where's master nodes in constraint equation?

rhoka

hi All,

currently i try modeling shell with connection to master nodes to apply nodal forces (normal, bending and twisting)

first try only normal Fz i modeling, the solver run without error but the results are zero at all.

where i goes wrong?

Victor

Thanks for pointing this out, rhoka. There's a limitation in that a node typically has to belong to an element to be included in the solution. I hope to address this in the next version.

To work around it, just create an element on that lone node. I did it here with a beam. It doesn't matter that the material properties aren't right for it because it's not carrying any load.

I also adjusted the coefficient of the slave DOF in the MPC to give it the average z displacement of other 16 nodes.

rhoka

thanks Victor for review & correcting my models,

i just have a little bit confused with equation options in MECWAY, usually rigid arm connection defined by list of slaved node and force applied at master. last the user required to define DOF's being considered.

Victor

Mecway's constraint equations are very low level so you can only specify DOFs, not entire nodes. To do all directions, you would have to make 3 equations, all identical except using different DOFs - x,y,z. This is where the text field comes in handy since you can copy and paste to duplicate it.

rhoka

thanks you Victor for explanations and implemented it.

may you wish to give additional information in documentation and provide some examples how to use MPC since it's essential in FEA modeling.

Victor

Good idea. I may need to add some more sophisticated examples such as joining a shell to a solid or spreading a load like you're doing. There's a very simple one at the end of the tutorials manual but it's just coupling two nodes.

davefo

Is there more guidance on using equations?

I was hoping to be able to define loads relative to other loads using equations but don't know where to start.

As an example if I define a load in the X axis of +50kN and apply it to one edge of a plate then have an equation in another defined load that is -ve of the related load applied to the opposite end of the plate.

 

Victor

Not much, though they're much the same as constraint equations in other software. You can use them to distribute forces by making the displacement of one node equal to an average of the displacements of two or more other nodes a bit like the linkage mechanism the distributes force along a windscreen wiper blade. 

The attached model shows a single 2 N force being split into two 2N forces in opposite directions and causing different displacements.

Another way to have equal magnitude forces in different directions is to use a single pressure load applied to both equal-area surfaces.

Please try to start new questions in new threads so it's easier for other people to find from the subject line.