Petrel Platform

Blueback ODiSI

Accurately invert seismic data directly to properties including uncertainty

Overview

Blueback ODiSI is a robust joint and simultaneous stochastic inversion method that produces high-quality images of reservoir properties, such as porosity, N/G, HPV and its associated uncertainties. It works using prior distribution models based on a rock physics model and a geologically realistic model that provide close control of the lithofacies proportions

Features

Produces high-quality images of reservoir properties and uncertainties. Process is strictly based on geological concepts. Modeling is done using Markov Chains for vertical statistics then by a RPM to assign petrophysical and elastic parameters. Salient QC guides the user through every step of the process. The inversion executes as a multi-threaded task within Petrel, on a cluster or in cloud.

Specifications

No low frequency model required No lateral/spatial constraints applied Jointly inverts to both facies and rock properties (Porosity, NtG) at the same time Simultaneously inverts multiple color inverted angle stacks Provides probable values and uncertainty (mean and stdv) and high/low percentile realizations Works by generating 1000s of statistically correct (i.e. follows real-well facies distributions and rock properties) pseudo-wells, generates a synthetic to each and matches to the seismic data The best matching pseudo-wells are then selected to provide mean lithofacies and rock property values and their uncertainties

Pseudo-well lithology columns using Markov chains

For each macro layer, statistics are defined for bed thickness and transitions to form the inversion prior. The proportion of each facies within each bed can be specified, and the generated ranges QC'd to ensure the inversion is searching the required model space.

ODiSI complements Rocks QI, fully integrated in Petrel

Results from Blueback Rocks QI, like rock physics models, fluid substitution, EEI angle analysis and color inversion can be used right away as input for Blueback ODiSI, along with any other relevant data that is available inside Petrel.

Invert in Petrel, optionally on a cluster or cloud

After defining all relevant parameters, inversion computation runs as a full multi-threaded process in Petrel. Optionally, the process can be staged to run on a cluster or in the cloud. Various outputs can be selected: reservoir properties like porosity or N/G, or lithofacies cubes, high/low case percentile cubes. Add lateral continuity with Telegony.

Powerful post inversion QC options available

Perform various post-inversion QC tasks. It allows to sort best-match pseudo wells (BMPW) and to create percentile volumes (P10, P90, etc.). When making 'cases', BMPWs are sorted witin each individual layer by NtG. It also allows to create maps from averaged inverted data between surfaces.

Robust and accurate estimation of properties

Method matches large numbers of pseudo-wells to seismic impedance volumes, trace-by-trace. Pseudo-wells, defined as stochastic simulations of 1D stratigraphic profiles containing physical properties, are sampled from a prior distribution based on a RP and geological model describing the vertical stats and lithofacies proportions. The best matches provide estimates of properties and uncertainties.

Comprehensive, workflow driven GUI

A logically and comprehensive interface supports the user defining all necessary parameters step by step. Continuous controls clearly inform the user about missing parameters or data. Different scenarios can be built and saved to run independent inversion cases.

Interactive control, visualization of parameters, data

Visualize and QC all relevant parameters and data while defining the optimum settings to build lithofacies columns. From these, petrophysical and elastic curves are derived for each macro-layer using a well-calibrated rock physics model.

Robust stochastic method close to geological constraints

For trace matching to impedance volumes, EEI synthetics are computed from the pseudo-well Vp, Vs and rho logs using the EEI formula and then filtered to seismic bandwidth. 1000s of realizations are computed for each seismic trace to optimally align pseudo-wells to the EEI synthetics with seismic traces. A set of best-match pseudo-wells is composited with the lowest RMSE values

Additional Information

Jointly developed with BP. For more detailed information, please kindly refer to "Stochastic Inversion by matching to large number of pseudo wells" (Geophysics Vol 81-2016) by Patrick Connolly and Mathew J. Hughes. Cegal also offer a consultancy service for performing ODiSI inversion studies.

Success

Petrel Platform

Blueback ODiSI

Quantity

I would like to

1

Platform

Developed by

ECCN

Domain

Challenges

Lifecycle Conventional