In achieving a good convergence plot, does it require to be all in straight plot? Or adding a number of iterations is required to get more plot stability?
Solved! Go to Solution.
Automatic Convergence Assessment assesses convergence, and stops the analysis when the solution is no longer changing. Typically we would run the analysis out enough iterations such that the Automatic Convergence Assessment kicks in.
If you're trying to manually assess Convergence to determine how much further to run the solution, the following discussion applies:
The average value plots can be displayed for each degree of freedom (velocity, pressure, temperature, etc) during this solve. A converged analysis will have the major velocity direction (u,v, or w) and pressure varying only an acceptable amount from the mean from each iteration to the next. Remember to look at the components individually since the graph is self normalizing for each component individuallly. The avergage value plot shows this during the solution. Monitor your maximum and minimum average values in these terms for a deviation that is within acceptable amounts for your deviation criteria.
Next review the summary file information. Look for acceptable mass balance from inlet to outlet.
Review the results and ensure they appear physically correct. Typically, errors in an analysis will be most pronounced in the pressure. Use isosurfaces to detect abnormally high or low pressures that may be indicative of a mesh issue. For example, an iso surface applied to a model with a mesh issue will typically appear immediately on the nodes that are a problem in the mesh. You can either refine the mesh or the geometry in the area of the isosurface at that point.