Accurate numerical simulation of soil heat transfer process using self-correcting dynamic grid method

Soil Vapor Extraction is an in-situ thermal remediation technology that heats contaminated soil through buried pipes. Accurate temperature simulation helps predict the remediation status of a site in advance, enabling cost estimation and remediation strategy design. However, fixed-grid methods struggle to maintain high spatial resolution, and existing dynamic mesh-related studies are mostly limited to refinement due to numerical stability issues, leading to a sharp increase in simulation costs as the process continues. To address these problems, this study proposes a self-correcting dynamic grid technique, which employs Lagrange interpolation to automatically adjust the mesh after splitting and before merging, thereby maintaining high spatial resolution while ensuring numerical stability. A comparative model was established using COMSOL, and validation was conducted with measured data from an actual remediation site. The results show that, compared to the COMSOL simulation, the proposed model reduces mesh data storage by 90 % while achieving high accuracy, with an error below 20 % between the simulated and measured cold-point temperature increments (85 % lower than COMSOL under 62 °C). The relevant research findings can provide guidance for the design, evaluation, and operation stages of soil remediation engineering.