Application Model: Proppant Transport in a Single Hydraulic Fracture Using MP-PIC CFD and a Graph Neural Network – Reduced Order Model (GNN-ROM)
This application model uses Barracuda Virtual Reactor to simulate proppant transport and bed formation within a single hydraulic fracture using the Multiphase Particle-in-Cell (MP-PIC) method. A non-Newtonian drag correction accounts for the shear-thinning behavior of the slickwater carrying the proppant, and the resulting settling and bed formation behavior are validated against the single-fracture experimental measurements of Zhou et al. (2023). The simulation captures the heel-skewed bed geometry and particle size segregation that govern how much of the fracture remains conductive after treatment. This validated CFD solution also serves as the benchmark for a Graph Neural Network reduced-order model (GNN-ROM), demonstrating a pathway toward fast, physics-based proppant transport prediction at the scale of full multi-fracture completions.



