爆破震动速度峰值的预测

PDF(993 KB)

振动与冲击 ›› 2009, Vol. 28 ›› Issue (10) : 195-198.

论文

爆破震动速度峰值的预测

杨佑发;崔波

作者信息 +

Intensity of Vibration Predicition to Blasting Induced Ground Motions

Yang Youfa;Cui Bo

Author information +

文章历史 +

摘要

结合在重庆测量的42组爆破数据，针对规程里所用的波速预测经验公式，采用经验公式法、基函数回归方法以及神经网络法对爆破地震效应进行了预测。基函数回归预测法要比经验公式预测法好，比经验公式迭代法稍差，但基函数回归法的使用要方便些。人工神经网络可用于爆破地震波的三向速度峰值预测，从检验样本值与预测结果值之间的相对误差可以看出，人工神经网络预测法的精度要高于基函数回归和经验公式法。同时，对于需要考虑影响震动强度多因素变量的情况，在神经网络中通过修改输入参量即可解决爆破多参量的问题。为爆破地震效应的预测及推广应用提供了有效途径。

Abstract

Engineering blasting will produce a strong ground vibration,which may affect damage on structure, it’s necessary to predict the intensity of vibration.Based on 42 groups of blasting data,the three different methods (the empirical formula method,the function regression method and neural network method) are presented to forecast the intensity of vibration. the function regression method is more effective than the empirical formula method, the neural network method can predict three direction velocity peak values of blasting wave,It is pointed out that the precision of neural network method is the best. the neural network method can solve the many-parameter problem by the way of modifying the input variable. which submits efficient path to predict the intensity of vibration.