复杂场源形态的海洋可控源电磁三维正演

在使用电偶极发射源的可控源电磁法(CSEM)勘探中,发射源的方位、长度、形状等对观测数据有重要的影响,然而现有的大部分三维数值模拟方法没有全面地将这些因素考虑进来,很多都只能应对非常简单的场源形态,例如单一方位的点电偶极子,这有可能显著降低模拟结果的准确性.本文实现了基于交错网格有限体积(FV)离散的海洋CSEM三维正演算法,能够模拟形态相对复杂的场源,包括任意方位的有限长直导线和弯曲导线发射源.该算法使用一次场/二次场方法,只需对二次场使用FV法求解,避免了场源的奇异性问题;一次场的计算为一维正演问题,使用准解析法求解,并且只要在计算一次场时考虑复杂的场源形态便可以实现同样场源的三维正演.通过与一维理论模型的解析解对比验证了三维程序的准确性,并针对三维理论模型进行了一系列正演测试,初步考察了场源形态对三维正演结果的影响.

Three-dimensional forward modeling of the marine controlled-source electromagnetic field with complex source geometries

In controlled-source electromagnetic (CSEM) surveys that utilize electric dipole sources, observations can be greatly influenced by the orientation, length and the shape of the source. However, most of the current 3D CSEM modeling algorithms do not take all these factors into account, and many are only able to deal with very simple source geometries such as a point dipole with a fixed orientation, which may cause significant inaccuracies in the modeling result. A 3D forward modeling code for marine controlled-source electromagnetic (CSEM) surveys has been developed, which permits simulations of complex source geometries.The solution is based on a staggered-grid finite volume (FV) discretization of the second-order Maxwell's equations. The primary/secondary field approach, in which only the secondary fields need to be solved using the FV method, is employed to avoid the source singularity problem. To accurately compute the primary fields over an 1D layered background model, the open source software Dipole1D is integrated into the 3D code. Due to Dipole1D's versatility and the fact that the primary field calculation is the only part of the 3D modeling that directly handles the source geometry, the 3D code is capable of simulating arbitrarily orientated finite-length wire sources. The resulting linear system of FV equations is solved by the quasi-minimum residual (QMR) method.To validate the 3D code, we first benchmarked FV solutions against quasi-analytic solutions for a canonical 1D reservoir model. The numerical solutions show excellent agreement with analytic solutions for various transmitter orientations for both a point dipole source and a 300 m wire source. Then we calculated responses of a 3D reservoir model similar to a well-studied one, varying horizontal sizes of the resistor. Reasonable results for the 3D model, which are consistent with published work, can be observed.Further, a suite of numerical experiments on the 3D model were conducted to investigate the influences of source geometries, and the observations include: (1) The full lateral extent of the 3D resistor can be well estimated from the responses normalized by responses of the background model without the resistor, by moving and rotating the transmitter. (2) Electromagnetic fields excited by wire sources with wire lengths ranging from 100 m to 500 m are compared with those by a point dipole, showing that the longer the source, the greater the differences, and the EM field amplitudes of 500 m wire source can be up to more than 4.5 times larger than that of a point dipole. Prominent differences occur not only in the region immediately around the source, but also in four narrow regions at oblique angles to the source. (3) While the 300 m straight wire source is slightly bended in four different ways, either in the horizontal plane or in the vertical plane, one or several components of the EM field change significantly.From results of all the numerical experiments, we can conclude that the presented 3D FV CSEM modeling code is accurate and versatile, capable of modeling relatively complicated source geometries, such as finite-length straight/bent wire sources with arbitrary orientations. In CSEM surveys, source geometries, i.e. the orientation, the length and the shape of the source can have remarkable influence on the observed data, thus accounting for source geometry effects is necessary.