Inversion of Surface Nuclear Magnetic Resonance by Regularization with Simulated Atomic Transition Method

Abstract

Initial water content impacts the accuracy and resolution of the inversion in surface nuclear magnetic resonance (SNMR). In order to solve this problem, a new method called as regularization combined with simulated atomic transition method (RSATA) is proposed. The inversion of SNMR is transformed into an unconstrained nonlinear global optimization problem, and it solved directly by RSATA without pre-assigning the initial water content distribution. The conjugate gradient linear iterative algorithm is adopted to look for local minimum and global extreme value when using RSATA, and has improved the inversion speed. Results show that this method is a very good solution to solve the effect of initial water content and it is also better than the existing methods on the operation efficiency and the accuracy of inversion.

DOI: http://dx.doi.org/10.11591/telkomnika.v11i9.3248