Study on 3‐D Resistivity Inversion Using Conjugate Gradient Method

吴小平,徐果明

DOI: https://doi.org/10.1002/cjg2.55

2000-01-01

Chinese Journal of Geophysics

Abstract:We present a 3‐D minimum structure inversion algorithm for d.c. resistivity data, using conjugate gradient (CG) relaxation techniques. Firstly, to solve the minimum structure inverse equations, the CG relaxation techniques only require the multiplications of the Jacobian matrix G and its transpose G T with ab itrary vectors, x and y respectively. When the Rodi method is introduced to compute the G , Gx and G T y can be obtained by one forward calculation in each inverse iteration. Therefore the inversion is computed much quickly. In this way we can also avoid direct computing the inverses of G and G T G and save the huge storage for G and G T G . In addition, as there are too many inversion parameters, which effectively erase unnecessary structures and obtain trustworthy results close to true model, we impose a smoothness constraint on the model in minimum structure inversion to obtain the smoothest solution. We combine these methods and techniques into 3‐D resistivity inversion and obtain very promising results with 3‐D synthetic data. We also developed new iterative technique to deal with the problem of very slow convergence rate in the traditional minimum structure inversion. This technique requires no more than ten iterations to reach the convergence.