Characterization of pores and microfractures in tight conglomerate reservoirs

Wang, Ziqiang; Ge, Hongkui; Zhou, Weili; Wei, Yun; Wang, Bei; Liu, Sai; Zhang, Hao; Du, Shanshan

doi:10.1016/j.ijhydene.2022.06.037

Cited by 7 publications

(20 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Finally, most related studies lack robust supervised ML methodologies (Abedini et al, 2018;Borazjani et al, 2016;Ghiasi-Freez et al, 2012;Mollajan et al, 2016;Sharifi, 2022;Z. Wang et al, 2022).…”

Section: Study Design Issues In Related Studiesmentioning

confidence: 99%

“…It is evident that studies claiming excellent performance of supervised ML for pore typing have not fully considered the true complexity of the task and instead report the results of highly curated data sets (Abedini et al, 2018;Borazjani et al, 2016;Ghiasi-Freez et al, 2012;Mollajan et al, 2016;Z. Wang et al, 2022).…”

Section: Biased Samplingmentioning

confidence: 99%

“…Wang et al, 2022), thereby obfuscating the underpinning drivers of their reported excellent model accuracies. Almost all related studies do not furnish details on hyperparameter tuning, perhaps as the default parameters produce excellent results (Abedini et al, 2018;Borazjani et al, 2016;Ghiasi-Freez et al, 2012;Mollajan et al, 2016;Z. Wang et al, 2022).…”

Section: Study Design Issues In Related Studiesmentioning

confidence: 99%

“…Only two studies have employed the binary approach within macrofractures in micro-CT models (Li et al, 2017;Singh et al, 2021), and one in the case of microfractures (Z. Wang et al, 2022). Li et al (2017) utilized an SVM to separate macrofractures from vugs using simple shape features, reporting an accuracy of 100%.…”

Section: Introductionmentioning

confidence: 99%

“…Supervised ML models are particularly sensitive to the nature of the labeled data. Most related studies are opaque on their sampling protocols, which raises questions as to whether the data was properly curated (Table S1 in Supporting Information S1) (e.g., Abedini et al, 2018;Borazjani et al, 2016;Ghiasi-Freez et al, 2012;Mollajan et al, 2016;Sharifi, 2022;Z. Wang et al, 2022).…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

An Object‐Based Approach to Differentiate Pores and Microfractures in Petrographic Analysis Using Explainable, Supervised Machine Learning

Jayachandran,

Gibbs,

Laya

et al. 2024

Earth and Space Science

View full text Add to dashboard Cite

Petrographic observations are vital for carbonate pore‐typing, linking geological frameworks to petrophysical behavior. However, current petrographic pore typing is manual, with the qualitative to semi‐quantitative results not easily fitted into quantitative subsurface characterization. Some recent studies have automated this process using supervised machine learning (ML) and deep learning (DL), focusing on simple pore morphological features, and have reported high classification accuracies for several complex pore types. However, there are concerns about the validity of these studies due to conceptual and technical flaws in their collective approach. This study was aimed at a more fundamental problem, classifying between open microfractures and open pores in petrographic thin sections using an object‐based approach and explainable supervised ML. We analyzed 18 carbonate thin sections from the USA, numerically representing them using five shape features: compactness, aspect ratio, extent, solidity, and formfactor. Using a labeled data set of 400 microfractures and 400 pores, we evaluated nine of the most widely used supervised models. All models showed high testing accuracies (89.58%–90.42%). Interestingly, complex non‐linear models did not significantly outperform simpler linear ones. Compactness and aspect ratio were the most informative features. However, the labeled data sets did not reflect the overall data set's complexity, which suggested that high accuracies in similar studies might be due to curated data sets rather than accounting for the true complexity of carbonate pore systems. The study concludes that simple shape features are ineffective for classifying carbonate pore types. It is hoped that this study will provide a foundation for more robust artificial intelligence‐assisted pore typing.

show abstract