随钻电磁波传播测井传感器参数优化设计

 随钻电磁波传播测井是随钻测井系列中最重要的一种测井方法，它通过记录电磁波信号的幅度比和相位差来反映地层介质信息。不同的频率、源距的测井仪器参数会产生不同的测井响应特性，选择适当的传感器参数能够提高其探测效率。应用有限元法对非均质地层进行正演模拟，得到随钻电磁波传播测井方法的纵向分辨率与径向探测深度。采用神经网络的方法辅助传感器参数优化设计，计算不同地层电阻率、频率、源距的纵向与径向函数。通过计算值分析得到优化后的传感器参数，并计算新传感器参数的视电阻率正演响应。计算结果虽然与正演计算数值有一定误差，但误差很小，并且在可以接受的范围内。随着正演模拟样本的增加，神经网络的方法能够有效辅助参数优化设计，降低计算次数。设计的新型传感器够有效反映地层电阻率。

Optimized Design for Sensor Parameters for Electromagnetic Wave Propagation in Logging While Drilling

Electromagnetic wave propagation is one of the most important method of logging while drilling. The formation medium information is obtained through the amplitude ratio and phase difference of the electromagnetic wave. Different logging response characteristics can be generated by different frequency, different source spacing. Detection efficiency can be improved by selecting proper sensor parameters. Finite Element Method(FEM) is used to analyze the forward problem for inhomogeneous formation. Vertical resolution and radial exploration depth are calculated. The optimized design is achieved by using the neural network. The functions of vertical resolution and radial exploration depth against formation resistivity, frequency and source spacing are calculated by using FEM and neural network to obtain the optimized parameters of sensor. It is shown that the forward simulation results agree well with the results calculated by the neural network, with errors in an acceptable range. The neural network is an effective auxiliary method for optimized designing, which can reduce the forward simulation calculation time. Formation resistivity can be effectively reflected through the use of the new design parameters.