Reservoir Characterisation and Modelling

Application of Digital Core Analysis

The results from this Project can assist in more reliable estimation of rock properties compared to conventional core experiments which can lead to a greater confidence in reservoir characterisation and plume migration behaviour by providing inputs to static and dynamic models.

Explore

Formation micro-imaging log automated interpretation

The project’s aim is to develop a systematic mapping alternative to resource-intensive manual interpretations and better leveraging of expensive, high-resolution Formation Micro-Imager (FMI) data to characterise the subsurface CO2 storage system.

Explore

Application of Digital Core Analysis

Using the digital core analysis (DCA) results that have been performed on core plugs from one of CO2CRC’s wells (CRC-3) and investigate their application in improving the CO2 injection modelling

Objectives

This Project used the digital core analysis (DCA) results that were performed on core plugs to investigate their application in improving the CO2 injection modelling from perspectives below:

  • Prediction of permeability in the model
  • Multiphase flow parameters and rock types
  • Directional multiphase flow parameters.

Results & outcomes

R
All multiphase flow parameters can be obtained from a single core sample using DCA. This not only reduces the cost of reservoir characterisation compared to conventional SCAL experiments but also helps to deploy more representative multiphase flow parameters in the reservoir modelling
R
DCA can measure porosity and permeability every 3 mm along the core at higher uncertainty, whereas conventional core analysis measure the data at a much coarser scale (every few 10cm at best) but with less uncertainty
R
DCA can measure anisotropic relative permeability data, which is an important parameter to use in reservoir characterisation and modelling
R
DCA can generate porosity and permeability for an important injection interval in for which data are missing from RCA

Publications

Cook P. J Geologically Storing Carbon, Learning from the Otway Project Experience, CSIRO Publishing 2014.

Ashworth P, Rodriguez S and Miller A, Jenkins C, Case study of the CO2CRC Otway National Project, Energy Transformed Flagship, CSIRO, 2011.

Sandrine Vidal-Gilbert, S, Tenthorey E,  Dewhurst, D, Ennis-King J, Hilli R, Geomechanical analysis of the Naylor Field, Otway Basin, Australia: Implications for CO2 injection and storage 2010, International Journal of Greenhouse Gas Control, Volume 4, Issue 5, 827-839

 

Underschultz, J., Boreham, C., Dance, T., Stalker, L., Freifeld, B., Kirste, D., Ennis-King, J., 2011. CO2 storage in a depleted gas field: An overview of the CO2CRC Otway Project and initial results. International Journal of Greenhouse Gas Control 5, 922-932.

Jenkins C, 2013, Statistical aspects of monitoring and verification, International Journal of Greenhouse Gas Control 13, 215- 229

 

Boreham, C., Underschultz, J., Stalker, L., Kirste, D., Freifeld, B., Jenkins, C., Ennis-King, J., 2011. Monitoring of CO2 storage in a depleted natural gas reservoir: Gas geochemistry from the CO2CRC Otway Project, Australia. International Journal of Greenhouse Gas Control 5, 1039-1054.

Dance T, Spencer L, and Xu J 2018 Geological characterisation of the CO2CRC Otway Project Site – What a difference a well makes. Link to poster.

Dance T, A Workflow for Storage Site Characterisation: A Case Study from the CO2CRC Otway Project Site. Conference Presentation AAPG 2009 Hedberg Conference.

Noble, R, Stalker, L, Wakelin S, Pejcic B, Leybourne M, Hortle A, Michal K; 2012, Biological monitoring of carbon capture and storage –’ A review and potential future developments. International Journal of Greenhouse Gas Control 10, 520 – 535

Formation micro-imaging log automated interpretation

The project’s aim is to develop a systematic mapping alternative to resource-intensive manual interpretations and better leveraging of expensive, high-resolution Formation Micro-Imager (FMI) data to characterise the subsurface CO2 storage system.

The technical basis of this project will be development of a sophisticated image recognition machine learning system using convolutional neural networks

Objectives

  • Classification of intervals intersected at new wells using the family of known lithotypes previously encountered
  • Identification of any new lithotypes not previously encountered for additional manual interpretation
  • Objective identification of stratigraphic boundaries and brittle strain features intersected by these new wells
  • A trained image texture recognition and classification model used for the purpose of satisfying objectives 1 & 2

    Rationale

    Lithology is an important property to consider when studying reservoirs. Based on this property rock types are defined where each rock type has specific petrophysical and mechanical properties directly linked to its exploitation capability. Lithofacies identification or rock typing using image logs, such as those from Formation Micro-Imager (FMI) is a time consuming and laborious task. Many attempts have been made in the past to automate this process; however, this is very much an open research problem.

    In this context Machine Learning (ML) and Artificial Intelligence (AI) has shown to be quite promising.

    Deep Convolutional Neural Networks (CNN) are a particular type of neural networks more suited to data types with some sort of distribution either spatial or temporal. These networks have shown to be very efficient in extracting data patterns and detecting boundaries where data trends change.

    In this research we aim to design and train a CNN suitable for automatic detection of lithotypes in uninterpreted wells based on available interpreted data.

    CO2CRC’s CRC-1 to -7 wells all had an FMI log taken as part of their formation evaluation. Presently, only CRC-1-3 have had an advanced FMI interpretation performed. This project will use this information to train an ML algorithm to automate the interpretation and test/demonstrate this on CRC-4-7 logs.