Time-Lapse Seismic at the Heidrun Field: Advantages of Cross-Disciplinary Work

摘要

A time-lapse repeated (4D) seismic survey was conducted in 2004 (1) over the southern part of the Heidran oil field, following the time-lapse seismic project performed from 1986 to 2001 (2). The Heidrun field is an oil field with an initial gas cap. The 1986-2001 time-lapse seismic project established time-lapse seismic as a useful technology for identifying remaining oil at the Heidrun field, contributing to the planning of four new wells in the field. The purpose of the 2004 repeated seismic survey was to improve reservoir understanding and drainage strategy in the Fangst and Upper Tilje reservoir units, which due to their location, production properties and timing of the repeat survey were favourable for time-lapse seismic. At the Heidrun field the majority of time-lapse seismic effects are fluid replacements, because the drainage strategy is pressure maintenance by water and gas injection. Around some of the water injectors we see indications of pore pressure buildup, but interpretation is in general based on oil replacement by water or gas, respectively. The 2004 time-lapse seismic data contributed to detail planning one well. The main contribution, however, was to improve drainage understanding and strategy. The results were used for the planning of several new drilling targets, and an important tool contributing to this was refinement of flooding maps. The 1986-2001 time-lapse seismic project was a conventional 3D versus Q marine 3D acquisition. The non repeatable noise level was relatively high, which lead to some difficulties in interpretation, especially of intra-reservoir changes. The 2001-2004 data were acquired as Q marine versus Q marine (Q on Q) surveys, thereby improving the repeatability of the data. Consequently, intra-reservoir changes were easier to interpret and seismic time-lapse inversion became a useful interpretation tool. The 2001 data were acquired in September and the 2004 data in June, enabling less than three years of production between the surveys, compared to the more than six years production changes observed in the 1986-2001 time-lapse seismic experiment. Consequently, fluid changes from 2001 to 2004 were much smaller than in the previous time-period, meaning that the data would hardly be interpretable if repeatability were not improved. After the 1986-2001 time-lapse seismic study, we generated flooding maps at three reservoir levels. Prior to the 2001-2004 study, we extended these to six, using all available data in the field. After having the 2004 data, this was further extended to nine levels. The time-lapse seismic data gave valuable input to five of the nine flooding map levels.