Parameter Prediction Of Hydraulic Fracture For Tight Reservoir Based On Micro-seismic And Dynamic Production Data

Hydraulic stimulation is an important measure for the development of tight reservoirs.In order to describe the distribution of hydraulic fractures,micro-seismic monitoring technology was introduced into petroleum fields.Micro-seismic events may reveal important information about rock properties and the static characteristics of hydraulic.However,this method is limited to reflect the distribution area of the hydraulic fractures and fails to provide specific parameters such as length and azimuthal angle.Therefore,micro-seismic is integrated with dynamic development data to predict the real hydraulic fracture parameters in this paper.Forward modeling micro-seismic wave(P-wave)propagation path based on shooting ray tracing method and double-difference method is adopted to construct the objective function,and then Simultaneous Perturbation Stochastic Approximation(SPSA)algorithm is applied for source location.Give the initial parameters of the hydraulic fracture according to the prediction of the micro-seismic events' distribution.On the basis of micro-seismic inversion results,we consider secondary modeling to calibrate the real parameters information of hydraulic fractures by history matching,where the DFM(Discrete Fracture Model)and EDFM(Embedded Discrete Fracture Model)is used for simulate.In addition,the fractal theory is used to study the generation and simulation of the fracture network,which make it more consistent with the real fracture morphology.In the aspect of history matching,we established the object function based on Bayesian theory and apply Hybrid SPSA Algorithm and Finite Difference Method are used to matching the production data(oil production,water production,etc.)by adjusting the azimuth and length of the hydraulic fracture.Theoretical analysis,the error of micro-seismic inversion results and real events is controlled below 15 m.The error is fluctuating in each direction,which is in accordance with the Gaussian disturbance distribution.In general,the inversion results are expected.In the case of history matching,the results show the oil/water production and the cumulative oil/water production matching well.Comparative analysis the initial shape,the true form and the final pattern,and the results show that the method can be more accurately approximated the real distribution of hydraulic fracture.Overall,the results clearly show the effectiveness of assimilating micro-seismic data into dynamic data to estimate the parameters of real hydraulic fractures,which has a guiding significance for construction site.