自组织神经网络在煤矿地质异常的横向预测中的应用
详细信息 本馆镜像全文    |  推荐本文 | | 获取馆网全文
摘要
自组织神经网络是一种学习能力强、收敛速度快的智能系统,可用于解决分类、聚类、识别异常等问题。依据煤田地震勘探中的地震波运动学和动力学特征,提取地震波的最大互相关系数、分形关联维、主频、频带宽度、主频带能量共五组特征参数,利用自组织神经网络横向预测地质异常。计算结果表明,该方法可行,可望成为预测地质异常的一种有效方法。
Self-organized nervenet is an intelligent system with strong study ability and rapid astringency which can be used in solving classification,clustering and identification of abmormity. It is used to forecast geologi- cal abnormity transeversely in this paper based on the characteristics of kinematics and dynamics of earth- quake wave, five groups parameters are selected such as the maximal correlatable coefficient of earthquake wave,relevant dimension,main frequency, bandwidth and the energy of main band. Results show that this method is feasible and can be an efficient method to forecast the geological abnormity.
引文
[1]张继先,宓霞.神经网络及其在工程中的应用[M].北京:机械工业出版社,1996.
    [2]蔡煜东,杨兵,汤军彪.自组织人工神经网络在煤层对比判别中的应用[J].物探化探计算技术,1995,17(4):81-87.
    [3]刘玲,黄玲,张小东,等.用自组织映射方法进行油气预测[J].石油物探,1994,33(4):56-64.
    [4]陈遵德,朱广生. Kohonen网络在油气横向预测中的应用[J].石油物探,1995,34(2):53-56.
    [5]蔡煜东,杨兵,汤军彪.自组织人工神经网络预测碳酸岩中岩溶洞穴规模[J].物探化探计算技术,1995,17(2):40-44.

版权所有:© 2023 中国地质图书馆 中国地质调查局地学文献中心