Publications

There is an increasing demand facing the petroleum industry for more realistic and detailed reservoir description in order to better predict reservoir performance. Geostatistical modeling is a state of the art answer to this demand. Geostatistical modeling characterizes spatial continuity as well as honoring all control data. A realistic description of spatial continuity is crucial for predicting reservoir performance. Faulting in a reservoir is an important feature affecting spatial continuity. Conventional geostatistical modeling either requires that faults be determined before hand or ignores faults altogether. This paper introduces a new tool in the geostatistical arsenal that maps faults by correlating well logs.The first step in this method is geostatistical correlation of well logs by variogram minimization. This establishes the stratigraphic alignment of the wells. Then the local dip is calculated from the stratigraphic alignments. Fault throw is then mapped as the discontinuity of the stratigraphic alignments.Two examples are included in this paper. One is from the Kern River Field in California. The first step in correlating the well logs identifies bad data in some well logs or coordinates. After correlating the wells, the dip and discontinuity are mapped. Discontinuity along dip shows the throw of strike-trending faults. Another example is from the Minas field in Indonesia. These examples show how this new method can help geologists characterizing a reservoir.