A question about Kinematic (RBE2) and Distributing (RBE3) couplings

[ymn]

Hello,

As the title says I have a question about couplings which can be something very fundamental yet it is puzzling me. Let's assume that we have a classic MPC with one reference point and on this reference point we are applying a constrain. When we try to solve this problem in a static case with Kinematic coupling the program solves it well. But when we change the coupling type to Distributing, the program does not solve the problem. As you can guess in this case there is one dependent node and bunch of independent nodes and when we put constraint on the dependent node, this constraint does not apply on independent nodes and rigid body motion occur. (I hope I am right until this point.) But when we apply force on the reference point and apply constraint at somewhere else on the model, the program solves it in both cases with obviously different results. At this point, for the distributing coupling usage, how the force which is applied on the dependent node can be transferred to independent nodes, why would independent nodes be influenced by dependent node now ? Because we observed that dependent node does not affect independent nodes in the previous case when we applied constraint on dependent node. So what has changed ? What is the difference between applying constraint and applying force onto a dependent node of a distributing coupling which in result makes the difference between solveable and unsolveable problem.

Thanks

 

[FEA way]

Can you attach some pictures showing such model ? By constraint do you mean boundary condition ?

 

[ymn]

yes exactly, I mean boundary condition, pin, encastre or whatever.

 

[FEA way]

These two coupling types differ in the way they work. Kinematic couplings eliminate degrees of freedom while distributing couplings are applied in average sense and may lead to indefinite stiffness matrix.