Use the following best practices to improve the accuracy of transient coupled analyses.
- Define an appropriate coupling step size.
Set the coupling step size to resolve the coupling-related timescales. For example, with thermal couplings, choose a coupling time step size that is long enough to resolve the thermal timescales.
If a participant's timescales are slightly smaller (for example, between 1x and 100x) than coupling-related timescales, then reduce the coupling step size to resolve to the smallest timescale in the participant physics. This often occurs for fluid-structure interaction applications.
If a participant's timescales are significantly smaller (for example, greater than 100x) than coupling-related timescales, then run that participant using a steady analysis. This often occurs for electromagnetic/thermal applications.
- Define motion consistently across coupling participants.
When motion is defined on a coupled region ensure that the motion is defined consistently for both coupling participants. (Currently, this is supported only for electromagnetic-thermal analyses between Maxwell and Fluent coupling participants.)
- Assess the convergence status of the analysis.
Identify and address any issues that may be affecting the convergence of the analysis.
Slow convergence:
If convergence is not realized for each coupling step within approximately 5 – 10 coupling iterations, then consider reducing the coupling time step size or introducing stabilization methods such as ramping and under-relaxation.
Participant-specific solution stabilization methods, such as those offered by Fluent, may also be employed. For details, refer to the corresponding participant product documentation.
Strong non-linearities:
If there are strong non-linearities within the coupled solution, then convergence within approximately 5 – 10 coupling iterations may not be possible. If this is suspected (and the previous recommendations are not sufficient to allow convergence), then increase maximum number of coupling iterations per step.
Note: Although maximum number of iterations may be used to limit the amount of work done per step, this is generally not recommended.
Premature convergence:
In some cases, convergence is realized prematurely (for example, within one iteration) because the non-linearities in the coupled solution have not yet been activated. If this is suspected, then set a minimum number of coupling iterations per step to trigger the non-linearities and force adequate resolution.