Solver for large models

General Questions
21

Little correction model 1,76 million of elm total time ca 12,2 hour sory for earlier wrong data
best regards

22

could you be more specific about tie constrains:
what are general behavior I mean nodes form slave surface are projected on master surface or in opposite direction ? how it is linked ? what is role of adjustment icon ? mesh for slave should be more dense or mesh of master surface should be more dense ? What role plays search distance acc to master and slave surface ?
best regards

23

Slave nodes are connected to master faces if the tolerance (orthogonal distance between the nodes and the adjacent face) criterion is fulfilled. When adjustment is enabled, the position of the slave nodes is modified such that they lie exactly on the master surface. When it comes to the assignment of master and slave surfaces, usually the same approach as with contact is advised - rigid/stiffer, more coarsely mesh surface should be master.

24

I have situation that slave surface is touching master surface , but tie constraint is not working (one from ca 350 ) search distance is set to 2 mm , Iā€™m wondering if I should use adjust option , I have question if slave surface should be smaller in projection direction to master surface I mean slave surface can not protrude from master surface

Best regards

image001.jpg

25

does master surface layout have inflence on tie constraint ? is it connection node to node or in other way ?

26

Apart from what I said, thereā€™s only this rule:

the projection of the slave node onto the master face must lie within this face or at any rate not
farther away (measured parallel to the face) than the default tolerance just defined

Tie constraint is realized as MPC connecting slave nodes with master faces. You can find the details in CalculiX Userā€™s Manual.

27

I would also advise reading the CalculiX manual. There are many interesting and informative things written and explained in it.

28

I did it but without success. after recalculation I have noticed lack of connection between plates

29

Will be in near or far future improvement for large models computations ? . As I wrote earlier I have problems with big models to solve , shell idealisations are not working some bugs and so on. I have used solver pardiso and after many failures I tried iterative choleski and Iterative scaling but all without success . It was solved 2 times but some constrains were not working

30

Thatā€™s a question for Guido Dhondt - CalculiXā€™s dev.