Convolutional Surrogate for 3D Discrete Fracture-Matrix Tensor Upscaling

Researchers have developed a convolutional neural network surrogate model to predict equivalent hydraulic conductivity tensors in three-dimensional fractured crystalline media, significantly reducing the computational expense of fine-scale discrete fracture-matrix simulations while maintaining high accuracy. This advancement enables more efficient and practical groundwater flow modeling for developers and engineers working with complex geological systems.

Read the full article at arXiv cs.LG (ML)

---

Want to create content about this topic? Use Nemati AI tools to generate articles, social posts, and more.