【摘要】 水库蓄水诱发地震虽然是水利工程建设伴生的少见现象,但有时会诱发6级以上地震,加之震中距小,震源浅,成为水利工程抗震设防的重要因素。已有震例还证明,水库规模越大,诱震的可能性越大,诱震的最大震级(M)也越高。我们曾提出用水库的“综合影响参数E”来表征水库的“规模”,并给出了M～E回归关系式。本文补充了部分震例资料,并把E以及最大库深(H)都视为随机量,经模式识别的分类判别后,把所用震例分为两组(第一组为M≥4.521个震例;第二组为全部37个震例),把变量E取为确定量,或取为随机量和把变量E、lnH取为两个随机量,与M进行了回归。回归结果:第一组的双随机量(M、E)模型的回归标准差为σ_M=0.610,变异系数V=0.118,三随机量(M,E,H)σ_M=0.577,V=0.111;第二组的双随机量的σ_M=1.028,V=0.234,三随机量时σ_M=1.032,V=0.235。说明用三随机量模型对可能发生M≥4.5级的新建水库最大诱震震级的预测是可用的。更多还原

【Abstract】 Although the reservoir—induced earthquake is a scarce phcnomenom associated with the water concervancy construction, it would induce earthquake of greater than 6, so it becomes an important factor of seismic protection of water conservancy engineering. The present events prove that the larger the reservoir scale, the greater the induced earthquake may be and the higher the maximum magnitude will become. We have proposed using " comprehensive effecting parameter E " for " scale " of reservoir and given the regression relationship between M and E. In this paper, we supply partial information of events, take M, E and reservoir maximum depth H as random variables and divide all the events into two groups (the first group as M≥4. 5 has 21 events; the second one totals 37 events) after classification of pattern recognition. The results of regression show that in the first group as tri—random (M, E and H) model, regression standard deviation is σ_M=0. 577, variation coefficient is V_M=0.111. It shows that tri—random model is very useful to predict the maximum magnitude of newbuilt reservoir which will induce earthquake of M≥4.5.更多还原