Inverse Modeling Of Hydrogeological Parameters Based On Aquifer Structure Information

Various hydraulic parameters of aquifers are widely used in the study of groundwater numerical simulation,pollutant transport,dynamic monitoring,environment evolution and water resource evaluation.The solution of hydrogeological problems depends to a great extent on the accurate description of aquifers heterogeneity,so as to predict the temporal and spatial distribution and changes of water and solute in geological formations.Under normal circumstances,a limited number of pumping or injecting tests on site and corresponding head response data are not sufficient to accurately describe important geological characteristics of aquifers.How to realize the more accurate identification and depiction of abnormal area boundaries in the absence of priori information is still a difficult problem for hydrogeological parameter inversion.Hydraulic tomography is an inversion interpretation technique to obtain the distribution of heterogeneous parameters from the observed data of multiple pumping or injecting tests.In view of the problem that hydraulic tomography is fuzzy to the boundary recognition of abnormal areas and the spatial distribution resolution of hydrogeological parameters is low,a least square inversion algorithm based on cross-gradient constraints is proposed.Through inversion theory analysis,algorithm platform development and numerical simulation research,abnormal area imaging and boundary feature depiction in hydrogeological parameter inversion of structure similar aquifers are studied and explored.In aquifers with certain geological structure characteristics,the spatial distribution of hydraulic conductivity and specific storage(or transmissivity and storativity)is structurally similar.The two types of hydrogeological parameters are simultaneously inversed using hydraulic tomography,and the cross-gradients function is used to constrain in the inversion process to enhance the structural similarity of different physical property parameter models,so as to achieve the purpose of structure-coupled.The research shows that the least square inversion algorithm based on cross-gradient constraints can improve abnormal area imaging and boundary recognition effects in hydrogeological parameter inversion,solve non-uniqueness of inversion,and finally get more accurate and higher resolution inversion results.The specific research work and conclusions are as follows:(1)A structure-coupled algorithm for interpreting head response data based on cross-gradient constraints and simultaneously inversing underground hydraulic properties has been developed.The algorithm is used for hydraulic tomography experiments in various dimensions(one-dimensional,two-dimensional or three-dimensional),various scales(laboratory scale or engineering scale)and various schemes(group wells pumping and injection or group wells observation).It integrates all head response data simultaneously inversing to obtain the spatial distribution of aquifer hydraulic characteristics.(2)We develop a modular hydrogeological parameter inversion algorithm development platform based on LiveLink for MATLAB module of COMSOL Multiphysics.The platform is suitable for the inversion study of hydrogeological parameters under the condition of two-dimensional saturated and transient flow,and it is introduced structural information of cross-gradient constraints indirectly into aquifers to improve parameter estimation results.It integrates the combination of pumping or injection and observation schemes with variable flow rate pumping or injection and any time scheme observation with group wells to conform various needs of hydraulic tomography experiments.(3)Two theoretical models(horizontal tabular orebody model and slant tabular orebody model)are set up for numerical experiments,which verify the stability of the development platform and the feasibility of the inversion algorithm.The abnormal area is set to high permeability and low water storage,and the background corresponds to low permeability and high storage,i.e.hydraulic conductivity and specific storage are negatively related and structurally similar.The head response data at the end of pumping tests are more likely to estimate hydraulic conductivity,and the data at the beginning are more likely to be more suitable for estimating specific storage.The relevant areas identified by pumping and observation wells at different locations are different,and the spatial and temporal distribution of sensitivity is similar,so head response data of different observation wells at different time can be combined to obtain the high-precision spatial distribution of aquifer hydrogeological parameters.(4)The inverse results match boundaries of the abnormal area well without cross-gradient constraints,but because of strong smoothing constraints,the transition zone of parameters occupies a large area.Hydraulic conductivity is basically the same as the order of magnitude of the reference model,and the difference between the maximal value and the minimal value of specific storage is less than that of the reference model.After considering cross-gradient constraints,the transition area is compressed.The abnormal area morphology of hydraulic conductivity and specific storage is more similar,and the inversion algorithm is more conducive to estimate hydraulic conductivity.