Abstract
Electrokinetic (EK) remediation is one of the few in-situ technologies that can effectively remove contaminants from low-permeability porous media. Process-based modeling, including the complex multiphysics and biogeochemical processes occurring during electrokinetic remediation, is instrumental to describe EK systems and to assist in their design. Multidimensional process-based models require long computational times, which limit their use for optimization and prediction of system performances over wide ranges of conditions. To overcome such limitation, in this work we develop an approach integrating a process-based model and a surrogate model based on machine learning. The approach provides fast predication of system performances and is therefore highly suitable for decision-making processes involving uncertainty analysis, sensitivity analysis and feasibility studies.
Original language | English |
---|---|
Publication date | 2021 |
Number of pages | 3 |
Publication status | Published - 2021 |
Event | IWA Digital World Water Congress 2021 - Online Duration: 24 May 2021 → 4 Jun 2021 |
Conference
Conference | IWA Digital World Water Congress 2021 |
---|---|
Location | Online |
Period | 24/05/2021 → 04/06/2021 |
Keywords
- groundwater remediation
- machine learning
- reactive transport modeling