Prediction of shear wave velocity using machine learning models and characteristics of the porous medium
Author Identifier (ORCID)
Mohsen Farrokhrouz: https://orcid.org/0000-0001-5169-6894
Alireza Keshavarz: https://orcid.org/0000-0002-8091-961X
Abstract
Shear wave (S-wave) velocity is an important parameter of rocks, as it is used for the evaluation of the morphology and saturation of the porous medium. Estimation of missing values of S-wave has been a focal point of interest in the energy sector for a long time. For this study, we collected a comprehensive set of data (about 1,200 data points) from different rock types and different machine learning (ML) models to find the best fit for the prediction of the S-wave. The ML models used were linear regression (LR), decision tree (DT), random forest (RF), extreme gradient boosting (XGBoost), and multilayer perceptron (MLP). Our results confirmed that considering the conceptual effect of porosity as well as compressional wave (P-wave) within the model gives a very good prediction of the S-wave with R2 = 0.996. The proposed model is independent of medium saturation, confining stress, and fluid type and depends on the type of the porous medium (e.g., rock type). The linear form of the ML model is quite simple and can be easily used, while the MLP model gives less error in estimation.
Document Type
Journal Article
Date of Publication
8-1-2025
Volume
30
Issue
8
Publication Title
SPE Journal
Publisher
Society of Petroleum Engineers
School
School of Engineering
Copyright
subscription content
First Page
4530
Last Page
4544
Comments
Farrokhrouz, M., Mohammadi, A., & Keshavarz, A. (2025). Prediction of shear wave velocity using machine learning models and characteristics of the porous medium. SPE Journal, 30(8), 4530–4544. https://doi.org/10.2118/228292-PA