关键词: 时移地震, 流体识别, 叠前反演, 岩石物理模型

Abstract:

This study focuses on the identification and dynamic monitoring of fluids in the development process of complex reservoirs, and conducts research on fluid seismic prediction technology based on  time-lapse seismic data. A fluid factor inversion method based on  time-lapse seismic data is proposed. Firstly, a time-lapse seismic rock physics model considering changes in pore pressure, saturation, and inelastic stress response is constructed to quantitatively describe the evolution mechanism of seismic response with fluid replacement and changes in reservoir conditions. On this basis, a joint inversion method for multi-period prestack seismic data has been developed. The Bayesian inversion framework is used to constrain and fuse multi-time seismic information, directly invert fluid factors that reflect fluid changes, and improve the stability and physical consistency of time-lapse signal extraction. The effectiveness of the proposed method in identifying fluid frontiers, abnormal response zones, and reservoir heterogeneity zones during injection and production processes has been verified through practical applications, especially in weak time-lapse signal backgrounds where it still has strong resolution ability.

Key words:  , time-lapse seismic, fluid identification, prestack inversion, petrophysical model