LAUSR

Abstract The purpose of this work is to present a fuzzy inference system in order to identify lithologies from wireline logs and core data from a specific borehole and transport this information to nearby uncored wells in the same oilfield. Input variables in this inference system are natural gamma ray log (GR) and porosity logs (density, neutron porosity and sonic) used to obtain M and N parameters from M-N plot in order to reduce the number of variables used and obtaining more discrete numerical intervals for the variables. The database of this inference system is built from the core information with evidence of the presence of recognised lithologies. The response of this inference system (output) is indicated along the depth of uncored boreholes where the lithologies occurring. The proposed methodology was applied to real well log data recorded in two boreholes in the Campo de Namorado (Bacia de Campos, Rio de Janeiro). In general, the system obtained 97% of accuracy, getting the conclusion that the methodology presented is applicable to the mapping of one reservoir layer (sandstone, for example) along oilfield, even at complex geologic scenarios which cause the thickness variations and of the depth of the layer of interest.