Permeability Inversion In Artificial Geothermal Reservoirs Based On The Tracer Test And Microseismic Data: Method And Applications

Hot Dry Rock is a renewable energy source which has the advantages of large reserves and wide spatial distribution.However,the inherent permeability of dry hot rock reservoirs is generally poor with the increase of reservoir depth.The dry hot rock reservoirs requires reservoir reconstruction by hydraulic fracturing to improve the fluid flow and heat transfer capabilities,which formed the the foundation for the internal hydrothermal circulation and heat extraction within the rock mass.In this process,accurate description of the spatial distribution characteristics of the permeability of dry hot rocks is crucial for optimizing well placement strategies and efficient extraction of deep geothermal energy.However,due to the large depth of the reservoirs and limited number of wells,the availability of direct monitoring data indicating the flow properties of dry hot rock reservoirs is extremely limited,making it challenging to apply most theories and techniques developed for characterizing shallow subsurface heterogeneity to the characterization of flow properties in hot dry rocks.To address this problem,this study fully considers the types of monitoring data commonly available in practical hot dry rocks geothermal development projects in enhanced geothermal systems(EGS)and establishes a theoretical framework and methodology for inverting flow properties based on tracer tests and microseismic monitoring data,which enables the fine characterization of flow properties in heat reservoirs exceeding 3 km in depth.The research focuses on artificial heat reservoirs with dense fracture network structures and uses the spatial distribution of permeability under the assumption of an equivalent porous medium,to express the spatial variations in flow properties of artificial reservoirs.The main research and findings are as follows:(1)Based solely on tracer test data,a deep learning algorithm is introduced for dimensionality reduction(DR)of the permeability parameters of hot dry rock artificial heat reservoirs to balance the number of uncertain reservoir parameters with the number of tracer test monitoring data,and the Markov Chain Monte Carlo(MCMC)stochastic inversion algorithm is implemented to form the DR-MCMC stochastic inversion method in a calculation program.Application at the Gonghe site showed that this method can effectively indicate wellbore permeability values and significantly reduce the uncertainty in permeability inversion.However,even with dimensionality reduction,the DR-MCMC method relies solely on tracer test data and has difficulty in accurately inverting the permeability of reservoirs in distant wells(especially those beyond the range of the tracer),resulting in larger errors and uncertainties in permeability inversion.(2)In order to improve the characterization accuracy of the permeability in artificial heat reservoirs(especially in distant wells),high-coverage microseismic monitoring data(MS)is introduced as a constraint.Based on the second-type Biot theory,the hydraulic diffusion coefficient is calculated using the temporal and spatial information of microseismic events as a prior estimate for the spatial distribution of permeability.Then,the MCMC algorithm is used with tracer test data to invert the conversion coefficient between the hydraulic diffusion coefficient and permeability,obtaining the posterior estimate of permeability and forming the MS-MCMC constrained inversion method.Application at the Gonghe site demonstrates that this method can better estimate the spatial distribution of permeability in artificial reservoirs,with significantly higher accuracy in characterizing the permeability in distant wells compared to the DR-MCMC method.It is suitable for inversion of the spatial distribution of permeability in artificial reservoirs when there is high-precision microseismic monitoring data.However,sensitivity analysis results indicate that the inversion results of MS-MCMC are highly sensitive to noise microseismic events.The presence of scattered noise microseismic events outside the hydraulic fracture development space can lead to overestimation of the volume of the fracture space and the values of permeability.Additionally,since this method assumes equal probability of fluid flow near wellbore in all directions,the directionality of the obtained spatial distribution of permeability is weaker than the real state.(3)In order to reduce the influences of microseismic noise events on the permeability inversion results and strengthen the characterization ability of the model for the directional spatial distribution of permeability,this paper further proposes and constructs a joint inversion algorithm for microseismic and tracer test data.The Iteration Ensemble Smoother(IES)with high computational efficiency is used for permeability iteration updating,forming the MS-Tr-IES joint inversion method.This method starts with a prior estimate of the original permeability,simulating the hydraulic fracturing process to obtain the permeability distribution and microseismic event distribution after fracturing.Based on the permeability formed by fracturing,the migration and transformation process of the tracer test is further simulated to obtain the calculated value of tracer concentration.Meanwhile,the calculations of microseismic events and tracer concentration are compared with the measurements,and the deviation is minimized as the objective function to optimize the original permeability and obtain the permeability distribution after fracturing.Since the model fully considers the hydraulic fracturing and tracer migration and transformation processes,the accuracy and reliability of the model are significantly improved compared to the MS-MCMC method.However,the computational efficiency is significantly lower than that of the MS-MCMC method.In this study,we have developed advanced permeability inversion methods for EGS by integrating characteristic monitoring data from typical EGS projects worldwide.These methods include the DR-MCMC algorithm for permeability tracing inversion,the MS-MCMC algorithm for permeability inversion with microseismic constraints,and the MS-Tr-IES joint inversion for permeability using microseismic and tracer test data.A framework for permeability inversion in artificial hot dry rock reservoirs based on microseismic and tracer test data is established.This framework is specifically designed to accurately characterize the spatial distribution of permeability in hot dry rock reservoirs with dense fracture networks,serving as a fundamental basis for predicting the hydrothermal evolution of reservoirs and achieving highly efficient heat extraction under various thermal energy extraction schemes.