Abstract
AbstractRecent advances in machine learning open new opportunities to gain deeper insight into hydrological systems, where some relevant system quantities remain difficult to measure. We use deep learning methods trained on numerical simulations of the physical processes to explore the possibilities of closing the information gap of missing system quantities. As an illustrative example we study the estimation of velocity fields in numerical and laboratory experiments of density‐driven solute transport. Using high‐resolution observations of the solute concentration distribution, we demonstrate the capability of the method to structurally incorporate the representation of the physical processes. Velocity field estimation for synthetic data for both variable and uniform concentration boundary conditions showed equal results. This capability is remarkable because only the latter was employed for training the network. Applying the method to measured concentration distributions of density‐driven solute transport in a Hele‐Shaw cell makes the velocity field assessable in the experiment. This assessability of the velocity field even holds for regions with negligible solute concentration between the density fingers, where the velocity field is otherwise inaccessible.
Highlights
Gaining a quantitative understanding of hydrological systems is difficult and relies on the availability of accurate measurements that are dense in space and time
We focus on a comparably simple problem as an example: velocity field estimation on density-driven active solute transport observed in a small-scale laboratory experiment within a
Adopted the network architecture of Zhu and Zabaras (2018) and extended the model as a surrogate for uncertainty quantification of transient multiphase flow in heterogeneous media. Whereas these studies focus on replacing the forward models of related physical systems using an encoder-decoder convolutional neural networks (CNNs), in our work we use similar deep learning methods to aim at the estimation of missing system quantities
Summary
A convolutional neural network is trained on density-driven solute transport simulations to incorporate the physical process representation. Missing velocity fields on experimental data are estimated utilizing the physical process representation learned on purely synthetic data. Light transmission measurements using a Hele-Shaw cell are conducted to infer spatially resolved solute concentration distributions. Velocity field estimation on density-driven solute transport with a convolutional neural network.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
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 callDisclaimer: 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.
