Abstract
Abstract. The computational costs associated with coupled reactive transport simulations are mostly due to the chemical subsystem: replacing it with a pre-trained statistical surrogate is a promising strategy to achieve decisive speedups at the price of small accuracy losses and thus to extend the scale of problems which can be handled. We introduce a hierarchical coupling scheme in which “full-physics” equation-based geochemical simulations are partially replaced by surrogates. Errors in mass balance resulting from multivariate surrogate predictions effectively assess the accuracy of multivariate regressions at runtime: inaccurate surrogate predictions are rejected and the more expensive equation-based simulations are run instead. Gradient boosting regressors such as XGBoost, not requiring data standardization and being able to handle Tweedie distributions, proved to be a suitable emulator. Finally, we devise a surrogate approach based on geochemical knowledge, which overcomes the issue of robustness when encountering previously unseen data and which can serve as a basis for further development of hybrid physics–AI modelling.
Highlights
Coupled reactive transport simulations (Steefel et al, 2005, 2015) are very expensive, effectively hampering their wide applications
We devise a surrogate approach based on geochemical knowledge, which overcomes the issue of robustness when encountering previously unseen data and which can serve as a basis for further development of hybrid physics–AI modelling
Kinetic control is imposed on all mineral reactions following a Lasaga rate expression from Palandri and Kharaka (2004), which is limited to only neutral and H+ mechanisms and constant reactive surfaces; it is independent of the actual amounts of minerals
Summary
Coupled reactive transport simulations (Steefel et al, 2005, 2015) are very expensive, effectively hampering their wide applications. While hydrodynamic simulations on finely resolved spatial discretizations, containing millions of grid elements, are routinely run on common workstations, the order of magnitude of the computationally affordable reactive transport simulations on the same hardware decreases by a factor of 10 to 100 as soon as chemical reactions are coupled in (De Lucia et al, 2015; Jatnieks et al, 2016; Laloy and Jacques, 2019; Leal et al, 2020; Prasianakis et al, 2020) This usually requires oversimplifications of the subsurface domain, reduced to 2D or very coarse 3D, and of the geochemical complexity as well. Our implementation of coupled reactive transport includes a hierarchical submodel coupling strategy, which is advantageous when different accuracy levels for the predictions of one subprocess are available
Sign-in/Register to view highlights & summary 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.
