I would like to continue my nonlinear heat transfer simulation from a certain step of an earlier (crashed) simulation. Especially when the earlier calculation stopped as result of divergence. In that case I would like to adapt the time steps and then continue from the step (n-1) before the crash. I read the user manual which states that 'save' and 'restore' can be used for steps as well as for filos files. However, where exactly should it be written?
My current code looks like this:
*HEATTR
MODEL MATRIX CAPACI CONSIS
BEGIN INITIA
NONLIN
TEMPER INPUT FIELD=1 FACTOR=1
END INITIA
BEGIN EXECUTE
BEGIN NONLIN
BEGIN ITERAT
METHOD NEWTON
MAXITE=20
TOLCON=1.E-6
END ITERAT
END NONLIN
SIZES 10.(5000)
SAVE LAST
END EXECUTE
BEGIN OUTPUT FEMVIE BINARY
BEGIN SELECT
NODES ALL
ELEMEN ALL / NODES ALL
STEPS 1:10/
END SELECT
TEMPERIAL COOR
FLUX INTPNT
END OUTPUT
*END
SAVE and RESTORE
The SAVE and RESTORE commands should be within the STEPS block. So instead of SIZES 10.(5000), write:
BEGIN STEPS
EXPLIC SIZES 10.(5000)
SAVE CONVER
END STEPS
The last converged step will be saved on a separate filos file named for example diana10.ff (for step 10). Then rerun the analysis from step 10 as follows:
BEGIN STEPS
RESTORE 10
EXPLIC SIZES 1.(5000)
SAVE CONVER
END STEPS
But if you need output for all loadsteps, you need to combine the output of both analyses, I think.
Does not work
Unfortunately this does not seem to work in *HEATTR ?