中文摘要：

以集集地震B类场地（相当于基岩场地）的强震记录为基础，通过分析v/a和ad/v^2（a、v和d分别为地面峰值加速度、速度和位移）的取值及其随地面峰值加速度的变化规律，提出考虑变化的拐点周期值，以实现不同地震动反应谱使用其自身的地面峰值标定讨论了伪加速度放大系数谱、伪速度放大系数谱和相对位移放大系数谱在所研究的周期范围内的变异性研究了3类放大系数的概率分布情况，并发现认为放大系数谱值服从对数正态分布的假设存在不妥之处．基于前面分析和讨论，推导出了在不同周期段采用不同放大系数的分区优化标准反应谱，该反应谱具有比现有规范中的设计谱更为一致的概率意义．将此设计谱与各规范中的设计谱作比较，提出了修改我国核电规范中设计谱的参考意见．

英文摘要：

Based on the strong accelerograms on site B from the Chi-Chi Earthquake, through the analysis of the values of v/a and ad/v 2 （a, v and d are peak acceleration, velocity and displacement of ground motion respectively） and their variation parttens against peak ground acceleration, alterable period value of inflection points was put forward to realize the calibration of different response spectra with its own peak ground motion parameters. The variability of spectra of amplification factors, namely amplification factor, pseudo-velocity and relative displacement, in the whole interested period, was also discussed. According to the analysis of probability distribution of amplification factors, the common assumption that amplification factor of pseudo-velocity is in accordance with lognormal probability distribution is questionable, Based on the above/previous analysis and discussion, the optimal design spectra, with different amplification factor applied in different period, are specified with more consistent probability. Comparison between this optimal design spectrum and the design spectra specified in codes was made to give reference comments on code revision.