Extended search
FORECASTING OF STRESS-STRAIN BEHAVIOR OF ROCK SAMPLES USING RECURRENT NEURAL NETWORKS
1 Schmidt Institute of Physics of the Earth of the Russian Academy of Sciences
2 Moscow Institute of Physics and Technology
Journal: Science and technological developments
Tome: 103
Number: 2
Year: 2024
Pages: 59–74
UDK: 004.032.26 + 550.8.05
DOI: 10.21455/std2024.2-4
Show citation
Egorov
I.V N.A. FORECASTING OF STRESS-STRAIN BEHAVIOR OF ROCK SAMPLES USING RECURRENT NEURAL NETWORKS // . 2024. Т. 103. № 2. С. 59–74. DOI: 10.21455/std2024.2-4
@article{Egorov
I.VFORECASTING2024,
author = "Egorov
I.V, N. A.",
title = "FORECASTING OF STRESS-STRAIN BEHAVIOR OF ROCK SAMPLES USING RECURRENT NEURAL NETWORKS",
journal = "Science and technological developments",
year = 2024,
volume = "103",
number = "2",
pages = "59–74",
doi = "10.21455/std2024.2-4",
language = "English"
}
Copy link
Copy BibTex
Keywords: neural networks, laboratory geomechanical testing, strength limit
Аnnotation: This study explores the application of for predicting the stress-strain behavior of rock samples under multistage testing conditions. Time series data of stress-strain obtained during pseudo-triaxial compression tests are utilized. The recurrent neural network model is trained on historical data to forecast the further deformation of the sample based on its porosity and deformation conditions. The results demonstrate the potential of neural networks in accurately assessing the mechanical properties of rocks when core sample availability is limited
