Soil reaction on anchor plate

[hansini]

Dear all,
I have a very basic problem regarding how abaqus calculate reaction forces of nodes with displacement boundary conditions. I am analysing a vertical anchor plate buried in sand. The anchor is given a 0.01 displacement. Reaction forces I get after geostatic step and load step are as below.But reaction forces of corner nodes are lower than middle nodes. I wander why? Geostatic stresses in soil are correct. Also during the analysis the direction of RF1 in left edge change directions. I am expecting a triangular RF distribution as theory suggests. Can anyone help me.
Thanks in advance

 

[Twana]

Hello Hansini,
The process is a little bit complicated and may be difficult to me to explain it in English, but I will do my best. You are applying displacements to the model therefore an expected pattern or distribution of the forces will not be achieved. The forces are created by applied displacements and if you can image the nodes in the model, when only one node, say node A, is applied by a specific displacement (say 0.1m), this node will affect its neighboring nodes and create some displacements in them as well. If B is a neighboring node to A, a displacement of (say) 0.01m may be created in B because of the node A applied displacement. When Abaqus calculates the reaction force of node B, it will calculate the force required to displace node B by 0.09m because 0.01m is already created by node A displacement. This effect is possible for all other nodes and sometimes a negative (opposite) force might be created in a node in order to achieve the required applied displacement .

 

[hansini]

Hi,

Thank you for the reply. If I understood you correct, although I applied equal displacements to each node, reaction forces calculated by ABAQUS consider different displacements at nodes. I am sorry if this is a dumb question. Does not ABAQUS calculate reaction forces by multiplying the inverse of the stiffness matrix by the displacement vector? How about the resultant of the reaction force?

 

[Twana]

Hi,

You are right for the resultant force but the individual forces do not follow the displacement pattern. A node in a continuum model is affected by other surrounding nodes. The point here is you must get the required displacements by applying all nodal reaction forces (in all directions) to the model. In addition, the material should be in the elastic range (there should be no plasticity).