I am attempting to create a model of speed core walls (concrete-filled panels) subjected to fire loads in abaqus, using two analyses. The first analysis involves heat transfer to obtain temperatures at the nodes, while the second analysis is a general static analysis where I apply an axial load and then incorporate the temperature. The issue I'm facing is non-convergence due to plate buckling and excessive deformations, especially at the beginning of the process. Any ideas on what I might be doing wrong? I've attached the input file for reference.
Engr. Tufail I got the materials curve from the Euro code, it's all in Pa, can you tell me if it's the concrete or the steel that's wrong?
Jose Fernandez, I have addressed a similar question before, although I don't recall the specific instance. If you're encountering issues like those described above, it's likely due to an incorrect material definition being employed. To resolve this, I recommend removing any temperature-dependent data from your analysis and systematically examining each parameter to identify which one has the most destabilizing impact on the simulation. In other words, when encountering problems in your FEA simulation as mentioned earlier, it's crucial to review and possibly refine your material definitions. Start by eliminating temperature-dependent properties from your analysis. Then, carefully analyze each parameter, one by one, to pinpoint which one is causing the most significant destabilization in your simulation results. This systematic approach will help you identify and rectify the root cause of the issues you're experiencing.
Engr. Tufail I tested all the properties separately; there was a slight error in thermal expansion, but it was not relevant to the model's convergence. My ultimate idea is to create a wall of steel plates filled with concrete and analyze the fire load from one side only. To simplify my problem and ensure that my entire model is working correctly, I started with a column with fire on one side. When testing the model, the problem arises due to the interaction. I need the plates to be able to separate as the simulation progresses, due to possible buckling and also because the steel expands faster than concrete. When I run a hard contact simulation, it doesn't converge, unlike when I use a tied constraint interaction, but this affects my results.
Do you have any ideas on how to improve convergence when elements separate?
Convergence issues in a sequentially coupled thermal model in Abaqus can arise due to various reasons, and addressing these issues typically involves a combination of troubleshooting techniques. Sequential coupling means that the thermal analysis is performed first, and then the results are used to update the structural analysis, or vice versa. Here are some common steps to troubleshoot convergence problems in Abaqus:
1. **Reducing Time Steps**: Try reducing the time step size for both thermal and structural analyses. Smaller time steps can help stabilize the solution and improve convergence. You can do this by modifying the step size and incrementation options in your analysis settings.
2. **Change Element Types**: Depending on your model and loading conditions, the choice of element types can affect convergence. Try using different element types for both thermal and structural analyses. Sometimes, using more specialized elements can improve convergence.
3. **Check Material Properties**: Ensure that your material properties (e.g., thermal conductivity, specific heat, density) are defined correctly and consistently between the thermal and structural analyses. Inconsistencies in material properties can lead to convergence issues.
4. **Initial Conditions**: Check the initial conditions for both the thermal and structural analyses. Proper initial conditions are essential for convergence. Make sure that temperatures and displacements are initialized correctly.
5. **Boundary Conditions**: Verify that your boundary conditions are correctly applied in both the thermal and structural analyses. Inconsistent or incorrect boundary conditions can lead to convergence problems.
6. **Contact and Interfaces**: If your model includes contact or interfaces between components, carefully review the contact settings. Adjust contact parameters, such as friction coefficients and penalty factors, to improve convergence.
7. **Restart Analysis**: Sometimes, it helps to perform a restart analysis. Save intermediate results and restart the analysis from a stable point.
8. **Convergence Criteria**: Adjust the convergence criteria in your analysis settings. You can make convergence criteria more or less stringent depending on the problem.
9. **Output Requests**: Reduce the number of output requests if you have too many. Excessive output requests can slow down the analysis and affect convergence.
10. **Check for Singularities**: Inspect your model for potential singularities or areas with extremely high stress or strain gradients. These can lead to convergence problems. Mesh refinement or localized modeling changes may be needed.
11. **Solver Options**: Experiment with different solver options. Abaqus provides various solver options, and some may work better for your specific problem.
12. **Consult Documentation and Support**: Refer to the Abaqus documentation and user manuals for guidance on specific convergence issues. You can also seek support from Abaqus technical support forums or communities.
It's essential to approach convergence issues systematically by isolating potential causes and gradually making adjustments. Keep in mind that convergence problems can sometimes be challenging to resolve, and it may require a combination of the above techniques to achieve a stable solution.
- Badges
- Science topic