LM-BP神经网络在泥页岩地层横波波速拟合中的应用

发布时间：2018-03-09 04:24

本文选题：横波波速　切入点：弹性波理论　出处：《中国石油大学学报(自然科学版)》2017年03期 　论文类型：期刊论文

【摘要】：首先依据弹性波理论对影响纵横波波速的参数进行分析,明确影响横波波速的参数主要包括密度、应力载荷及应变量。根据分析结果,分别测试不同岩性、饱和状态、围压及轴压条件下的岩石纵横波波速。最后以实验结果为最初样本,通过训练LM-BP神经网络,对横波波速实验结果进行拟合,拟合平均相对误差为2.22%。结果表明,岩性、含气性及应力状态是影响纵横波波速主要因素,利用LM-BP神经网络的多条件拟合横波波速具有更高的精度。
[Abstract]:Based on the elastic wave theory, the parameters affecting the velocity of compressional and shear waves are analyzed, which include density, stress load and strain. According to the analysis results, different lithology and saturation state are measured respectively. The longitudinal and shear wave velocities of rocks under confining pressure and axial pressure. Finally, taking the experimental results as the initial sample and training the LM-BP neural network, the experimental results of shear wave velocities are fitted and the average relative error of fitting is 2.22. The results show that the lithology, The gas-bearing property and stress state are the main factors affecting the velocity of compressional and shear wave, and the multiple conditions of LM-BP neural network are used to fit the velocity of S-wave with higher accuracy.
【作者单位】：油气藏地质及开发国家重点实验室;成都理工大学能源学院;中国地质大学能源学院;
【基金】：国家自然科学基金项目(41572130)
【分类号】：P618.13;P631.81

【相似文献】