| Abstrakt: |
Abstract: The prediction of thin-bed thickness has been generally recognized as one of the challenges in seismic exploration. Based on joint constraints of pseudo- and real-well, a seismic method for predicting the thin-bed thickness is presented in this paper. By analyzing geological and petrophysical characteristics of thin beds and surrounding rocks, a geological model, which is consistent with the geological features and represents the variation of thin-bed thicknesses, is established to conduct the forward modeling. Using the model, pseudo-well samples are “created” by extracting thin-bed thicknesses and their corresponding seismic attributes from the forward modeling. On the other hand, real-well samples formed by real thin-bed thicknesses and their seismic attributes are extracted from the wells and seismic data. A joint sample set combining the pseudo- and real-well samples is then used to fit the correlation between seismic attributes and thinbed thicknesses quantitatively, which can finally be used to predict the distribution of thinbed thickness. By making a full use of the pseudo- and real-well data in establishing the quantitative relationship between seismic attributes and thin-bed thicknesses, especially in the early stage of oil and gas exploration when only few wells are available, this method increases the number of valid samples and could largely reduce ambiguity in interpretations of seismic attributes, which may improve the accuracy of the prediction significantly. The method presented here was applied to predict the thin-bed thickness of Carboniferous formation in Fuling district, and a new well drilled in the area confirmed the accuracy of the prediction. |
