I am performing ballistic simulations using NSPH in ABAQUS/Explicit and I frequently encounter the following error:
ABAQUS/explicit rank 0 encountered a SEGMENTATION FAULT
The material behavior is modeled with the Johnson Cook. In certain cases, the stable time increment drastically decreases to the order of 5e-13, while in other cases the analysis proceeds normally before abruptly terminating with the segmentation fault.
Has anyone experienced similar issues when using NSPH with damage models in ABAQUS/Explicit? Any insights into possible causes or strategies to mitigate this problem would be greatly appreciated.
Tell us more about your model. I know this kind of error from models with a user subroutine. Do you use a VUMAT?
Which Abaqus version are you using? Do you use double precision?
I am using a VUMAT implementation with MMC, MJC, CL, and other damage models. Initially, I started with double precision for some parameters, but I have now switched them to single precision. I am working with ABAQUS 2022.
After switching to single precision, I noticed that the analysis runs for a longer time with the same values (e.g., for bulk viscosity), although I sometimes still encounter a similar error. In addition, I observed a drop in total energy during the simulations.
To address this, I increased the linear bulk viscosity, which led to a rise in artificial viscosity. However, when I do not increase the linear bulk viscosity, the results deviate further from the experimental residual velocities.
- Badges
- Science topic
- Similar topics
- Engineering
- Materials Engineering
- Powders