1887

Abstract

Summary

The study focuses on the characterization of source rocks in the Vaca Muerta formation in Argentina, utilizing portable X-ray fluorescence (XRF) data and compact neural networks to predict total organic carbon (TOC) content. Annually, between 300 and 500 meters of core samples are analyzed, generating 3000 to 5000 portable XRF measurements and approximately 300 TOC measurements using Rock Eval 6. TOC is a crucial parameter in the exploration and characterization of unconventional reservoirs.

The proposed method combines carefully curated geochemical data and expert knowledge to train a neural network with a dataset of 1081 samples. Each sample includes specific REDOX elements from the XRF as inputs and the measured TOC obtained through pyrolysis as the output. The training process involves several stages, including dataset splitting, normalization, and defining the neural network architecture.

Results indicate that it is possible to predict TOC with an accuracy of 88%, allowing for the generation of continuous, high-density TOC profiles. This methodology accelerates data acquisition and provides a significant density of data to improve models and support decision-making

© EAGE Publications BV
Loading

Article metrics loading...

/content/papers/10.3997/2214-4609.202539056
2025-03-24
2025-12-08

  • From This Site

    /content/papers/10.3997/2214-4609.202539056
    dcterms_title,dcterms_subject,pub_keyword
    -contentType:Journal -contentType:Contributor -contentType:Concept -contentType:Institution
    10
    5
Loading full text...

Full text loading...

References

  1. Algeo, T.J., Maynard, J.B., [2008]. Trace-metal covariation as a guide to water-mass conditions in ancient anoxic marine environments. Geosphere4: 872–887.
     [Google Scholar]
  2. Helz GR, CV Miller, JM Charnock, JL Mosselmans, RA, Pattrick, CD Garner y DJVaughan, [1996]. Mechanisms of molybdenum removal from the sea and its concentration in black shales: EXAFS evidences. Geochimica et Cosmochimica Acta60, p.3631–3642
     [Google Scholar]
  3. Larriestra, C.y Larriestra, V., [2013]. La construction de nuevos paradigmas en recursos no convencionales (black shale), Rev. Petrotecnia, IAPG, Oct. 2013, p.70–79.
     [Google Scholar]
  4. Marakushev, A., Marakushev, S., [2006]. Nature of Specific Geochemical Features of Oil. Doklady Akademii Nauk, 2006, Vol. 411, 1, p. 111–117
     [Google Scholar]
  5. Nawratil, A., Gomez, H. y Larriestra, C, [2012]. Key Tools for Black Shale Evaluation: Geostatistics and Inorganic Geochemistry Applied to Vaca Muerta Formation, Neuquen Basin, Argentina, AAPG, Search and Discovery Article # 41028.
     [Google Scholar]
  6. Norrish, K., Hutton, J.T., [1969]. An accurate x-ray spectrographic method for the analysis of a wide range of geological samples. Geochimica et Cosmochimica Acta33, 431–453
     [Google Scholar]
  7. Potter, P.Maynard B. y DepetrisP., [2005]. Mud and mudstones. Introduction and overview. Springer Verlag297 p.
     [Google Scholar]
  8. Schiuma, M. y Larriestra, C., [2011]. Cuenca del Golfo de San Jorge Log geoquimico de pozo: una nueva vision de viejos problemas, VIII Congreso Exploratión y Desarrollo de Hidrocarburos, 8–12 Nov., Mar del Plata, Argentina
     [Google Scholar]
  9. TribovillardN, ARiboulleau, TLyons y FBaudin, [2004]. Enhanced trapping of molybdenum by sulfurized organic matter of marine origin as recorded by various Mesozoic formations. Chemical Geology213, p. 385–401.
     [Google Scholar]
  10. TensorFlow, G. [2015]. Large-scale machine learning on heterogeneous systems. Google Research, 10, s15326985ep4001.
     [Google Scholar]
  11. Chollet, F. [2021]. Deep learning with Python. Simon and Schuster.
     [Google Scholar]
/content/papers/10.3997/2214-4609.202539056
Loading
Accurate TOC Modeling in Vaca Muerta formation with Compact Neural Networks and Portable XRF Data
Conference Proceedings 2025, 1 (2025); https://doi.org/10.3997/2214-4609.202539056
/content/papers/10.3997/2214-4609.202539056
/content/papers/10.3997/2214-4609.202539056
Loading

Data & Media loading...

Most Cited This Month Most Cited RSS feed

This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error