摘要
为在射孔时间未知条件下达到微震速度模型矫正的目的,采用遗传算法交叉互换搜索最优初值和Levenberg-Marquardt算法局部锁定最优解的方式来构建微震速度模型。通过6层层状模型及井下不同数量检波器进行试算,随着检波器数量增加,构建的速度模型准确度越高,当检波器数量增加到20个时,误差约13.4×10^(-3),目的层检波器数量越多,构建目的层速度误差越小,第5层检波器由1增加到3个时,精度提高9.1×10^(-3),并且反演震源与实际震源距离相差10 m,说明此方法在射孔时间未知条件下可以良好构建速度模型。
To achieve the goal of correcting velocity models when the perforation time is unknown,we proposed a method that uses genetic algorithm to get eligible starter and uses Levenberg-Marquardt algorithm to lock optimal solution for constructing velocity model.Through numerical examples on a six layered model with different quantity detectors under mine,we found that the more the number of receivers,the higher the accuracy of constructed velocity model.When the number of receivers reaches 20,the error is about 13.4 × 10^-3.And the error of constructed velocity models reduces when we increase the receiver number.When we increase the number of receivers in fifth layer from 1 to 3,the precision improved 9.1 × 10^-3,and the distance of inversed source location and real source location is 10 m,which demonstrates that our method can construct good velocity model when the perforation time is unknown.
作者
孙嘉骏
曾晓献
SUN Jia-jun ZENG Xiao-xian(College of Instrument Science and Electrical Engineering, Jilin University, Changchun 130026, China)
出处
《世界地质》
CAS
2016年第4期1127-1132,共6页
World Geology
基金
国家科技重大专项(2011ZX0560-003)资助
