【摘要】 本文用某地区46口井的实测孔隙度、纵横波速度资料，建立了该地区孔隙度与纵横波速度关系的线性公式。由此，仅用纵横波速度便可预测孔隙度而不需要任何其它参数，并且精度明显高于wyllie公式的预测精度；但此线性公式预测的孔隙度有平均化现象。以46口井实测资料供神经网络学习，用神经网络作函数逼近，由此建立了该地区孔隙度与纵横波速度关系的非线性公式，非线性公式明显地更逼近于实测数据，提高了精度、并消除了孔隙度预测值的平均化现象。更多还原

【Abstract】 By using the observed data of porosity, P- and S-wave velocity from 46 wells in an area, we estab-lish a linear formula on the relation between porosity and P- and S-wave velocity. Thus, the porosities can be predicted by use of P- and S-wave velocities rather than any other parameters, and the precision is obviously higher than the prediction precision of Wyllie formula. But the porosities predicted by the lin-ear formla have the averaging phenomenon. The observed data from 46 wells have been used for the neural network to leam. Then the neural network is used to perform function approximation. So the nonlinear formula on the relation between porosity and P- and S-wave velocity in the area has been built.The nonlinear formula is clearly more approximate to the observed data. Therefore the precision is im-proved, and the arveraging phenomenon of the predicted porosity values is eliminated.更多还原