Exploring Optimization Synergies: Neural Networks and Differential Evolution for Rock Shear Velocity Prediction Enhancement

Year 2024, Volume: 6 Issue: 1, 21 - 28, 15.05.2024

Fares Chanane

Abstract

Accurate prediction of rock shear velocity is paramount for various applications, including geothermal energy extraction, CO2 storage, hydrogen storage, and geomechanics. This study introduces an innovative approach to rock shear velocity prediction by integrating neural networks optimized through the differential evolution algorithm. The dataset comprises critical well-logging parameters, including depth, gamma ray, photo-electric factor, neutron porosity, and density. Neural networks are trained to model intricate relationships between these well-logging parameters and rock shear velocity. The application of the differential evolution optimization algorithm, with tuned parameters (population size: 50, crossover probability: 0.8, differential weight: 0.9, and convergence criteria: 0.001), refines neural network parameters. This fine-tuning optimizes the model's ability to capture nuanced variations associated with diverse geological formations, strategically balancing exploration and exploitation within the optimization process. Validation against a comprehensive dataset reveals a notable improvement in rock shear velocity prediction accuracy compared to traditional methods, with an average increase of 15%. Results demonstrate the synergistic effect of specific well-logging parameters and the strategic configuration of differential evolution parameters. A detailed analysis of the differential evolution process highlights how the algorithm explores the solution space, guiding the neural network toward more optimal configurations. The enhanced predictive performance is attributed to the differential evolution algorithm's ability to efficiently search the parameter space, adjusting neural network weights and biases. The population-based approach, governed by the crossover probability and differential weight, facilitates a dynamic exploration of potential solutions. The convergence criteria ensure the algorithm refines the neural network until a satisfactory predictive model is achieved, reducing convergence time by 20%. This research contributes a robust tool to the geophysical community, facilitating precise subsurface structure characterization. The strategic inclusion and optimization of well-logging parameters, coupled with an insightful adjustment of differential evolution parameters, underscore the method's effectiveness in real-world geological contexts. The proposed approach proves valuable for resource exploration, reservoir management, and geological risk assessment, marking a significant advancement in rock shear velocity prediction methodologies.

Keywords

Mchine learning, Fracture porosity, Optimisation, Tigh rocks, Ahnet

References

• Aihar, A., Bouabdallah, N., Ifrene, G., Irofti, D., 2023. Comparing Fishbone Drilling and Hydraulic Fracturing in Ultra-Low Permeability Geothermal Reservoirs.
• Beekman, F., Badsi, M., van Wees, J.D., 2000. Faulting, fracturing and in situ stress prediction in the Ahnet Basin, Algeria—A finite element approach. Tectonophysics, 320(3-4), 311-329.
• Bouabdallah, N., Latrach, A., Aihar, A., Fadairo, A., 2023. Machine Learning Algorithms for Predicting Liquid Loading in Gas Wells.
• Ifrene, G.E., Irofti, D., Khouissat, A., Pothana, P., Aihar, A., Li, B., 2023. Fracture Roughness Characterization from 360 Unrolled Core Images in a Sandstone Reservoir, Case Study, Algeria, Hassi Messaoud. ARMA US Rock Mechanics/Geomechanics Symposium, ARMA-2023-0511.
• Ifrene, G., Irofti, D., Ni, R., Egenhoff, S., Pothana, P., 2023. New Insights into Fracture Porosity Estimations Using Machine Learning and Advanced Logging Tools. Fuels, 4(3), 333–353.
• Irofti, D., Ifrene, G.E., Aihar, A., Bouabdallah, N., Khouissat, A., Djemai, S., 2023. Characterization of a Tight Gas Reservoir Using the Integration of Electrofacies and Fracture Aperture, Ahnet, Algeria. ARMA US Rock Mechanics/Geomechanics Symposium, ARMA-2023.
• Laalam, A., Mouedden, N., Ouadi, H., Chemmakh, A., Merzoug, A., Boualam, A., Djezzar, S., Aihar, A., Berrehal, B., 2022. Prediction of Shear Wave Velocity in the Williston Basin Using Big Data Analysis and Robust Machine Learning Algorithms. ARMA US Rock Mechanics/Geomechanics Symposium, ARMA-2022.
• Aihar, A., Bouabdallah, N., Ifrene, G., Irofti, D., 2023. Comparing Fishbone Drilling and Hydraulic Fracturing in Ultra-Low Permeability Geothermal Reservoirs.
• Akono, A-T., Ulm, F-J., 2012. Fracture scaling relations for scratch tests of axisymmetric shape. Journal of the Mechanics and Physics of Solids 60 (3), 379-390.
• Beekman, F., Badsi, M., van Wees, J-D., 2000. Faulting, fracturing and in situ stress prediction in the Ahnet Basin, Algeria—A finite element approach. Tectonophysics 320 (3-4), 311-329.
• Bouabdallah, N., Latrach, A., Aihar, A., Fadairo, A., 2023. Machine Learning Algorithms for Predicting Liquid Loading in Gas Wells.
• Ifrene, G.E.H., Irofti, D., Khouissat, A., Pothana, P., Aihar, A., Li, B., 2023. Fracture Roughness Characterization from 360 Unrolled Core Images in a Sandstone Reservoir, Case study, Algeria, Hassi Messaoud.
• Ifrene, G., Irofti, D., Ni, R., Egenhoff, S., Pothana, P., 2023. New Insights in Fracture Porosity Estimation using Machine Learning and Advanced Logging Tools.
• Imani, M., Nejati, H.R., Goshtasbi, K., Nazerigivi, A., 2022. Effect of brittleness on the micromechanical damage and failure pattern of rock specimens. Smart Structures and Systems 29 (4), 535-547.
• Irofti, D., Ifrene, G.E.H., Aihar, A., Bouabdallah, N., Khouissat, A., Djemai, S., 2023. Characterization of a Tight Gas Reservoir Using the Integration of Electrofacies and Fracture Aperture, Ahnet, Algeria.
• Irofti, D., Ifrene, GEH., Pu, H., Djemai, S., 2022. A Multiscale Approach to Investigate Hydraulic Attributes of Natural Fracture Networks in Two Tight Sandstone Fields, Ahnet, Algeria. https://doi.org/10.56952/arma-2022-0450.
• Laalam, A., Mouedden, N., Ouadi, H., Chemmakh, A., Merzoug, A., Boualam, A., Djezzar, S., Aihar, A., Berrehal, B.E., 2022. Prediction of Shear Wave Velocity in the Williston Basin Using Big Data Analysis and Robust Machine Learning Algorithms. 56th US Rock Mechanics/Geomechanics Symposium.
• Latrach, A., Bouabdallah, N., Aihar, A., 2023. PS Machine Learning Application for Shear Velocity Prediction in Ahnet Basin, Algeria.
• Logan, P., Duddy, I., 1998. An investigation of the thermal history of the Ahnet and Reggane Basins, Central Algeria, and the consequences for hydrocarbon generation and accumulation. Geological Society, London, Special Publications 132 (1), 131-155.
• Mehrgini, B., Izadi, H., Memarian, H., 2019. Shear wave velocity prediction using Elman artificial neural network. Carbonates and Evaporites, 34, 1281–1291.
• Perron, P., Le Pourhiet, L., Guiraud, M., Vennin, E., Moretti, I., Portier, É., Konaté, M. 2021. Control of inherited accreted lithospheric heterogeneity on the architecture and the low, long-term subsidence rate of intracratonic basins. Bulletin de La Société Géologique de France, 192 (1).
• Pothana, P., Garcia, F.E., Ling, K., 2024a. Effective Elastic Properties and Micro-mechanical Damage Evolution of Composite Granular Rocks: Insights from Particulate Discrete Element Modelling. Rock Mechanics and Rock Engineering 1-45.
• Pothana, P., Garcia, F. E., & Ling, K., 2024b. Understanding Effective Elastic Properties and Stress-Strain Evolution in Composite Granular Rocks Using Discrete Element Modeling.
• Pothana, P., Ifrene, G., Ling, K., 2023. Stress-Dependent Petrophysical Properties of the Bakken Unconventional Petroleum System: Insights from Elastic Wave Velocities and Permeability Measurements. Fuels 4 (4), 397-416.
• Pothana, P., Ifrene, G., Ling, K., 2024. Integrated Petrophysical Evaluation and Rock Physics Modeling of Broom Creek Deep Saline Aquifer for Geological CO2 Storage. Fuels, 5(1), 53–74.
• Singh, S., Kanli, A. I. 2016. Estimating shear wave velocities in oil fields: a neural network approach. Geosciences Journal 20, 221-228.
• Wang, J., Cao, J., Yuan, S., 2020. Shear wave velocity prediction based on adaptive particle swarm optimization optimized recurrent neural network. Journal of Petroleum Science and Engineering 194, 107466.
• Zazoun, R.S., 2001. La tectogenese hercynienne dans la partie occidentale du bassin de l’Ahnet et la region de Bled El-Mass, Sahara Algerien: un continuum de deformation. Journal of African Earth Sciences 32 (4), 869–887.