LM-BP神经网络在泥页岩地层横波波速拟合中的应用

Application of LM-BP neural network in simulation of shear wave velocity of shale formation

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。

Using elastic ＂wave theory, the parameters such as density, stress, and strain that affect the velocity of P-wave andS-wave are analyzed. The velocities of P-wave and S-wave are tested subsequently in different lithology, saturation state, ambient pressure and axial pressure conditions. Finally, the average relative error is estimated as 2. 22% utilizing the LM-BPneural network fit with experimental results. The results show that the lithology, saturation state and stress state are key factors that influence the relationship of the P-wave and S-wave velocity. To obtain higher accuracy, the LM-BP neural networkcan be used to fit the S-wave speed under multi-condition.