Hello everyone,

I am simulating a geotechnical slope using Abaqus/Explicit CEL. In my model, Zone 2 (red marked region in the figure) represents the toe of the slope. My analysis has three steps:

1. Step 1 (Geostatic / initial stress build-up): the slope is stable.
2. Step 2 (Toe erosion): I weaken the toe region (Zone 2).
3. Step 3 (Post-erosion response): I expect instability / deformation triggered by toe weakening.

How I implemented toe weakening

Because this is CEL, I cannot easily isolate toe vs non-toe regions the way I can in a purely Lagrangian mesh. So I am controlling material properties using a field variable updated in a VUSDFLD subroutine. My logic is:

• In Step 1: FIELD(1) = 1 (intact properties)
• In Step 2: FIELD(1) = 2 (reduced strength properties for toe erosion)

Then in the material definition, I provide Mohr–Coulomb properties with dependencies, so that when FIELD(1)=2, the toe strength parameters are reduced.

Issue

Even after Step 2 (toe erosion), I do not observe any significant deformation or failure mechanism, and Step 3 also shows no clear “toe erosion effect.”

However, I noticed two numerical symptoms right after Step 2 starts:

• Kinetic Energy (KE) becomes zero (or drops to near zero) after Step 2
• Total energy becomes negative

This makes me suspect that either:

• the model is reaching a static equilibrium and no instability is being triggered, or
• the sudden property drop is causing a numerical/energy issue and the response is suppressed.

My current hypothesis

I think I may need to reduce toe strength gradually instead of switching from intact to residual instantly. But I am unsure how to do this correctly in CEL, because I also have other state variables controlling behavior (e.g., plastic strain, mean stress, etc.), and I do not want to interfere with those.

Questions

1. In Abaqus/Explicit CEL, why would kinetic energy go to zero immediately after a strength reduction step?
2. What does it indicate when total energy becomes negative in this context?
3. What is the best way to implement gradual toe weakening in CEL?
   • Should I ramp a field variable using time (e.g., FIELD=1 → 2 gradually)?
   • Or should I apply an equivalent “erosion” through another method (e.g., material removal, density reduction, body-force change, etc.)?
4. If gradual weakening is recommended, how can I do it robustly in VUSDFLD, especially when other SDVs are used for strain-softening?

Any guidance on how to diagnose this (energy balance, hourglass, stable time increment, damping, etc.) and how to implement gradual toe erosion would be greatly appreciated.

Also I got this warning message during running in toe erosion step, “***WARNING:  Step 3, Increment    57:    1 material point(s) failed to

            converge in the material constitutive routines.

Thank you.