Analysis on Correlation Model Between Fracture Network Complexity and Gas-Well Production: A Case in the Y214 Block of Changning, China DOI Creative Commons
Zhibin Gu, Bingxiao Liu, Wang Liu

et al.

Energies, Journal Year: 2024, Volume and Issue: 17(23), P. 6026 - 6026

Published: Nov. 29, 2024

The fracture network of the Y214 block in Changning area China is complex, and there are significant differences productivity different shale gas wells. However, traditional machine learning models have problems such as missing key parameters, poor fitting effects low prediction accuracy, which make it difficult to effectively evaluate impact crack complexity on productivity. Therefore, Pearson correlation coefficient was used analyze between evaluation mineral content, horizontal stress difference, natural fractures production. Combined with improved particle swarm optimization (IPSO) algorithm support vector (SVM) algorithm, a index (FNI) model proposed networks, verified by comparing performance results from other two models. Finally, actual average daily production fracturing sections calculated analyzed. showed that density factor controlling (the 0.39), factors weak. In process data, determination, R², IPSO-SVM-FNI training set increased 8% 24% compared models, effect greatly improved. based R² test 22% 20% accuracy also significantly concentrated, its main distribution range [0.2, 0.8]. section higher FNI production, positive Indeed, research provide some ideas references for reservoirs.

Language: Английский

Tunable discrete fracture network for dynamic analyses of rock landslides by material point method DOI
Jingsong Yan, Yawen Wu,

Qirui Gao

et al.

Computers and Geotechnics, Journal Year: 2025, Volume and Issue: 182, P. 107154 - 107154

Published: Feb. 19, 2025

Language: Английский

Citations

0
Denoising Diffusion Probabilistic Model-Based Multivariate Parameter Distributions for Rough Discrete Fracture Network Modeling DOI

Shuyang Han,

Jiajun Wang,

Dawei Tong

et al.

Research Square (Research Square), Journal Year: 2025, Volume and Issue: unknown

Published: April 1, 2025

Abstract Fractures significantly influence rock mass geotechnical behavior, necessitating precise characterization of their geometric parameters. Traditional modeling approaches, based on standard statistical descriptions and random simulations, often disregard parameter correlations assume smooth fractures, compromising accuracy. This study introduces a Denoising Diffusion Probabilistic Model (DDPM) to capture dip direction, angle, trace length, aperture, roughness generate discrete fracture network (DFN) data. By integrating fractal dimensions non-uniform rational B-splines (NURBS) tensor products, our approach accommodates roughness, enhancing overall realism. Validation real-world datasets using Kullback–Leibler(KL) divergence Wasserstein distance indicates that DDPM outperforms generative adversarial networks (GAN), variational autoencoders (VAE), normalizing flow (NF), Monte Carlo methods, achieving average KL/Wasserstein reductions 72.44%/57.08% against other models 74.84%/36.83% Carlo. Furthermore, the modeled rough fractures accurately match real traces, confirming improved fidelity DFN simulations.

Language: Английский

Citations

0
Transient free surface flow through 3D fracture networks: PVI approach and geological-entropy-based exploration on spatial disorder DOI
C. B. Li, Jianlong Sheng, Zuyang Ye

et al.

Computers and Geotechnics, Journal Year: 2025, Volume and Issue: 184, P. 107300 - 107300

Published: April 28, 2025

Language: Английский

Citations

0
An equivalent fracture length-based numerical method for modeling nonlinear flow in 2D fracture networks DOI
Jie Liu, Zhechao Wang, Liping Qiao

et al.

Computers and Geotechnics, Journal Year: 2024, Volume and Issue: 176, P. 106753 - 106753

Published: Sept. 19, 2024

Language: Английский

Citations

0
Analysis on Correlation Model Between Fracture Network Complexity and Gas-Well Production: A Case in the Y214 Block of Changning, China DOI Creative Commons
Zhibin Gu, Bingxiao Liu, Wang Liu

et al.

Energies, Journal Year: 2024, Volume and Issue: 17(23), P. 6026 - 6026

Published: Nov. 29, 2024

The fracture network of the Y214 block in Changning area China is complex, and there are significant differences productivity different shale gas wells. However, traditional machine learning models have problems such as missing key parameters, poor fitting effects low prediction accuracy, which make it difficult to effectively evaluate impact crack complexity on productivity. Therefore, Pearson correlation coefficient was used analyze between evaluation mineral content, horizontal stress difference, natural fractures production. Combined with improved particle swarm optimization (IPSO) algorithm support vector (SVM) algorithm, a index (FNI) model proposed networks, verified by comparing performance results from other two models. Finally, actual average daily production fracturing sections calculated analyzed. showed that density factor controlling (the 0.39), factors weak. In process data, determination, R², IPSO-SVM-FNI training set increased 8% 24% compared models, effect greatly improved. based R² test 22% 20% accuracy also significantly concentrated, its main distribution range [0.2, 0.8]. section higher FNI production, positive Indeed, research provide some ideas references for reservoirs.

Language: Английский

Citations

0