基于粒子群优化的双频激电数据联合反演

崔益安^{1, 2}, 李溪阳^{1, 2}, 向恩明^{1, 2}，柳建新^{1, 2}，朱肖雄^{1, 2}, 纪铜鑫^{1, 2}, 佟铁钢^{1, 2}

(1. 中南大学 地球科学与信息物理学院, 长沙 410083；
2. 中南大学 有色资源与地质灾害探查湖南省重点实验室, 长沙 410083)

摘 要：双频激电法是一种非常有效的探测极化目标体的勘探方法，可以同时观测获取视电阻率和极化率数据。为了提高双频激电观测数据解译的可靠性，利用粒子群优化算法对电阻率数据和极化率进行联合反演。在对异常目标体采用旋转椭球体进行几何近似模拟的基础上，基于Core-Core散射理论实现双频激电法的快速数值模拟计算，为粒子群优化联合反演提供正演基础。在对粒子群优化算法参数进行分析设定的基础上设计了联合反演算法，并采用加入不同程度噪声的模拟双频激电数据对算法进行了实验测试。测试结果表明：粒子群优化联合反演算法能有效实现电阻率与极化率数据的联合反演，具有很强的抗噪声能力，且算法收敛快、稳定性好。进一步的实测数据联合反演测试还表明：该算法具有较低的模型初始信息依赖性，在给定较大搜索空间的条件下仍能反演出较为满意的结果模型，具有较好的实用性。

关键词：联合反演；双频激电；粒子群优化算法；视电阻率；极化率

Joint inversion of dual frequency IP data using PSO

CUI Yi-an^{1, 2}, LI Xi-yang^{1, 2}, XIANG En-ming^{1, 2}, LIU Jian-xin^{1, 2}, ZHU Xiao-xiong^{1, 2}, JI Tong-xin^{1, 2}, TONG Tie-gang^{1, 2}

(1. School of Geosciences and Info-Physics, Central South University, Changsha 410083, China;
2. Hunan Key Laboratory of Non-ferrous Resources and Geological Hazard Detection,
Central South University, Changsha 410083, China)

Abstract:Dual frequency IP method is very effective to detect polarizable targets even in mountain terrains. Both resistivity and polarizability data can be observed during dual frequency IP measurement. In order to interpret dual frequency IP data quantitatively and more accurately, a joint inversion was proposed based on particle swarm optimization (PSO). Through simulating anomalies by ellipsoids, the forward of dual frequency IP can be given based on Core-Core theory, which provides the foundation to implement joint inversion by PSO. After selecting appropriate PSO algorithm parameters, the joint inversion is designed meticulously. Then synthetic data with noise in different degrees is used to test the joint inversion algorithm. The results of testing show that the joint inversion is effective to invert dual frequency IP data with high tolerance to noise and fast convergence. Further field data tests demonstrate that the PSO inversion is not sensitive to the initial model information. It can get a satisfactory result model even in a large given search space. That is very meaningful for interpreting field data in dual frequency IP observation.

Key words:joint inversion; dual frequency IP; particle swarm optimization; resistivity; polarizability