Research on the ability to explore the edge of electromagnetic removal of electromagnetic remains based on HIE-FDTD simulation
Lanqiang Zhou, Sinan Fang, Zeyu Zhang, Guilan Lin, Chengxiang DengAbstract
Accurate detection of residual oil and gas in deep anisotropic reservoirs is critical for improving the efficiency of energy resource utilization. However, conventional cross-well electromagnetic approaches encounter significant challenges when modeling sub-metered scale geological structures characterized by strong resistivity anisotropy and thin-layer stratifications. To address these limitations, a hybrid implicit-explicit finite-difference time-domain algorithm incorporating a thin plate mesh, enabling high-precision 3D forward modeling of low-frequency cross-well electromagnetic fields. Numerical experiments demonstrate that the proposed method effectively resolves sub-meter-scale geological structures while maintaining computational efficiency. In sandstone-mudstone formations (background conductivity: 0.05 S/m), cross-well electromagnetic measurements achieved optimal signal-to-noise ratios for residual oil detection. By studying the response mechanisms of fine cross-well residual oil to cross-well transient electromagnetic fields, it was found that when a low-conductivity body is located in the center of the cross-well electromagnetic transmitter-receiver coils, the y component of the magnetic field strength (Hy) shows the change process from negative to positive, which serves as a robust diagnostic for residual oil and gas. Further investigation of anomalous bodies with varying conductivity shows that oil-bearing and water-bearing layers can be distinguished according to the onset timing of the negative Hy anomaly. Therefore, this method provides a reliable and high-resolution approach for the accurate localization of cross-well residual oil and gas resources, with substantial strategic value for strengthening national energy security and promoting the efficient development and utilization of oil and gas resources.