I am currently working on a seepage analysis using the Coupled Eulerian-Lagrangian (CEL) approach in Abaqus Explicit. As you are aware, Abaqus does not provide built-in capabilities for seepage flow within the CEL framework. To overcome this, I am employing a thermal-stress analogy, where heat flow is used to represent hydraulic head gradients—effectively simulating seepage flow from higher to lower head regions.
The approach works well under the assumption of homogeneous thermal conductivity, where the head distribution (analogous to temperature) remains smooth and predictable. However, my current objective is to analyze a system with spatially varying thermal conductivity, representing different soil permeability zones.
In doing so, I am observing irregularities in the head (temperature) distribution, and the results do not appear physically consistent. I suspect this might be due to limitations in how material property variation is handled or due to improper boundary representation under variable conductivity.
I would appreciate your insights on the following:
- What would be the best way to accurately simulate seepage using this thermal analogy under non-uniform conductivity?
- Would the use of a user subroutine (e.g., VUMATHT, VUMAT or VDFLUX) be necessary or beneficial to define the thermal properties or fluxes dynamically?
- Are there alternative methods within the CEL or Explicit framework to better represent seepage without relying solely on the thermal analogy?
Your guidance on this would be immensely helpful as I move forward with the model development.
