论文部分内容阅读
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。
LWDWD logging is the most important logging method in logging while drilling. It records formation media information by recording the amplitude ratio and phase difference of electromagnetic wave signals. Logging instrument parameters with different frequencies and source distances will generate different logging response characteristics, and selecting the appropriate sensor parameters can improve the detection efficiency. Finite element method is used to simulate the heterogeneity strata. The vertical resolution and radial depth of the electromagnetic wave propagation while drilling are obtained. The method of neural network is used to assist the optimization of sensor parameters to calculate the longitudinal and radial functions of resistivity, frequency and source distance in different formations. The optimized sensor parameters are obtained through the analysis of the calculated values and the apparent resistivity forward response of the new sensor parameters is calculated. Although the calculation results have some errors with the forward calculation value, the error is small and within the acceptable range. With the increase of forward modeling samples, the neural network method can effectively assist the parameter optimization design and reduce the calculation times. The new sensor designed effectively reflects the formation resistivity.