Syntethic Correlation Petrel

Schlumberger Public

Synthetic seismogram correlation (Petrel 2009.1.1)

Introduction & Objective

● Petrel misses a correlation tool for synthetic seismogram ● In this presentation we will explain a work around to correlate a synthetic

Reflection coefficients ● Three processes will be used in Petrel 2009.1.1 1) Seismic to well tie 2) Genetic Inversion 3) Train estimation model

The objective is to have a numeric correlation indicator for synthetic seismogram generation 2

Schlumberger Public

seismogram to a real seismic trace based on either Acoustic Impedance or

Overview

● Generate synthetic seismogram ● Output Reflectivity Coefficient and Acoustic Impedance logs

● Use Correlation analysis tool under train estimation model (NN) ● Check the correlation factor to quality check the synthetic seismogram

3

Schlumberger Public

● Run G.I to obtain R.C inverted cube (Linear and Non-Linear)

Workflow scheme Seismic to Well Tie RC

RC Train estimation model

4

AI

Schlumberger Public

AI

Genetic Inversion

Step 1 Seismic to well tie ● Generate logset ● Wavelet extraction ● A.I and R.C output Schlumberger Public

5

Step 2 Genetic Inversion ● Generate Acoustic Impedance inverted dataset ● Generate Reflection Coefficients inverted dataset ● A.I and R.C logs QC Schlumberger Public

6

Step 3 Train estimation model ● Train AI and check the Correlation analysis ● Train RC and check the Correlation analysis ● A.I and R.C logs QC Schlumberger Public

7

Summary & Conclusion

● The proposed workflow gives a relatively accurate numeric indicator for correlation. ● A.I and R.C are major components of a seismic trace that could be used for correlation

in G.I and also using AI or RC rather than the synthetic trace itself

8

Schlumberger Public

● The proposed workflow is preliminary and not 100% accurate due to the uncertainty