A general question:
We're using COMSOL to solve a system of PDE's and comparing them to our in-house code. The variables in said PDE's are inherently in different scales (i.e. three variables are O(1) while others are O(e-5)). In our in-house code (which uses finite elements), this scaling issue leads so some terrible condition numbers. However, COMSOL is able to give answers that are in line with what we expect. My question is this: does COMSOL detect a scaling issue and, if so, how does it correct for it? For reference, the scaling option is set to: "Automatic"
We're using COMSOL to solve a system of PDE's and comparing them to our in-house code. The variables in said PDE's are inherently in different scales (i.e. three variables are O(1) while others are O(e-5)). In our in-house code (which uses finite elements), this scaling issue leads so some terrible condition numbers. However, COMSOL is able to give answers that are in line with what we expect. My question is this: does COMSOL detect a scaling issue and, if so, how does it correct for it? For reference, the scaling option is set to: "Automatic"
