中文摘要：

时间域激电法2.5维正反演算法相对较成熟,但当实测数据量较大,正反演网格剖分较精细,以及傅氏反变换波数选取较多时,常规的反演算法计算效率低,影响了算法的实用性.本文讨论如何提高时间域激电法2.5维正反演计算效率问题.以＂等效电阻率＂为基础,利用有限单元法实现了激电法2.5维正演;采用正则化反演方法实现了2.5维激电数据反演;讨论了正反演算法的并行性,实现基于MPI的正反演并行计算;对正则化反演中拉格朗日乘子搜索方案进行了优化.经理论地电模型算例测试表明,所实现的正反演算法运行稳定,结果可靠,计算效率高.

英文摘要：

2.5-dimensional forward modeling and inversion of time domain induced polarization method is relatively mature.The conventional inversion process will be very slow,affect the practicability of the algorithm.The affecting factors include large volume observed data,small grid subdivision, major wavenumberused in inverse Fourier transform and etc.In this paper,we discuss how to improve the computing efficiency of 2.5-dimensional forward modeling and inversion of time domain IP method.Based on the concept of＂equivalent resistivity＂,we employ Finite element method to accomplish IP forward modeling.And then we apply the inversion method including the regularization constraint for the model to achieve 2.5D inversion of induced polarization data.We also studied and accomplished the forward modeling and inversion parallel computation based on MPI.Moreover,we optimized the search scheme of Lagrange multiplier in regularized inversion.We tested the forward modeling and inversion algorithm on theoretical models and the results show that the algorithm is robust.The results are reliable and the computing efficiency is greatly improved.