中文摘要：

对近场速度脉冲型地震而言,其速度时程的主脉冲形状通常是不对称的,为了克服该现象对量化识别速度脉冲型地震动带来的影响,基于经验模态分解（EMD）良好的自适应分解能力,该文建立了以本征模态函数（IMF）重构为基础的速度脉冲量化识别方法。使用该方法进行信号重构之后,能够有效实现信号的去噪平滑,进而通过峰点法（PPM）较为准确地确定脉冲周期,最终通过其能量大小定义脉冲特征,该文还将该方法拓展到了多脉冲识别,并在脉冲方向的显著性方面进行了验证。研究结果表明：相比其他手段,基于IMF重构的脉冲量化识别方法能够更有效地识别不对称脉冲记录,具有更好的工程适用性。

英文摘要：

The configurational asymmetry of main pulses always hinders the quantitative identification for near fault pulse-like earthquake records. On account of adaptive decomposition capacity provided by the empirical mode decomposition（EMD）, a method using the Intrinsic Mode Function（IMF） reconstruction was presented in this paper for pulse-like identification. An IMF reconstruction approach was developed to denoise and smooth the signal, then the peak-point method（PPM） was applied to define the period of single main pulse, and the energy of single pulse was the ultimate step to identify pulse-like ground motion and the corresponding criteria were composed. Furthermore, the approach was extended to the identification for multi-pulses-like records. Finally, pulse-like significance in arbitrary orientations was tested to verify the method. Results show that the proposed method is more applicable in earthquake engineering due to its advantages in configurational asymmetrical pulse determination.