2-D Occam inversion study of snmr method on unstructured grids based on local encryption
The Surface Nuclear Magnetic Resonance(SNMR)method,as a direct means of detecting subsurface water,exhibits high sensitivity in exploring aquifer properties.Traditional SNMR inversion techniques can provide information on the distribution of water content within a certain depth range,but they are limited in their resolution of water boundaries under finite pulse excitations.In order to overcome these challenges and enhance the capability to detect the distribution characteristics of water-bearing structures,this study employs numerical simulation methods.Initially,three geological models—ancient river channels,aquifer faults,and water-rich Karst formations are established.Unstructured mesh partitioning and localized mesh refinement are applied to these models.Comparative analyses of the Occam inversion imaging effects under minimum model constraints and directional constraints are conducted.Subsequently,the study validates the numerical simulations through field experiments in the Qinghai-Xizang Plateau.The results demonstrate that local mesh refinement on high-water-content areas using unstructured grids effectively enhances the resolution of water structure boundaries.Additionally,incorporating directional constraints along with prior information leads to more accurate inversion results.
Surface Nuclear Magnetic Resonance(SNMR)Local encryptionUnstructured gridsOccam inversionDirection constraint