Constructing sub-scale surrogate model for proppant settling in inclined fractures from simulation data with multi-fidelity neural network

Particle settling in inclined channels is an important phenomenon that occurs during hydraulic fracturing of shale gas production. In order to accurately simulate the large-scale (field-scale) proppant transport process, constructing a fast and accurate sub-scale model, or surrogate becomes critical issue. However, mapping between physical parameters velocity complex, which makes model construction difficult. Previously, particle has usually been investigated via high-fidelity experiments meso-scale numerical simulations, both are time-consuming. this work, we propose new method, i.e., multi-fidelity neural network (MFNN), construct could greatly reduce computational cost while preserving accuracy. The results demonstrate with MFNN can need for data thus by 80%, accuracy lost less than 5% compared surrogate. Moreover, applied macro-scale simulation, shows significant yields results. framework opens novel pathways rapidly predicting reservoir applications. Furthermore, method be extended almost all simulation tasks, especially high-dimensional tasks. • A developed as fractures. Both low-fidelity incorporated training process achieves satisfactory significantly reduces resources. Macro-scale simulations implemented efficiently using fracture inclination angle influence on transport.