LAUSR

Norway’s Ringhorne Field is a faulted anticline which produces oil from Triassic (Statfjord) and Paleocene (Hermod) sands. It is located on the Utsira High. Geochemical studies of the produced oil indicate the oil is generated from mature Upper Jurassic marine shales in the adjacent Viking Graben. However, it has not been clear how oil migrated into the Triassic reservoirs and charged the overlying Paleocene reservoirs. Gas chimney detection using a proven neural network technique was used to detect the vertical hydrocarbon migration pathways on normally processed seismic data. The processing results were then validated using a set of criteria to determine if they represented true hydrocarbon migration rather than seismic artifacts. The chimney processing results using this traditional (shallow) neural network was compared with convolutional neural network (deep learning) results and geo-mechanical modeling on key lines. Key reservoirs were delineated using a stochastic (elastic) inversion approach. Reliable chimneys were then visualized in the vicinity of the producing reservoirs. The results showed pathways by which the Triassic fluvial sands received charge, and how these reservoirs had flank leakage to provide charge to shallower Paleocene reservoirs. This approach has now been used over hundreds of fields and dry holes in the Norwegian North Sea and worldwide as analogs to assess hydrocarbon charge and top seal risk predrill.