Research And Application Of Seismic Predicting Methods For Deep Stratigraphic-Lithologic Reservoirs

Our old oilfields have gone into the second and even third recovery.So we need to explore into deep areas to make old oilfields maintain and even increase production.But,until now,our country study scantly for the deep oil and gas fields,which are still a difficult problem.Yongjin oilfield in Junggar Basin and buried deeply has huge reserves,which is a typical stratigraphic-lithologic reservoirs.Its reservoir is deep and single group is thin to lead to predict difficultly.In this issue,we,firstly,refer to numerous references of deep stratigraphic-lithologic reservoirs and fully understand and analyze the whole work area in accordance with the geological conditions,the logging data and seismic facies.This is key to the structure and lithological interpretation and the foundation to reservoir prediction and comprehensive evaluation of oil and gas reservoirs.Considering the problems of the limitations of seismic,logging data in the work area,Target processing for the data is essential.Due to the presence of the underlying coal seams in this work area,the signal of the reservoir interfered by the signal of coal seams.In this paper,the MP algorithm is applied to remove the coal seams.And HHT transform is applied to solve this problem of low resolution,which achieves good results.Characterizing finely the range of reservoir is tireless pursuit goal of geologists,which is guiding for developing wells?In this paper,we prefer amplitude attribute from under through of overlying layer by conducting forward modeling for different models.On the basis of spectral decomposition,extract amplitude attribute slice from under through at different frequencies to recognize pinch-out point and achieve good results.Then,using a variety of attributes,combined with the technology of integration of multiple attributes,describe qualitatively the reservoir of the work area,which laid a good foundation for the subsequent quantitative prediction of reservoir.After the completion of the qualitative description of the reservoir,geologists are more concerned about the thickness of the reservoir.In this paper,multivariate regression analysis,BP neural network,combined with the results of qualitative prediction,predict quantitatively the thickness of reservoir.Considering the less the less logging data in work area,the accuracy of the above method is limited.Finally,impedance inversion data is used to predict the thickness of reservoir and the result is better.