Cross-equalization for Time-lapse Study in Balol Field, India

Summary A successful 4D imaging requires that the non-repeatable noise due to acquisition or processing between successive seismic surveys should be minimized to bring out 4D signal unambiguously. However, even well-designed, dedicated imelapse studies are contaminated by several sources of non-repeatable noise such as differences in source waveform, varying near-surface conditions, changes in geometry/offsets due to logistics and differences in source/receiver positioning, processin g algorithms/flows, data dependent operators, velocity models and statics. The accurate matching of different datasets is a key issue in a time-lapse study to extract useful 4D signal. In this paper, we present a case study to match two 3D datasets acquired at an interval of 12 months to monitor a thermal EOR process in the northern part of Balol Field of Cambay basin, India. The first 3D survey was conducted prior to the commencement of in-situ combustion process in the already identified set of injector wells and therefore was meant to serve as base 3D survey. The second 3D survey spaced 12 months thereafter was to become 1 st monitor survey for studying the changes in the reservoir due to this thermal process. The matching procedure begins with identically processed PSTM gathers of the two datasets as inputs that undergo a step-by-step processing flow for gain correction, frequency balance, differential statics, phase adjustment and finally designing moving window match filter. The matched datasets are then differenced to derive 4D signal. Appropriate QC at each stage of the 4D workflow ensures that the desired 4D effect is preserved as the two datasets become increasingly comparable and look-alike in the non-reservoir zones where ideally no changes are expected. The final difference volume clearly shows high amplitude 4D anomalies around injector wells due to combustion/air injection. The possible fairways of flue gases and/or propagation of combustion can also be brought out through horizon/time slices and vertical sections.