Time-lapse Seismic Analysis Through Prestack Inversion and Rock Physics

Time-lapse seismic reservoir monitoring is based on the fact that production/injection significantly changes the reservoir's fluid saturations, pressure, and temperature, which subsequently alters the acoustic response of the reservoir rock, and this can be detected using seismic data. Changes in a reservoir's saturation and pressure can be derived through high-resolution seismic inversion and subsequent petroelastic calibration.

An integrated workflow for analysis of time-lapse data comprises:

• Prestack amplitude versus offset (AVO) inversion
• Derivation of a rock physics model
• Calibration to reservoir properties

Prestack inversion using the Aki-Richards (1980) three-term method can be used to derive acoustic and shear impedances and density. Data with a good signal-to-noise ratio up to incidence angles of 42 degrees are key to the success of this approach.

The acoustic contrasts (compressional, shear, and density) from the AVO analysis are conditioned by space -adaptive wavelet shaping, and implicit spectral inversion to a common, broadband, symmetrical, zero-phase wavelet.

Seismic relative acoustic impedance is obtained using a sparse spike-type inversion that requires no a priori information; it is unbiased because it is independently determined on a trace-by-trace basis. The absolute acoustic impedances are obtained after the addition of the low-frequency component from the available well data.

A rock physics model is constructed to describe the elastic response of the reservoir to variations in rock matrix, fluid saturation, and reservoir pressure. The model is calibrated to available core and wireline data. This is then used to determine saturation and pressure changes associated with a given set of seismic attributes.

Inversion to reservoir properties of fluid saturation and pressure is carried out using a proprietary technique that consists of three stages:

• Building a geological model over the zone of interest and populating it with the reservoir properties (for example lithology, porosity, and saturation) using all the available data (including well, core, reservoir simulation, horizons, and faults). This involves a combination of deterministic, geostatistic and stochastic methods.
• Forward modeling of the reservoir properties to acoustic and shear impedances using the appropriate rock-physics model.
• The forward-modeled reservoir properties are matched to the inverted acoustic impedances using a linear sequential matching process. The final results were a set of high-resolution seismically and geologically constrained reservoir properties.

The combination of high-resolution seismic surveys with excellent repeatability, high-quality inversion and rock physics-based calibration results in a detailed quantitative 4D analysis of a hydrocarbon reservoir. The results shows that AI and SI change as low as 2% can be detected and accurately related to variation in fluid saturation and/or pressure.