top of page

Contact person: Vita Kalashnikova

 

Rune Inversion (April 2019) - Artificial Intelligence 

The very first seismic Post-Stack Inversion algorithm that allows to estimate Velocity and Density independently; possibly, with no well logs data required! Inverted Density and Velocity used as a base for Vclay, Porosity and Reservoir quality attribute - (1-Vclay)*Phi computations. Read theory here.

The algorithm was developed to utilise big volumes of seismic data that are publicly available on the Norwegian continental shelf in only the post-stack version. It helped to efficiently process mega grids (merged seismic surveys) to rock properties of P-wave velocity, density, the volume of clay, best quality sands and porosity, see example here.

The Rune Inversion algorithm in the process of independent P-wave velocity (Vp) and Density search:

 

What is the difference between conventional seismic inversion and Rune Inversion?  

  • An important unresolved aspect in the seismic inversion of all known algorithms is the determination of the absolute non-coupled velocity and density. Such rock properties restoration within the developed physics and the known solution's directions remains impossible. The accurate density estimation leads to saturation prediction. 

  • PSS-Geo Developed AI-driven seismic inversion approach that is based on a global optimisation algorithm and constraints and allows an estimation of the rocks' non-coupled P-wave, S-wave, and density. The approach applies to both post-stack and pre-stack seismic data. The first case results in P-wave velocity and density, and it requires seismic reflected signal simulation. The second case adds an estimation of the S-wave velocity and uses the real seismic reflected signals. The P-wave velocity can be computed as a full band compared to the FWI, which is computationally extensive and, therefore, of a limited band. Also, FWI does not resolve for density, while the proposed algorithm does.

  • To predict fluid, especially for post-stack data cases, we use Machine Learning algorithms - learning at the wells' target parameter, Saturation, at the particular formation and recognising the target parameter at seismic and derivatives volumes. 

 

Rune Inversion  post-stack seismic result: P-wave velocity and Density 

The example below shows a random line from the Elephant project Tampen-Tjalve area, the Norwegian continental shelf, North Sea.
P-wave velocity and Density are the results of the post-stack Rune Inversion. The inversion can be run to full time section.

Inverted Density and Velocity used as a base for Vclay, Porosity and Reservoir quality attribute - (1-Vclay)*Phi computations.

P-wave Tampen, NCS
Density Tampen, NCS

 

The inverted P-wave velocity and Density are used as a base for Volume of Clay computation, Porosity and Best Reservoir Sands attribute computations. These computation must be adjusted for different formation. Below is a result with approximate valid data window. Read theory here.

ResAtt Tampen, NCS
Porosity Tampen, NCS

 

Public vintage data. 

 

Example 1

Cross-seismic time section going through the oil&gas well 35/11-23 NPD fact page. Data was deghosted prior to inversion.   

Jurassic sands are standing out on reservoir quality attribute - yellow color, about 2.1s.

 

Example 2

Cross-seismic time section going through the AErfugl and Skarv oil&gas fields. NPD fact page.

Sands are standing out on reservoir quality attribute.

colorbar
bottom of page