Please note, that in the actual version, the error on the element level is evaluated using default integration rule. For example, in case of ZZ error estimator, the error (L2 or energy norm) is evaluated from the difference of computed and ``recovered'' stresses, which are approximated using the same interpolation functions as displacements). Therefore, in many cases, the default integration rule order is not sufficient and higher integration must be used on elements (consult element library manual and related NIP parameter).
The sets of parameters (errorestimatorparams field) used to configure each error estimator are different
[regionskipmap #(ia)] | |
vartype #(in) | |
minlim #(rn) maxlim #(rn) | |
mindens #(rn) maxdens #(rn) defdens #(rn) | |
[remeshingdensityratio #(rn)] |
[regionskipmap #(ia)] | |
normtype #(in) | |
requirederror #(rn) | |
minelemsize #(rn) |
Borek Patzak