Hi Udi: The strain should be calculated by definition. For example the linear definition is just [(change in length)/(original length)]. You should be able to probe the displacement (say UX) of two nodes on an element. The ratio of the difference in the displacements, divided by the original (undisplaced) distance between the two nodes, should give the strain between the two nodes. For the whole element it's probably an average between all nodes that belong to the same element. In other FEA software I have been able to connect a soft spring of known length to two points on the surface of a finite element model (to simulate a strain gage), and that worked to show the hand-calculated strain in the soft spring agreed with the strain showing on the finite element color plot. I hope that helps a little.-Tony
Thanks for your reply.
We know the practice of linear strain calculation. The question was how does Simulation calculate the 10 Node Tetrahedron element strain.
I wander how it works in simulation and if the following description is how it works.
For element strain, is it work by developing sum of all dLi/Li where Li is distance between all nodes couples in an element and then find a "mean"?
Then for Node strain, is it develop sum of all dLj/Lj where Lj is distances to all "neighbor" nodes, and then find a mean?
This can be done also for X,Y and Z directions...
Simulation solves for displacement at the nodes. It then calculates the strains from the nodal displacements. I am not sure how it calculates the element strains, if it is just the average of the 10 nodal strains for the 10 node tet or if it calculates the centroid of the element and extrapolates the strain from the nodes.
Also not sure how it determines the strain between the nodes when you do a strain plot using node values. If it is linear between the nodes or uses adjacent nodes for smoothing.
What I do if I am comparing FEA to srtain gage data is create split lines in my model so that I have a node exactly where the strain gage is located.