The Research Of Fracture Prediction Method Based On Inversion Error Analysis

At present,the exploration targets are changing from conventional reservoirs to unconventional reservoirs and special reservoirs.Fractured reservoirs,as a typical complex reservoir,have great exploration potential.There are many methods of fracture prediction,mainly indirect seismic method.Seismic methods for fracture prediction mainly include coherence cube technology,edge detection technology,non-linear technology,shear wave splitting technology,vertical seismic profile technology and converted wave fracture prediction technology.In this paper,inversion error is proposed for fracture prediction.The theoretical basis is that the more complex the structure is,the more complex the shape of seismic wave field can be obtained by numerical simulation.The lower the accuracy of the inversion results,the greater the inversion error.Therefore,the inversion error characteristics can be used to predict the fracture area.Firstly,the basic theory and propagation characteristics of fractured media(HTI media)are studied,and the influence of Thomsen parameters on the velocity is verified intuitively by elastic wave group,three-dimensional velocity curves of phase velocity and two-dimensional velocity curves.Then a series of medium models with cracks are established.The propagation law of seismic wave in medium with cracks is analyzed by using finite difference numerical simulation and adding transparent boundary conditions.In this paper,Bayesian inversion algorithm is used for inversion.The theoretical basis of Bayesian inversion is described in detail,and the specific process of inversion is given.In order to improve the accuracy of inversion and the accuracy of fracture prediction by inversion error,the traditional Bayesian inversion algorithm is improved in this paper.The main improvement points include the improvement of the initial inversion impedance model,the prior probability distribution and the likelihood function.Finally,a set of high precision Bayesian inversion algorithm is obtained.The stability and applicability of the algorithm are tested through its anti-noise analysis.The improved Bayesian inversion algorithm is used to invert the fractured model.The inversion results are inferior to the actual results.Error analysis is carried out and error profiles are generated.Then the error profiles are compared with the records of actual fracture development zones.The corresponding locations of fracture development zones are found on the error profiles,and the performance characteristics of inversion error profiles in fracture development zones are explored.On this basis,fracture prediction is carried out.Through the inversion of the actual work area,the characteristics of error profile corresponding to the fracture development area are explored,and the inversion error analysis method suitable for fracture prediction is summarized.