Seismic Reservoir Inversion Method Research And Application

In the light of the important position for Seismic reservoir inversion in the modern oil and gas exploration, this article discusses the main principles and methods of the poststack and prestack seismic inversion. AVO inversion is set as the main object of study for Prestack inversion, and this paper uses the Matlab software for AVO forward modeling and inversion modeling. For post-stack inversion this paper studies sparse spike inversion and geological statistics inversion, and uses Jason software for the actual inversion of the seismic data. The comparison inversion results for sparse spike inversion and geological statistics inversion shows that the geological statistics inversion is better than the sparse spike inversion. The former has higher resolution, and the inversed gamma section fit well at the well and the gather near the log. The former fits more the rule of the geology, and can be used better for the reservoir prediction.Firstly this paper introduces the basic principles of the post-stack sparse spike inversion, geological statistics inversion and AVO inversion, and gives the derivation of the formula. This paper also gives the derivation of conversion formulas for rock physical parameters in the AVO inversion. For post-stack sparse spike inversion, it primarily bases on the principle that the reflection coefficients are sparse. It calculates the minimum reflection coefficient which makes the synthetic seismic and real seismic fittest for each other. Geological statistics inversion mainly uses the geological statistics method, and it can use sequential Gaussian simulation for simulating. For getting the unknown simulate distribution node value from the known distribution node values, it needs solving kriging equation to find the mean and the variance of the unknown node. After determining the characteristics of an unknown node distribution statistics, we randomly select a value from the distribution as the predicted value for the unknown node. For AVO, the reflection coefficient changes with the angle, and it associates with the media on both sides of the arguments. Thus, the arguments of the media on both sides can be inverted from the reflection coefficient, such as longitudinal and transverse wave velocity and density.With the theory of the Poststack and prestack inversion, the third chapter of this paper describes the inversion programming process using Matlab software for linear inversion of post-stack and pre-stack AVO, whose purpose is to study the post stack inversion of prestack inversion theory further. The part of stacked linear inverse modeling gives the relation between the damping coefficient and variance between the synthetic seismic and the real seismic records. For AVO prestack inversion simulation this paper mainly uses approximate formula of Aki and Richard. From the modeling results, it can conclude that the reflection coefficient get from approximate formula of Aki and Richard and real Zoeppritz formula are nearly the same when the incidence is less than 50 degrees, whereas the incidence is larger than 50 degrees, the difference between the two reflection coefficient will be large.The last part of the thesis conduct the post-stack sparse spike inversion and geological statistics inversion for real seismic data, for this real area, compared the two inversion result, geological statistics inversion is better than the sparse pulse inversion.