Quantitative Identification And Thecharacteristics Of Near-fault Pulse-like Ground Motions

Several severe earthquakes occurred close to metropolises in the 1990 s have inflicted heavy casualties and considerable economic losses on the human society. Since then, lots of scientists began to pay their attention to near-fault pulse-like ground motions, which are considered to have caused the above sufferings. Because of being much closer to the earthquake fault, near-fault pulse-like ground motions are believed to contain much more seismic energy than nonpulse-like ground motions. The structural damages are usually severer than expected when a building structure is situated on a site that is close to an earthquake fault. Therefore, it is of significant practical meaning to perform comprehensive studies over the characteristics and the effects of near-fault pulse-like ground motions. Better understanding of such information can be helpful to guanrantee t he security of those engineered structures located close to the earthquake sources. Four main topics are discussed in this study: quantitative identification of near-fault pulse-like ground motions, the characteristics of pulse-like ground motions, the narrow-band modified ground motion attenuation relationship and earthquake damage potential represented by vector-valued intensity measures. The main contents and conclusions are as the following:(1) Quantitative identification of near-fault pulse-like ground motions. In light of the limitations of the current pulse-classifying methods, an energy-based pulse identification approach is proposed here. To romove the interference from high-frequency contents, a simplified pulse model is first adopted to extract the velocity pulse from the orginal velocity time history. The pulse period, pulse- starting and-ending time points are then determined by the peak-point-method, whose validity is confirmed by comparing with two other methods. The relative cumulative energy, which is determined by the pulse-starting and-ending time points, is used as the indicator for identifying a velocity pulse. The proposed method is further used to detect pulselike ground motions with forward-directivity effects. It is believed that if the pulse-starting time precedes the mid-point of the ? on time axis, the record will be considered as a forward-directivity-induced ground motion.(2) Characteristics of near-fault pulse-like ground motions. Based on the knowledge that there is a need to separate pulse-like and nonpulse-like ground motions, the amplification effects of vocity pulses on the acceleration specra are discussed in detail. It is verified that the period region where the amplification effects exist would move towards the right along the period axis, and this region can become broader as the increase of the pulse period. The differences of the two pusel-extraction methods for investigating pulse-directivity effects are performed. It is demonstrated that the LSF method is much more stable and efficient in extracting velocity pulses than is the wavelet analysis. Finally, the correlation between the relative pulse energy and the structural resposes are confirmed. Under pre-defined condition, the relative pulse energy can be used as a reliable parameter for representing the earthquake intensity measure.(3) Narrow-band modification of ground motion attenuation relationships. By conducting statistical anlaysis, the period region where the amplification effects exist is quantitatively analyzed. It is indicated that the starting period, the maximum period, the ending period and the amplification factor are all correlated well with the pulse period. Based on the statistical results, the analytical expressions of the period region where the amplification effects exist and the amplification factor are presented. It is considered by comparing with an existing narrow-band modification model that the spectral values modified by the existing model can be larger than the proposed model.(4) Earthquake damage potential represented by vector-valued intensity measures. By the correlation analysis between the ground motion intensity measures and the overall structural damage indices, two independent intensity measures are selected to represent the earthquake damage potential. It is confirmed that the two intensity measures can not only guarantee the structural reponse prediction precision, but also decrease the prediction dispersion that is induced by the complex earthquake ground motions.