Abstract
ABSTRACTThe seepage equation plays a crucial role in fields such as groundwater management, petroleum engineering, and civil engineering. Currently, physical‐informed neural networks (PINNs) have become an effective tool for solving seepage equations. However, practical applications often involve variable flow rates, which pose significant challenges for using neural networks to find solutions. Inspired by Deep Operator Network (DeepONet), this paper proposes a new model named Simulation Net (Sim‐net) to deal with unsteady sources or sinks problems. Sim‐net is designed to simulate and solve seepage equations without the need for retraining. This model integrates potential spatial and temporal features based on spatial pressure distribution and well bottom–hole pressure, respectively, which serve as additional signposts to guide neural networks in approximating seepage equations. Sim‐net exhibits transfer learning capabilities, enabling it to handle variable flow rate problems without retraining for new flow conditions. Numerical experiments demonstrate that the trained model can directly solve seepage equations without the need for retraining, indicating its superior applicability compared to existing PINNs‐based methods. Additionally, in comparison to the DeepONet, Sim‐net achieves higher accuracy.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: International Journal for Numerical Methods in Fluids
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
