Study On Characterization And Prediction Methods Of Seismic Attributes Of Non-Homogeneous Reservoirs

Non-homogeneous reservoirs,as sand and gravel fan,buried hill and so on.Their lithology and physical properties are changing fast.Division of their deposition periods and prediction of their reservoir distribution characteristics are the difficulty of oil exploration and development work.Although conventional methods of seismic spectral imaging could solve the problem,but it's hard to guarantee the resolution.So we need to find out a better method of time-frequency spectrum analysis.The natural crevice is a main factor of capacity of non-homogeneous reservoirs.Using seismic data to predict crevice is the key point of reservoir prediction,so we need to go on the study of improving the accuracy of characterization and prediction pethods of seismic attributes of non-homogeneous reservoirs.In this thesis,we make use of the feature of seismic signal spectrum analysis in the small windows of thin interbed and carry out the study based on constrained least squares spectral inversion algorithm.And we obtain the spectrum with higer resolution both in time and frequency domain.Then we develop the study of an edge detection algorithm based on the angle sweep which can keep edge features.This work can show anomaly boundary of underground geological body better.Then we based on the automatic angle weep of three-dimensional seismic data to analyze two-dimensional surface.And the study of two-dimensional surface fitting of binary quadratic fitting equation.Then we gain the curvature properties.Further more,we carry out the study of formation surface recognition,stratigraphic discontinuity and crevice prediction.The study in this thesis direct to the feature of non-homogeneous reservoirs.By way of improving the resolution of seismic spectrum imaging and the study of characterization of seismic attributes based on the angle sweep edge detection,we gain the information of crevicedevelopment and stratigraphic lithology and reservoir properties.It can direct the exploration and development of non-homogeneous reservoirs.