Dynamic Coupling Modeling of Rock Electrical-Physical Properties Under Temperature-Pressure Constraints for Reservoir Parameter Prediction
Abstract
With the continuous deepening of exploration and development, permeability is a key parameter for evaluating unconventional reservoirs. Unconventional reservoirs have low porosity and low permeability characteristics, and the strong nonlinear characteristics of seepage
flow affect the validity of conventional permeability model relationships. Single theoretical models have limitations in characterizing permeability for complex lithology. Therefore, comprehensive algorithms are used to extract sensitive parameters and their combinations related to
permeability, so as to achieve intelligent permeability classification and estimation based on multiple physical properties parameters. Classical
machine learning models such as K-Nearest Neighbors (KNN), Random Forest, and Support Vector Machine (SVM) can all be used to predict
permeability.
??In this study, the Random Forest model, a simple Bagging integration algorithm, is employed to describe the nonlinear effects between
parameters such as elasticity, induced polarization, and permeability. This approach allows us to establish a permeability classification and
prediction model based on multiple physical properties parameters. In this study, the potential relationships between multiple physical properties parameters and permeability were investigated, and the importance of features was analyzed. The adaptability of intelligent algorithms in
prediction and classification effects was also introduced. By combining algorithms, the objective of intelligent classification and estimation of
permeability was achieved. In the prediction study of permeability, after eliminating outliers and adjusting model parameters, the prediction
loss MSE value of the model was 0.0925, and the classification score of the classification model was 0.879. These results verified the reliability of the classification method and permeability prediction model, providing guidance for the subsequent reservoir productivity evaluation in
the study area.
flow affect the validity of conventional permeability model relationships. Single theoretical models have limitations in characterizing permeability for complex lithology. Therefore, comprehensive algorithms are used to extract sensitive parameters and their combinations related to
permeability, so as to achieve intelligent permeability classification and estimation based on multiple physical properties parameters. Classical
machine learning models such as K-Nearest Neighbors (KNN), Random Forest, and Support Vector Machine (SVM) can all be used to predict
permeability.
??In this study, the Random Forest model, a simple Bagging integration algorithm, is employed to describe the nonlinear effects between
parameters such as elasticity, induced polarization, and permeability. This approach allows us to establish a permeability classification and
prediction model based on multiple physical properties parameters. In this study, the potential relationships between multiple physical properties parameters and permeability were investigated, and the importance of features was analyzed. The adaptability of intelligent algorithms in
prediction and classification effects was also introduced. By combining algorithms, the objective of intelligent classification and estimation of
permeability was achieved. In the prediction study of permeability, after eliminating outliers and adjusting model parameters, the prediction
loss MSE value of the model was 0.0925, and the classification score of the classification model was 0.879. These results verified the reliability of the classification method and permeability prediction model, providing guidance for the subsequent reservoir productivity evaluation in
the study area.
Keywords
Permeability; Random Forest; Porosity
Full Text:
PDFReferences
[1] Zhu, Y., & Toksz, M. N. (2012). Experimental study of the frequency-dependent complex resistivity of reservoir rocks under high pressure. Geophysics, 77(4), D163-D172.
[2] Revil, A., & Glover, P. W. J. (1997). Theory of ionic-surface electrical conduction in porous media. Physical Review B, 55(3), 1757-1773.
[3] Archie, G. E. (1942). The electrical resistivity log as an aid in determining some reservoir characteristics. Transactions of the AIME,
146(1), 54-62.
DOI: http://dx.doi.org/10.70711/itr.v2i4.7697
Refbacks
- There are currently no refbacks.
