Abstract：Identifying the natural properties of the fluid is one of the significant tasks involving reservoir evaluation, and the consequential results would directly affect the understanding and development of oil-gas filed. By taking the advantage of a variety of logging curves as parameters, the stylebook hierarchical decomposition cross-plot and decision tree methods are utilized to identify the fluid natural properties in a Sudanese oil-gas field. Finally a classification modeling is obtained. The results indicate that (1) The accuracy of a single cross-plot could be reached as high as 80%; however, there is still an abundant room for the improvement if the stylebook hierarchical decomposition cross-plot is used in spite of shortcomings. (2) The parameters could be automatically selected in decision tree by calculating the weight of the parameters, thus the accuracy could be further reached at 86.7%. Combination implementation of the cross-plot and decision tree is able to improve the accuracy up to 100%, and the shortcomings are able to be made up. (3) According to the conventional understanding, an uniform oil-gas interface is expected to occur in the Sudanese filed, however with the applications of the decision tree, the filed now is regarded as a layered reservoir, the oil-field has achieved better results by adjusting the development strategy which is depended on the results of the decision tree.
王礼常;王志章;陶果. 样本分解交汇图法与决策树方法识别苏丹某油田流体性质[J]. , 2012, 30(2): 22-27.
WANG Lichang;WANG Zhizhang;TAO Guo. Fluid Nature Identification of an Oil-gas Field in Sultan Using Stylebook Hierarchical Decomposition Cross-plot and Decision Tree Methods. , 2012, 30(2): 22-27.