Identification and Characterization of Naturally Fractured Reservoirs Using Conventional Well Logs

In petroleum exploration and production, fractures are one of the most common and important geological structures, for they have a significant effect on reservoir fluid flow. Despite their importance, detection and characterization of natural fractures remains a difficult problem for engineers, geologists and geophysicists. This paper presents a technique for the identification and characterization of naturally fractured reservoirs using conventional well logs. Logs are the most readily available source of information, however they are seldomly used in a systematic manner for quantitative analysis of naturally fractured reservoirs. Since all well logs are affected in one way or another by the presence of fractures, a Fuzzy Inference System is implemented in this study to obtain a fracture index using only data from conventional well logs. Additionally, a self-consistent model from O’Connell and Budiansky for the prediction of elastic properties of fractured porous rocks is inverted using genetic algorithms to obtain crack density and crack aspect ratio. The proposed algorithms are tested using data available from the Mills McGee # 1, an Austin Chalk formation well in Milam County, Texas. The results obtained are compared with core information available.