A method for enhancing well-log resolution of thin lithological heterogeneities using wavelet transform and automated machine learning

Aug 4, 2026

Clastic reservoirs contain complex lithologies, and thin low-permeability heterogeneities can strongly affect CO2 flooding strategies and oil recovery. These layers often produce weak well-log responses, making recognition difficult with conventional interpretation methods. A hierarchical workflow is proposed that combines discrete wavelet transform to enhance responses with an automated machine-learning framework for multi-algorithm parameter optimization and nonlinear mapping between lithology and log signals. Dividing recognition into three levels highlights lithological contrasts and improves differentiation accuracy.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages13

Reading Time33 min

AccessFree to download

Introduction

Permeability barriers in clastic reservoirs

Depositional-origin thin muddy/silty layers

Diagenetic permeability barriers (cementation)

Advances in OrigiGeo-Enal articlenergy Research Vol. 17, No. 2, p. 149-161, 2025 A method for enhancing well-log resolution of thin lithological heterogeneities using wavelet transform and automated machine learning Wei Li 1 ,2 , Lei Liu 1 ,2 , Dali Yue 1 ,2*, Jian Gao3 , Shanyan Zhang3 , Luca Colombera4 ,5 1 State Key Laboratory of Petroleum Resources and Engineering, China University of Petroleum, Beijing 102249, P. R. China 2 College of Geosciences, China University of Petroleum, Beijing 102249, P. R. China 3 Research Institute of Petroleum Exploration & Development, PetroChina, Beijing 100083, P. R. China 4 Department of Earth and Environmental Sciences, University of Pavia, Pavia 27100, Italy 5 School of Earth and Environment, University of Leeds, Leeds LS2 9JT, United Kingdom | Keywords:
Well-log interpretation permeability barrier clastic reservoir thin lithology flow baffle
wavelet transform
Cited as:
Li, W., Liu, L., Yue, D., Gao, J., Zhang, S., Colombera, L. A method for enhancing well-log resolution of thin lithological heterogeneities using wavelet transform and automated machine learning. Advances in Geo-Energy Research, 2025, 17(2): 149-161 .
[https://doi.org/10.46690/ager.2025.08.06](https://doi.org/10.46690/ager.2025.08.06) | Abstract:
Clastic reservoirs exhibit complex and diverse lithologies. Some lithological heterogeneities, occurring as thin but effectively low-permeability units, have pronounced impact on CO2 flooding schemes and oil recovery. Thin low-permeability units within permeable sandbodies typically exhibit weak well-log responses, and are therefore of difficult recognition using conventional well-log analysis methods. To address this challenge, a hierarchical method is proposed for interpreting thin lithological heterogeneities by integrating wavelet transform and machine learning. The discrete wavelet transform enhanceswell-log responses of thin heterogeneities. An automated machine-learning framework is designed, which integrates multiple algorithms and achieves automated parameter optimization. This machine-learning method is then applied to well logs to establish anonlinear mapping model between lithology and well-log responses. Additionally, the hierarchical nature of the workflow highlights lithological contrasts, facilitating a more accurate lithological differentiation by dividing the recognition of thin heterogeneities into three levels. Benefiting from these three advantages, the proposed method offers potential to significantly enhance the accuracy of well-log interpretations. The results demonstrate that this method yields accurate identification of lithological units as thin as 0.2 m for muddy beds and 0.3 m for diagenetic units, achieving a recognition accuracy exceeding the conventional well-log interpretations. This method also shows significant potential for broader applications, including the identification of other types of geological entities of limited thickness, and determination of reservoir parameters at fine scales. | | --- | --- | 1. Introduction Clastic reservoirs exhibit complex internal architectures, characterized by various sedimentary and diagenetic features that influence their porosity, permeability, and overall effectiveness as hydrocarbon reservoirs (Stanistreet and Stoll- hofen, 2002 ; Yue et al., 2018 ; Nyberg et al., 2023) . Among these features, the presence of permeability barriers within sandbodies is significant as a control on reservoir behavior (Li et al., 2011), with a particularly pronounced impact on CO2 flooding schemes and oil recovery (Wang et al., 2011) . These permeability barriers can be broadly categorized into ∗Corresponding author. E-mail [address](address: wei_li@cup.edu.cn)[: wei_li@cup.edu.cn](address: wei_li@cup.edu.cn) (W. Li); [liulei@student.cup.edu.cn](liulei@student.cup.edu.cn) (L. Liu); [yuedali@cup.edu.cn](yuedali@cup.edu.cn) (D. Yue); [gj940116@petrochina.com.cn](gj940116@petrochina

FAQs

Why are thin lithological heterogeneities hard to identify using conventional well-log analysis?

Thin low-permeability units within otherwise permeable sandbodies typically generate weak well-log responses, reducing contrast in standard interpretation workflows.

How does the proposed method improve well-log resolution for thin heterogeneities?

A discrete wavelet transform enhances the well-log responses of thin heterogeneities before machine learning maps log signals to lithology.

What is the role of the automated machine-learning framework in the workflow?

The framework integrates multiple algorithms and performs automated parameter optimization, enabling an effective nonlinear mapping model between lithology and enhanced log responses.

Aug 4, 2026

Upload by Seraphina