Seismic Inversion
A successful Inversion predicts properties of existing and future wells based on seismic data. A typical seismic Inversion process Involves a series of carefully Interlinked steps, rather than a single algorithm. A number of factors control the quality of the final Inversion output.
These Include:
Quality of the well and seismic data
We use our GWLA and Gather Conditioning workflows to optimize the quality of the Individual wall log and seismic data sets and help generate the bast possible wall tie.
Adequate separation in the elastic domains
Even given good quality data, the underlying rock physics describing rock and fluids properties, and seismic response must be sufficient to allow for a clear relationship to be identifiable in the seismic domain. We use our modeling work/lows to thoroughly Investigate these relationships prior to seismic inversion.
Calibration
We use our well based rock physics model to calibrate our impedance and/or facies models to rock and fluid properties providing a quantitative measure of subsurface rock and fluid properties.
Simultaneous Elastic Impedance Inversion
RSI uses a pre-stack, constrained stratigraphic inversion for simultaneously extracting elastic properties from amplitude versus angle (AVA) Information. The Inversion algorithm was developed by the Institut Françals du Pétrole (IFP) In collaboration with consortium members.
The Inversion is a global and 3D model-based algorithm, inverting three or more angle stacks simultaneously to yield p-and 5-impedence volumes and an optional density volume. Extraction of a wavelet for each angle takes into account wavelet variations between angles. Signal-to-noise ratios in the seismic data and parameter uncertainties can be Incorporated directly into the Inversion, minimizing error in the results.
Geological constraints allow control of the confidence in each parameter (e.g. Impedance or density), and the lateral continuity of the resulting structure with a stronger correlation in the layer direction. These constraining models are created from structural frameworks, well-logs, and geostatistical kriging methods.
During the Inversion, the objective function is minimized using an iterative gradient descent method in which AVO Information, calculated from Aki & Richards modeling, can be used to explicitly specify confidence in the observed seismic data. Rock physics transforms developed during the GWLA and seismic modeling phases can be applied to the resulting Impedance volumes to provide quantitative measures of reservoir properties such as clay content, saturation and porosity.
The result is a unique and powerful inversion formulation ensuring optimum results under a variety of geological and petrophysical/rock physics scenarios.