the support vector regression model. Then the constructed model is tested using entire earthquake records from the same period as well as the 2010 Kaohsiung earthquake with a magnitude of 6.4 ML. The effects on the performance of the regression models using different P-wave features and different lengths of time window to extract these features are studied. The results illustrate that, if the first three seconds of the vertical ground acceleration are used, then the standard deviation of the predicted peak ground acceleration error for the entire 15 years of earthquake records tested is 20.89 gal. The time window can be shortened, e.g., to 1 second, increasing the prediction error slightly, in order to lengthen the lead-time before destructive shear waves reach the station.