MSC Sinda > Solvers, Time Step and Result Accuracy > Introduction
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX''">XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX''">   
Introduction
Choosing an appropriate steady state or transient solver and good control constant values can be essential to obtaining an accurate MSC Sinda thermal solution. FE-based models can become quite large and the solver/control constant selection can have an even more dramatic effect on accuracy and/or run-time.
When running FE-based models, default solvers and control constants are pre-selected, but can always be overridden by the user. Typically, SNSOR is the default steady state solver, and SNDUFR the default for transient. Though these defaults usually suffice, the need to make different choices is not uncommon, most often caused by radiation or some other nonlinearity in the model.
The user is encouraged (urged) to review Section 3.7 in the MSC Sinda User’s Guide and Section 2 in the MSC Sinda Library Reference Guide. These sections contain discussion of the various control constants and solution methods.
The user should never trust the first apparently successful MSC Sinda solution for any real-work model. Whether the method was steady state or transient, the solution needs to be verified before the model can be trusted. Several methods are available for verifying the solution.
Obtain nearly identical results with a different numerical method
Obtain nearly identical results with tighter convergence (steady state or transient)
Obtain nearly identical results with a smaller time step (transient)
Obtain nearly identical results with more rays or different random seeds (external radiation solvers)