| Groundwater is an important source of drinking water for residents,as well as a geological force or hazard factor that permeates various production fields.The use of high-tech methods to accurately detect groundwater and conduct resource assessment and sustainable development is currently an important issue to be addressed.Surface nuclear magnetic resonance(SNMR)is the only geophysical exploration method in the world that directly detects groundwater and has advantages that other geophysical exploration methods do not possess.Traditional SNMR methods adopt a working mode of placing coils on a horizontal surface for depth measurement and use analytical methods to calculate the excitation electromagnetic field at each point in the underground half space.If the terrain of the measurement area has undulations,the coils are still approximated as being deployed under a horizontal terrain,and the calculated excitation electromagnetic field will produce large errors.In response to SNMR detection in undulating terrain,this paper uses the method of calculating the excitation electromagnetic field through the segmented integration of the harmonic electric dipole(HED)along the path.Non-structured grids are selected as the finite element partition strategy,and two-dimensional spatial grids under various terrain conditions are established.Different water-containing models are set in the partition space for2-D Occam inversion numerical simulation research.Finally,joint detection of a water-bearing sand layer in a certain area was carried out using high-density electrical resistivity method and SNMR method in field experiments,and 2-D imaging was conducted using Occam inversion method to preliminarily verify the feasibility of 2-D Occam inversion for measured data.The main research work and achievements of this paper are as follows:1.For the calculation of coil responses in undulating terrain,the coil can be segmented based on the undulating condition of the coupling ground,and the HED integral can be adopted to obtain the excitation magnetic field of each segment and perform vector superposition,thereby correcting the calculation error caused by terrain undulation.This approach achieves accurate calculation of the excitation magnetic field and kernel function distribution of SNMR coils in undulating terrain.2.In the 2-D SNMR forward and inverse simulations with finite element method and topography,the non-structured mesh generated by Distmesh mesh generator can achieve high-quality partitioning of complex underground spaces that are difficult to fit with traditional structured meshes.The spatial attribute data output based on this partitioning model can be used to calculate the response of SNMR electromagnetic fields and the distribution of kernel functions more accurately,providing a good basis for 2-D inversion simulation.3.For the 2-D Occam inversion method of SNMR,controlling the linear search process and model smoothness is a key step in determining the success of the inversion.The core of the method is to determine the search direction and step size by selecting appropriate initial regularization parameters,adopt reasonable smoothing strategies to constrain the model,and optimize the inversion results.In the numerical simulation process,the 2-D Occam inversion results under horizontal,single oblique,and embankment terrain conditions were discussed.The results show that,with the optimization of the transmission parameters,the reconstruction of the preset water-containing models in several terrains by the 2-D Occam inversion is excellent,but the performance under high noise conditions is not ideal.Therefore,before applying the Occam inversion method,it is necessary to suppress the noise in the original data to obtain a high signal-to-noise ratio initial signal to achieve better inversion results.4.The ERT and SNMR methods were used for outdoor data collection in field experiments,and the HED integral method was used to calculate the coil response and kernel function under undulating terrain.Combining the initial data collected by field measurements,the 2-D Occam inversion imaging was performed,and the inversion profile was cross-validated with the traditional 1-D inversion histogram and high-density resistivity method inversion results,which preliminarily proved the feasibility of the 2-D Occam inversion method for field data. |
