NEW APPROACH: USING RELATIVE INVERSION WITH SPECTRAL DECOMPOSITION TO DISTINGUISH THIN LAYERS IN THE 33-SERIES SAND RESERVOIRS OF THE WIDURI FIELD, SOUTHEAST SUMATRA, INDONESIA

摘要

Widuri Field is a mature oil field which consists of seven reservoir intervals and located in shallow water (20-50m) off the coast of Southeast Sumatra. One of reservoir intervals is known as the 33-series sand which is seismically picked as Gita horizon in TalangAkar Formation (TAF) at early Miocene. This reservoir has contributed 15% of total oil production. The reservoir is interpreted as a complex compartmentalized channel, where distributary channels, mouth bars and delta front shoals are observed. The 33-series reservoir interval are divided into two different reservoir sandstones (known as the 33-4 and 33-6 sands), where limestone stringers are also present in between. The main challenge of the 33-series reservoir interval is how to separate into single body sand with low tuning thickness issue, moreover both of the channels are most likely lies on same seismic even as well as limestone presence. The distribution of the channel sands is very critical to study the continued success of the water-flood program. Conventional seismic interpretation was conducted to distinguish these two sands. However as both reservoirs fall within a single seismic event in time this is usually difficult. Therefore, interpretation was conducted in the Time-Frequency domain because interference patterns were observed in the amplitude spectrum of the seismic which appeared to correlate with the thin layers we were trying to map. It was a concern that interference patterns were sub-seismic resolution however further investigation was undertaken. The Spectral Decomposition method was applied and seismic geomorphology obtained. In time-frequency responded over a short window of data and the results were interpreted using RGB blending. However, working with the constant frequency volumes derived from the seismic data were unable to distinguish the 33-4 and the 33-6 sands because they had very similar thicknesses. To resolve the thickness ambiguity between the 33-4 and 33-6 sands and map them a relative impedance inversion was conducted and the phase volumes from this inversion analyzed for seismic geomorphology. With this method, it is now possible to distinguish 33-4 and 33-6 sands in reservoir interval which has validated by analyzing on over 300 wells. The method demonstrates the value of combining relative impedance with spectral decomposition to plan a successful water-flood. This study has further identified some new interesting leads/prospects that are not visible on the conventional seismic interpretation. Therefore, it is hoped to extend the Widuri Field life cycle.