The De-Skeleton Inversion With The Integration Of Wells And Seismic And Its Application

Penberty and Shaughnessy pointed out that 70% of the world's total reserves of oil and gas distribute in the loose sand stratum,so it is necessary to study on the unconsolidated sandstone reservoir.Compressibility,the reciprocal of bulk modulus,theoretically this property of the fluid is much larger than the compressibility of the rock,in particular for gas.Based on fluid substitution equation,we can get rid of the skeleton's effects and inverse this attribute to identify gas reservoir.This paper adopts two ways to research the fluid's compressibility.One is based on the research of the related parameters of Gassmann equation,such as porosity,rock matrix modulus and dry rock modulus.The other is an approximate equation derived from the exact expression proposed by Sabrina to calculate the effective elastic modulus for the mixture,and with which we can estimate the fluid's compressibility.For porosity,we use two methods to predict it,which are double-wave method and BP neural network.Because the former is extremely sensitive to the effective pressure and the shallow loose sandstone has very strong anisotropy due to the high content of shale,and hence all these limit the application of this method.In this paper,we apply the BP neural network to establish the nonlinear relationship between the porosity and well logging attributes,and then combined with the prestack inversion data to predict porosity.Equivalent medium model,modified linear fitting equation,approximate formula are used to study on the rock matrix modulus.With little rock samples,the equivalent medium model is a simple model of the ideal,therefore it is difficult to apply to actual.Also the modified linear fitting equation will weaken the difference between data.This article uses the last way to deal with the rock matrix modulus.With regard to dry rock modulus,we handle it in two different ways.One is the improved unconsolidated sand model.As we all know,the traditional model is a combination of the critical porosity model and the lower boundary of HS's elastic theory.But the study finds that pressure coefficient attached to the critical porosity model is inaccurate,so it is modified in order to adapt to the actual environment.Meanwhile,the critical porosity model just provides the simple definition of the effective pressure.In this paper,the effective pressure can be effectively predicted by the method presented by Yu.With the help of the upper boundary of HS's elastic theory,we improve the unconsolidated sand model.The other is an adaptive method which can not only be applied to predict the dry rock modulus but also can be used to estimate the shear wave velocity,what's more,the predicted shear wave velocities are more accurate than the ones predicted by Xu-White model.Combining with well data and pre-stack inversion data,we can predict porosity,rock matrix modulus and dry rock modulus,then the fluid's compressibility can be calculated effectively.However the results are not very good,because in the process of the study of the above three parameters,there exist different degrees of approximation and it will produce the accumulation of errors.In this paper,an approximate equation is presented to predict the fluid's compressibility,which can be applied to the well data.Based on the calculated results of the well data,we can estimate the whole work area of the fluid's compressibility through the BP neural network.As a result,the gas reservoir can be identified effectively.